Islamicate Celestial Globes
Their history, Construction, and Use
            


EMILIE SAVAGE-SMITH with a chapter by ANDREA P. A. BELLOLI
SMITHSONIAN  INSTITUTION  PRESS

SERIES PUBLICATIONS OF THE SMITHSONIAN INSTITUTION
     Emphasis upon publication as a means of "diffusing knowledge" was expressed by the first
Secretary of the Smithsonian. In his formal plan for the Institution, Joseph Henry outlined a
program that included the following statement: "It is proposed to publish a series of reports,
giving an account of the new discoveries in science, and of the changes made from year to year
in all branches of knowledge." This theme of basic research has been adhered to through the
years by thousands of titles issued in series publications under the Smithsonian imprint,
commencing with Smittisoriian Contributions to Knowledge in 1848 and continuing with the
following active series:
Smithsonian Contributions to Anthropology
Smithsonian Contributions to Astrophysics
Smithsonian Contributions to Botany
Smithsonian Contributions to the Earth Sciences
Smithsonian Contributions to the hAarine Sciences
Smithsonian Contributions to Paleobiology
Smithsonian Contributions to Zoology
Smithsonian Folklife Studies
Smithsonian Studies in Air and Space
Smithsonian Studies in f^istory and Technology
     In these series, the Institution publishes small papers and full-scale monographs that report
the research and collections of its various museums and bureaux or of professional colleagues
in the world of science and scholarship. The publications are distributed by mailing lists to
librahes, universities, and similar institutions throughout the world.
     Papers or monographs submitted for series publication are received by the Smithsonian
Institution Press, subject to its own review for format and style, only through departments of the
various Smithsonian museums or bureaux, where the manuscripts are given substantive review.
Press requirements for manuscript and art preparation are outlined on the inside back cover.
Robert McC. Adams
Secretary
Smithsonian Institution

SMITHSONIAN    STUDIES    IN    HISTORY    AND    TECHNOLOGY   •    NUMBER    46
     Islamicate Celestial Globes:
Their History, Construction, and Use
Emilie Savage-Smith
with a chapter on iconography by
Andrea P.A. Belloli

Smithsonian Institution Press
Washington, D.C.
1985

ABSTRACT
Emilie Savage-Smith, with a chapter by Andrea P.A. Belloli. Islamicate Celestial Globes: Their
History, Construction, and Use. Smithsonian Studies in History and Technology, no. 46, 354
pages, 87 figures, 7 tables, indices, 1985.—Islamicate celestial globes made as early as the
eleventh century are found in museums and private collections today. There are also references
in classical Greek and Roman literature to earlier globes that are no longer extant. These
globes are of interest to the history of astronomy, of art, and of technology.
   The globe presently in the National Museum of American History of the Smithsonian
Institution, which is a fine example of a seventeenth-century Mughal Indian globe, was selected
for detailed analysis and serves as the focus for this monograph. The first part of the study
compares this particular globe with other known Islamicate globes and places the development
of such globes within the historical perspective of the earlier Greco-Roman world from which
it drew many of its traditions. An historical survey is given of all references and artifacts from
the Greco-Roman and Islamic world that can have bearing on our knowledge of the design,
construction, and use of such globes. The nature and general characteristics of three basic
types of Islamicate celestial globes, and their probable uses as well as methods of construction,
are the subjects of the second chapter of the study. Photographs of selected Islamicate globes
from the thirteenth to the nineteenth centuries, as well as line drawings based on written
descriptions, accompany the historical and analytical discussion.
   The fourth chapter on iconography analyses the constellation figures on the Smithsonian
globe from the perspective of an art historian. This chapter was contributed by Andrea P.A.
Belloli.
   The second major part of the study presents a discussion of the star names engraved on the
Mughal globe, tracing the origins of the terms in Greek mythology or early Bedouin constel-
lation outlines. The discussion of each constellation is accompanied by a photograph of the
constellation as depicted on the Smithsonian globe. An account of lunar mansions is included
as background to early Bedouin asterisms, which greatly affected later Islamicate star names
and eventually "modern" western star names.
   The sixth section presents an extensive descriptive catalogue of the 126 Islamicate celestial
globes known to scholars prior to 1982. The references in the other sections to particular
globes are keyed to the entry numbers in this catalog. Following the catalog are tables comparing
the features of the globes and transcriptions of the signature inscriptions. Six entries (Nos.
127-132) were added to the catalog while the study was in press.
Copyright © Smithsonian Institution     1984
OFFICIAL PUBLICATION DATE is handstamped in a limited number of copies and is recorded in
the Institution's annual report, Smithsonian Year. COVER: Globe No. 38, the Smithsonian Globe.
Library of Congress Cataloging in Publication Data
Savage-Smith, Emilie.
Islamicate celestial globes, their history, construction, and use.
(Smithsonian studies in history and technology ; no. 46)
Bibliography: p.
Includes index.
1. Globes, Celestial—History. 2. Globes, Celestial— India—History. 3. Astronomy, Arabic—
History.
  1. Belloli, Andrea P.A. II. Title. III. Series.
QB66.S28    1984    522'.7    83-20197
  
Contents
FOREV^^ORD		iv
PREFACE		vi
INTRODUCTION		1
Historical Overview
1. A History of Celestial Globes in the Greco-Roman and Islamic
Worlds		3
2. The Nature, Use, and Construction of Islamicate Celestial Globes	61
The Smithsonian Globe
3. Description and Attribution		96
4. The Constellation Figures on the Smithsonian Globe		99
5. A History of Star Names, Based on the Smithsonian Globe		114
Classical Greek and Pre-Islamic Sources		114
Lunar Mansions		119
The Forty-eight Constellations		132
Extant Islamicate Celestial Globes
6. Descriptive Catalog		213
7. Tables of Basic Characteristics of the Globes		278
8. Major Inscriptions on the Globes		284
NOTES		297
SELECTED BIBLIOGRAPHY  		312
INDICES		329
Manuscripts Cited		329
Collections		330
Makers   		334
Patrons and Proprietors   		336
Arabic Terminology		337
Greek Terminology		344
General Index		345
Ul

FOREWORD
  This comprehensive study of Islamicate globes was initiated in the most
casual manner one spring day more than ten years ago in my office as Deputy
Director of the National Museum of History and Technology (now the
National Museum of American History).
  The Museum had just acquired a most interesting celestial globe but there
was no one on hand who could read its inscriptions. At the time there was
no one on the professional staff with a knowledge of the Arabic language.
The lack was soon remedied, however, with the discovery that Mrs. Emilie
Savage-Smith, a Smithsonian post-doctoral Fellow working in the Division of
Medical Sciences, was familiar with the language. When called upon, she
readily provided the information needed.
  But the story did not end there. Mrs. Savage-Smith became intrigued with
the globe, and her preliminary research revealed that this was a category of
instruments that had been neglected by scholars. Realizing that a study of
Islamicate globes presented an unusual opportunity to make a contribution
to the literature, she undertook a more detailed examination of the example
in the Museum, and then began a search for others with which to compare
it. Little did she realize at the time that she was launched on a project that
was to preoccupy her for the next decade and would take her to far places.
  As a first step, arrangements were made with the Smithsonian's Conser-
vation-Analytical Laboratory to make a thorough technical analysis of the
Museum's globe to determine the techniques of construction and nature of
the metal. The results of this analysis are included in Chapters 2 and 3 of
this paper.
  Meanwhile the author sought out other globes in American public and
private collections, and then extended her search to repositories in England,
France, Egypt, and other countries. As each globe was examined, photo-
graphed, and documented, the information collected was made part of a case
study by means of which it became possible not only to establish the dates
and places of origin, but also to relate Islamicate globes in historical perspec-
tive with the production of scientific instruments derived from the Greco-
Roman tradition in the Islamic world.
  The project became a rewarding adventure in historical detective work as
the author continued to locate and document the 126 globes that are
cataloged in this work, consulting with authorities on Islamicate astronomy
and instrumentation in the United States and abroad. Almost a full ten years
passed before the search was completed and the findings brought together
for publication in this paper.
This monograph may not be the last to be written on the subject, but as
iv

NUMBER 46

the first it breaks new ground. Not only will it supplement studies of science
in Islam, but will prove to be an invaluable resource for historians of science
and of scientific instrumentation in general.
Silvio A. Bedini
Keeper of the Rare Books
Smithsonian Institution

PREFACE
  This historical survey of the design and construction of Islamicate celestial
globes and their Greco-Roman antecedents pays particular attention to the
products of seventeenth-century Mughal India. The globe that serves as the
focus of the study is one of the finest examples of precision workmanship
among all the 126 extant Islamicate celestial globes and is, moreover, a
seamless hollow cast globe. The fact that many Islamicate globes are hollow
seamless metal spheres and that this technique was a speciality of a particular
workshop of instrument makers in Lahore seems not to have been noticed
until this study was begun 12 years ago.
  In order to properly assess the products of the Lahore workshop, as much
information as possible was needed concerning other Islamicate celestial
globes. This author's effort to assemble this information resulted in the
catalog of all Islamicate celestial globes known as of the spring of 1982 to be
extant. The many references to celestial globes throughout this study are
keyed to the entry numbers in this catalog, which comprises chapter 6 of this
study. It is hoped that with this publication even more celestial globes now
in small museums and private collections will come to light and be made
available for study by scholars so that a fuller picture will emerge of the
development of the design and construction techniques. While the study was
in press six entries (Nos. 127-132) were added to the catalog; these globes
do not figure, however, in the general analyses and comparisons made
elsewhere in the book.
  In order to appropriately designate the non-religious aspects of the com-
plex of cultural traditions in a society historically associated with the religion
of Islam, some term other than "Islamic" is needed. I have chosen to use
Islamicate, for whereas the word "Islamic" refers to subjects directly related
to, growing out of, or affected by the religion of Islam, Islamicate can be
used to refer to objects or cultural features that are not related directly to
the religion but are often based on traditions taken over from other cultures
and nurtured and developed by Muslims and non-Muslims alike. When
speaking of science and technology in the Islamic world, Islamicate is a
particularly appropriate term, as this topic of celestial globes illustrates. These
globes represent a tradition of instrument design inherited from the Hellen-
istic, Roman, and Byzantine worlds, and except for minor points of design
and considerable progress in construction techniques, that tradition, unaf-
fected by religious belief or dogma, remained essentially unchanged through
the end of the nineteenth century. Among the makers of the instruments
and the authors of treatises on celestial globes there were Christians, Muslims,
and Hindus, and the languages used in their works included Arabic, Persian,
Indo-Persian, Ottoman Turkish, and even Sanskrit. Yet all are representative
of a secular culture shared by an enormous area from Spain to the Indus that
vi

NUMBER 46	vii
was marked by the domination, at one time or another, of the political-
religious structure of Islam, in what has been loosely termed the Islamic
world, or Islamdom. All the extant globes, whether from Spain or India,
whether made by Muslim or non-Muslim, whether inscribed in Arabic,
Persian, or Sanskrit, reflect an aspect of the culture of Islamdom.
  The examination of the Mughal celestial globe's design and construction
and that of its historical antecedents came about after the National Museum
of American History (then the National Museum of History and Technology)
acquired (at first on loan) the anonymous and undated globe in 1972. That
globe became the primary object of this investigation. My study and analysis
of the globe was instigated and encouraged by Silvio A. Bedini, presently
Keeper of the Rare Books and former Deputy Director of the National
Museum of History and Technology; Dr. Uta C. Merzbach, Curator of the
Division of Mathematics; and Deborah J. Warner, Associate Curator of the
Division of Physical Sciences, where the globe is presently located. The
laboratory analysis of the globe was carried out by Maurice Salmon, formerly
of the Conservation and Analytical Laboratory of the National Museum of
American History. For reasons completely beyond the control of the present
author, the analysis has not been published. It is hoped that it will be
subsequently published in detail and interpreted by someone more knowl-
edgeable then myself in such matters.
  I was most fortunate in being able to have an art historian, Andrea P.A.
Belloli of the Los Angeles County Art Museum, contribute to this study a
chapter on the iconography of the Smithsonian globe. Using art historical
evidence she analyses the products of this Mughal workshop and attributes
the globe to one particular designer and scientific instrument maker within
this large workshop—arriving at the same conclusion to which I was forced
independently by the examination of the linguistic and technical features of
the globe. I wish to express my special thanks to her for undertaking this
chapter and so rigorously meeting all appointed deadlines. She bears no
responsibility, however, for whatever errors or misinterpretations are to be
found elsewhere in this study.
  The nature of the study has necessarily involved a large number of people
who have given assistance in various ways. The author is particularly indebted
to the enormous assistance given through conversations by Francis R. Mad-
dison, Curator of the History of Science Museum, University of Oxford, who
made many valuable suggestions based on his extensive knowledge of Islam-
icate scientific instruments. In addition, both he and Alain Brieux of Paris
most generously placed at my disposal all the files for their comprehensive
repertoire of signed or dated Islamicate scientific instruments, which they
are preparing jointly.
  Among the persons who gave liberally of their time to supply needed
information and answer queries I would particularly like to mention Anthony
J. Turner of Le Mesnil-le-Roi, France, and Roderick and Marjorie Webster,
Curators of the Adler Planetarium, Chicago, as well as acknowledge the kind
assistance of R.G.W. Anderson, formerly Assistant Keeper of the Department
of Technology, the Royal Scottish Museum, Edinburgh, and now with the
  
viii	SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY
Science Museum, London; Madame Marthe Bernus-Taylor, Conservateur a
la Section Islamique, Musee du Louvre; D.J. Bryden, formerly of the Whipple
Museum of the History of Science, University of Cambridge; Professor D.J.
de Solla Price of Yale University whose untimely death occurred just before
publication; Dr. Muammer Dizer, Director of the History of Science Museum,
Kandilli Observatory, Istanbul; Michael Robinson, Assistant Keeper, Depart-
ment of Art, Ulster Museum, Belfast; J.S. Simons, Technician, Museum of
the History of Science, University of Oxford; Professor Cyril Stanley Smith,
Massachusetts Institute of Technology; and Christopher Terrell, Curator of
Hydrography of the National Maritime Museum, London. Helpful sugges-
tions regarding one or more of the signature inscriptions were given by Dr.
Ismail K. Poonawala, Department of Near Eastern Languages, UCLA, and
James R. Reid and Mehri Reid of the Von Grunebaum Center for Near
Eastern Studies, UCLA, and some brief linguistic enquiries were answered
by Julian Baldick of Wolfson College, University of Oxford; Beatrice F. Manz
and Menuchechr Mohandessi of Harvard University; Serge Obolenski of the
U.S. Foreign Service; and Mary Chase Smith of the University of North
Carolina, Chapel Hill. The author also wishes to thank Judy Erickson von
Gunten, who translated Spanish texts for my use and drew sketches of
constellation figures from several globes.
  Many museums and private collectors assisted by graciously allowing me
to examine the globes or by supplying photographs: Madame Marthe Bernus-
Taylor, Conservateur a la Section Islamique, Musee du Louvre; Alain Brieux
of Paris; Michael V. Butler, Curator of Physics, Cranbrook Institute of
Science, Bloomfield Hills, Michigan; Marcel Destombes of Paris, who kindly
let me examine his very important personal collection of Islamicate globes;
Jan Dolman, Assistant Curator, the Time Museum, Rockford, Illinois; Wa-
fiyya Ezzi (deceased), former director, and Abu El-Ra^Cif Ali Yousuf current
director, the Museum of Islamic Art, Cairo; Mr. and Mrs. David H.H. Felix
of Philadelphia; Kenneth A. Lohf, Librarian for Rare Books and Manuscripts,
Butler Library, Columbia University, New York; A.E.R. North, Department
of Metalwork, and A.G. Mitchell, Indian Section, of the Victoria and Albert
Museum, London; Miss Dorothy E. Miner, Librarian, Walters Art Gallery,
Baltimore; R.H. Pinder-Wilson, former Deputy Director of the Department
of Oriental Antiquities, British Museum, London; D. Morgan Rees (de-
ceased), former Keeper of the Department of Industry, the National Museum
of Wales (now the Welsh Industrial and Maritime Museum, Amgueddfa
Diwydiant a Mor Cymru), Cardiff; A.N. Stimson, Curator of Navigation, and
Miss Drucilla Bates, of the National Maritime Museum, London (Greenwich);
A.C. Thompson, Keeper, and R.G.W. Anderson, Assistant Keeper, of the
Department of Technology, the Royal Scottish Museum, Edinburgh; The
Reverend F.J. Turner, S.J., Librarian, Stonyhurst College, Lancashire, Eng-
land; and Antony Vincent, Department of Astronomy and Geophysics, the
Science Museum, London.
  In addition to those mentioned earlier, others at the Smithsonian have
given advice and assistance, including Esin Atil of the Freer Gallery of Art;
William T. Chase of the Conservation Laboratory of the Freer Gallery of
  
NUMBER 46	IX
Art; George Norton, museum specialist of the Division of Physical Sciences
of the National Museum of American History; and Robert Organ, Chief of
the Conservation and Analytical Laboratory of the National Museum of
American History. I also wish to express my appreciation to Rebecca Curzon,
publications coordinator of the National Museum of American History, who
consistently showed interest and courtesy in handling the publication of this
study; Alice Gergely of the National Museum of American History who
patiently typed the tables; and Theresa J. Slowik, technical publications editor
of the Smithsonian Institution Press, who meticulously groomed and pre-
pared the manuscript for press.
  There are, of course, many scholars whose previous writings were of
fundamental usefulness to this study; among them I would particularly
mention P. Kunitzsch, whose studies of Arabic star nomenclature must be
extensively employed in any discussion of the topic. Furthermore, any ex-
amination of Islamicate celestial globes is necessarily indebted to the studies
of some of the early globes by Marcel Destombes, who at the time of his
recent death was nearing completion of an analysis of Islamicate celestial
globes prior to AD 1500, giving special attention to the star names.
  In addition to the photographs of the Smithsonian globe prepared by the
photographic laboratory of the National Museum of American History, there
are other photographs used in this publication, a number of which were most
generously supplied by Alain Brieux of Paris, while the Musee du Louvre,
the Whipple Museum of the History of Science, the National Maritime
Museum (London-Greenwich), the Royal Scottish Museum, the Time Mu-
seum of Rockford, Illinois, and Columbia University also furnished photo-
graphs of items in their collections. Three photographs of unknown origin
are drawn from the files of the History of Science Museum, Oxford, and the
Whipple Museum of the History of Science, Cambridge. The author wishes
to thank the Adler Planetarium, Chicago; Alain Brieux of Paris; the British
Museum; Columbia University, New York; the Cranbrook Institute of Sci-
ence, Michigan; Musee du Louvre; the Museum of the History of Science,
Oxford; the National Maritime Museum, London (Greenwich); the Royal
Scottish Museum; Stonyhurst College, Lancashire, England; the Time Mu-
seum, Illinois; the Victoria and Albert Museum; the Welsh Industrial and
Maritime Museum; and the Whipple Museum of the History of Science, for
granting permission to publish photographs.
  To the constant support and inspiration of M.B.S., friend, lover, scholar,
crank, this study is dedicated.
Emilie Savage-Smith



     Islamicate Celestial Globes:
Their History, Construction, and Use
Emilie Savage-Smith
INTRODUCTION

  The globe that is the particular focus of the
study is an anonymous and undated hollow metal
seamless sphere. It shows about 1022 inlaid silver
stars and has 48 constellation figures and numer-
ous star names engraved on its surface. The
attribution of this globe to a particular maker
was arrived at independently by an historian of
science and an art historian (Andrea P.A. Belloli).
Thus, the conclusions given here may serve as a
case study of how scholars having different view-
points and employing different methodologies
can approach the same body of evidence and
arrive at the same conclusions without knowl-
edge of one another's findings. Both scholars,
one using art historical evidence and the other
linguistic and technological arguments, agreed
upon a definite attribution to one particular de-
signer and scientific instrument maker within a
large workshop. The globe is a fine example of
the Mughal Indian celestial globes of the seven-
teenth century, a period that has received rela-
tively little attention and is thoroughly within the
tradition of medieval Islamicate scientific
thought. This particular globe is, moreover, one
of the finest examples of precision instrument
making among all the extant Islamicate celestial
globes. This globe is also a good example of
constructing a globe by casting a hollow seamless
Emilie Savage-Smith, Gustave E. von Grunebaum Center for
Near Eastern Studies, University of California, Los Angeles, CA
90024.

sphere using the cire perdue (lost wax) method—
an intricate process in which a particular family
workshop of astrolabe makers in Lahore excelled
during the sixteenth and seventeenth centuries.
There is some evidence that the technique might
have developed outside of northwestern India
and as early as the thirteenth century, though
unfortunately all the evidence for this theory
rests on globes whose authenticity is question-
able. The seamless nature of certain Islamicate
metal globes was not recognized until this study
was undertaken. These Islamicate celestial globes
are the only known examples of hollow cast
seamless spheres.
  NOTE ON TRANSLITERATION AND DATES.—
With the Arabic I have tried to follow for the
most part the orthography rather than the pro-
nunciation, emphasizing the element of reversi-
bility in the transliteration. I have employed bas-
ically the Library of Congress System with a few
of the variations given by Hodgson (9-10). A
macron rather than a circumflex has been used
to indicate long vowels. The Arabic article is
never assimilated to the following noun. The
douh\eyd^ after kasrah is written as iy rather than
iyy, and the hamzah at the start of a word is
omitted. The final a written in Arabic with ayd^
{alif maqsurah) is written as -a. The feminine
ending of a noun or adjective is indicated by a
td'' marbutah, which is a combination of an h with
the diacritical marks of a ^ and pronounced some-
times as an h and sometimes as a t and sometimes
  
1

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

not at all. In this study it is usually written as an
h, except when the word occurs before a noun
in the genitive or construct state {iddfah), in
which case it is written as a t. For example, by
itself the word for "work" or "opus" would be
san"ah, but in the statement "the work of Yiinus
ibn al-Husayn al-Asturlabl" it would be written
as san"at Yunus ibn al-Husayn al-Asturldbi.
  In the case of transliterating Persian and Indo-
Persian, 1 have tended to transliterate the words
with the same alphabet used for the Arabic with
the addition of four letters p, ch, zh, and g. In
this procedure I deviate from the system em-
ployed by the Library of Congress. Thus I write
Diya" al-DTn instead of Ziya^ ad-DTn. It is with
some regret that I do this, for I recognize that
Persian is a separate language despite the fact
that it is written with the Arabic alphabet. The
choice was made, however, because I am a better
Arabist than Persianist and because Arabic is the
basic language of all the globes, except for the
two with Sanskrit inscriptions, with the Persian
and Indo-Persian (and in one case Ottoman
Turkish) occurring, except for a few minor
words, only in the signature inscriptions. The
reader will find in Chapter 8 all the signature
inscriptions on the globes written in Arabic and
Persian alphabets.
  Throughout the study, dates given in the Is-
lamic era, that of the Hijrah, are designated by
H, while those of the common Christian or Gre-
gorian calendar are indicated by AD (anno Dom-
ini). Whenever a reference occurs to a century
with no designation of calendar, the common
Christian era is intended. Because this study is
written in a European language, 1 have tended
to give only the equivalent Christian dates when
greater specificity seemed unnecessary. When
referring to the date of a particular product, I

place the Hijrah date first followed by the Chris-
tian date. Thus in the historical sections a date
of a product will have the form 622 H/AD 1225-
1226. The actual date on Islamicate scientific
instruments and treatises is not given in the
Christian era (with one exception on a single
globe), but in that of the Hijrah or some other
Eastern calendar (see Table 5 in Chapter 7 for
all the calendrical eras used on Islamicate globes).
  NOTE ON REFERENCES AND BIBLIOGRAPHY.—
The references in this book use a shortened form
that has been devised to overcome the difficulties
of dealing with classical texts and scientific works
within the same system. Each work listed in the
Bibliography is alphabetized under its shortened
form—usually the author's or editor's name, and
either an abbreviated title or the year of publi-
cation.
  These short forms are used in the notes to
Chapters 1-5 and in the Citations sections in the
Catalog. In the discussion of the Constellations
in Chapter 5, where a limited number of treatises
are used repeatedly, the short forms are used in
the text itself and refer the reader directly to the
Bibliography.
  For classical works, the traditional system of
book and line numbers has been used; e.g.. Ho-
mer//., 11,15 refers to Book II line 15 of Homer's
Iliad. For Aratus's poem, Phaenomena, line num-
bers are given. Thus these references can be
found easily in any edition. Where a particular
translation has been quoted, the translator is
specified and page numbers are given. For more
contemporary sources the short form of the ref-
erence is followed by volume and page numbers
as necessary; e.g., EI [2], 4:387 refers to page
387 in Volume 4 of the second edition of the
Encyclopaedia of Islam.
  
                                Historical Overview
1.    A History of Celestial Globes in the Greco-Roman and Islamic Worlds
From Zeus let us begin; him do we mortals never leave unnamed . . . for himself it was who set
the signs in heaven, and marked out the constellations, and for the year devised what stars
chiefly should give to men right signs of the seasons, to the end that all things might grow
unfailingly.
Aratus
Phaenomena

  Patterns of stars and their collective movement
across the sky have been of interest to people
from the earliest historical periods. The stars
have been commonly, though perhaps not uni-
versally,^ perceived as though attached to the
inside of a hollow sphere enclosing and rotating
about the earth. It is therefore reasonable that
the earliest attempts to represent in a model the
arrangement and movement of the stars were by
means of a celestial globe. The earth, which was
known to be spherical from early classical antiq-
uity, was imagined at the center of the globe,
while the stars were placed on the globe so that
the resulting model presented the stars as seen
by an observer outside the sphere of fixed stars.
Consequently, the relative positions of the stars
on a celestial globe are the reverse, east to west,
of their appearance when viewed from the sur-
face of the earth.
  In antiquity, the idea of constructing a physical
model to represent certain celestial phenomena
appears to have been peculiarly Greek, as was
the attempt to represent the known world on a
terrestrial globe, if we can accept as authentic
the few references to terrestrial globes.^ Tradi-
tion has it that Anaximander of Miletus (610-
547 Bc) made a map of the earth and sea and
made a globe {(T<patpa), possibly terrestrial.^
  According to Cicero,'' who reported the state-
ments of the Roman astronomer Gaius Sulpicius
Gallus of the second century BC, the first globe
was constructed by Thales of Miletus (sixth cen-
tury BC). The account continues:
Subsequently, Eudoxus of Cnidos [ca. 390-340 BC], Plato's
pupil, as Gallus said, marked on the globe the stars that are

fixed in the sky. Many years after Eudoxus, Aratus adopted
from him the entire detailed arrangement of the globe and
described it in verse, not displaying any knowledge of as-
tronomy but showing considerable poetical skill.^
  Eudoxus was an astronomer who appears to
have had considerable influence on later Greek
astronomers.^ Unfortunately nothing remains to-
day of his writing except some possible frag-
ments. Whether or not Eudoxus actually used a
celestial globe, he did compile a descriptive list
of constellations, with their calendric risings and
settings, which he prepared in two versions titled
'EvoTTpa (mirrors) and ^aivbueva (phenomena).
  The latter version was no doubt the basis for
the extraordinarily influential astronomical
poem Phaenomena (^aLvbueva) by Aratus of Soli
in Cilicia (ca. 315-240 BC).' This poem was un-
dertaken at the request of Antigonus Gonatas
who, as ruler of Macedonia, formed around the
court an active literary circle.^ The Phaenomena
surveys in very general terms with no precise
indications of positions or distances the northern
(including zodiacal) and southern constellations,
the circles of the celestial sphere, and the risings
and settings of the fixed stars, followed by a long
section, subtitled dioarjiilaL (omens from the sky),
on indications of weather phenomena. Approxi-
mately 48 distinct asterisms are described, in-
cluding the Pleiades, which are considered sepa-
rately from Taurus because of their importance,
and a constellation called uiScop (water), which may
be the group later considered by Ptolemy to be
the external stars of the Southern Fish.^ In ad-
dition to the Milky Way {yaXa) the circles of the
celestial sphere consist of the northern and south-
  
SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

ern tropic circles (the Tropic of Cancer and
Tropic of Capricorn), the celestial equator {la-
rjixepLvos KVKXOS, or equinoctial), and the ecliptic
{KVKXOS Xo^bs or oblique circle sometimes called
^wldios KVKXOS or zodiacal circle). Sometimes Ar-
atus speaks of the circle of the zodiac as being
only the ecliptic circle, and at other times as a
zone extending about 6° to either side. Certain
asterisms are described as being always visible,
lying within the circle of perpetual visibility {S
ael (pavepos KVKXOS), which indicates a certain geo-
graphical latitude.
  The most frequently suggested latitude is that
of about 37° or 38°, approximately that of Ath-
ens;'*^ however, R. Boker has prepared a detailed
technical analysis of the celestial globe that Ara-
tus (and Eudoxus before him) may have used
while composing the Phaenomena. He concluded
that the celestial globe had the human constella-
tion figures portrayed from the back (as if facing
in toward the globe) and was designed for a
particular geographical latitude, which he esti-
mated to be between 32°30' (that of the ruins
of Babylon south of Baghdad) and 33°40' (that
of Tyrus in Phoenicia is 33° 18' and Sidon
33°34'), with the star positions relative to the
equinoxes corresponding to the beginning of the
tenth century BC, plus or minus 30 to 40 years."
  This significant example of didactic Hellenistic
poetry consisting of 1154 hexameters'^ achieved
great popularity, prompting many commentaries
(the names of 27 are known)'^ and later Latin
translations. These include one by Cicero (frag-
ments of this youthful work still survive); another
from Germanicus Caesar (born 15 BC), who was
adopted by his uncle Tiberius and stood in direct
line of succession; and another by Avienus, pro-
consul of Africa in AD 366. Aratea is the name
given to the Latin versions of Aratus's Phaeno-
mena, and they cover a period of some four
hundred years.'^ In the ninth century the poem
was translated into Arabic.''"
  While it is quite likely that even before the
time of Eudoxus other early poems had been
written on constellations, such as that by Cleos-
tratus of Tenedos,'*" none were as influential as
Aratus's versification of Eudoxus's Phaenomena.

An extant treatise on the mythologies associated
with the constellations entitled KaracrTepiofioL is
questionably attributed to the geographer, math-
ematician, and literary critic Eratosthenes of Gy-
rene (ca. 275-194 BC).'' Though he is known to
have written a treatise by this title, the extant
prose fragment is probably spurious. It describes
42 constellations with the stars numbered in
each, but with no stellar coordinates. In the later
Roman world this work was an important source
concerning the various myths related in antiquity
to the constellation forms.'*
  It is highly unlikely that either Aratus or his
predecessor Eudoxus made any methodical astro-
nomical observations. The poem by Aratus seems
to be describing a celestial globe, though one is
never specifically mentioned. If we can accept
the testimony offered by Cicero and his sources
and that of the seventh-century Byzantine writer
Leontius, who speaks of the greater simplicity of
Aratean globes {aparela aipalpa),^^ celestial
globes were employed by Aratus and possibly
Eudoxus; the precise nature of these globes is
not known today. There is no evidence that the
globes were mounted in meridian and horizon
rings as was to be the classical form throughout
the Christian and Islamic eras. The globe was
probably a sphere representing a number of
prominent stars and 42-48 constellation outlines
drawn around groups of the stars. The globe
would have had great circles representing the
ecliptic and the celestial equator, with two circles
parallel to the equator indicating the tropic cir-
cles and two additional equatorial parallel circles
indicating, the area of the heavens that would
always be visible (or invisible) at whatever loca-
tion the maker of the globe lived. There is no
mention of colures or meridians on these early
globes, nor any indication that they were to be
rotated around the equatorial poles within a ho-
rizon ring to illustrate the movement of the
heavens, though the later writers Manilius of the
first century AD and Leontius of the seventh
century place similar globes within a horizon
ring.
  While the Phaenomena is a smoothly written,
easily readable poetic introduction to the con-
  
NUMBER 46

Stellations, it is, as Cicero implies, astronomically
deficient. Criticisms were made of Aratus's (as
well as Eudoxus's) positioning and descriptions
of the constellations and stars by Hipparchus in
his Commentary on the Phaenomena of Eudoxus and
Aratus {'E^'q-y^qais TCCV 'Aparou Kal ^vdb^ov (paivo-
n'evoov).^^ This commentary is the only preserved
work by this great early astronomer, who was
born at Nicaea in Bithynia about 190 BC.^' It has
been frequently stated that Hipparchus's catalog
of stars (which is not extant as an independent
work, but is contained in part in his commentary
on Aratus) was virtually repeated by Ptolemy
three centuries later in his Almagest, with appro-
priate changes in longitude. It has been conclu-
sively shown,^^ however, that the Hipparchan
catalog must have been quite different. In the
Commentary on Aratus Hipparchus is concerned
with describing only those stars, based on the
latitude of Rhodes, that rise and set, thus ignor-
ing the northernmost constellations, which are
always visible. He uses no single system of coor-
dinates and most significantly no ecliptic coordi-
nates, and he makes no attempt to classify the
stars by magnitude. For these and other reasons,
although Hipparchus was apparently carrying
out observations and was concerned with accu-
rate descriptions and positions, his catalog of
stars was certainly not repeated by Ptolemy,
whose later catalog was fundamentally different.
The number of stars, as well as the number of
distinct constellations, in Hipparchus's catalog is
ambiguous, but it seems there were between 45
and 47 constellations with no more than 850
stars.
  Hipparchus can also be credited with the dis-
covery of the precession of the equinoxes, the
slow movement of the equinoxes against the
background of the fixed stars, which he assumed
was at least 1 ° per hundred years.^'* It has long
been assumed^^ that early astronomers, even be-
fore Hipparchus, commonly used celestial globes.
Although there is no concrete evidence, such
usage must be inferred from statements such as
those found in Aratus's poem. It is probable that
the type of celestial globe used by Hipparchus,^^
if indeed he used one, was a simple sphere not

mounted in rings, designed only to display the
relative positions of the stars and to indicate areas
never visible and always visible at his location.
  There seems also to have been an interest in
terrestrial globes in the second century BC, if we
are to believe the statement of the geographer
Strabo, writing a century later, who says^' a
terrestrial globe was constructed by the Homeric
scholar Crates of Mallos^* about 170 BC. Strabo
himself goes on to say that a terrestrial globe
should be no less than 10 feet {deKa irodicv) in
diameter and that if the maker is unable to
construct one of such a size he should instead
draw a map on a plane surface seven feet across.
Elsewhere in his Geography Strabo says that the
"earth as a whole is sphere-shaped {(TipaipoeLdijs),
but not sphere-shaped as though from a turning-
lathe {eK Topvov) but rather has some irregulari-
ties."^^ From this we may assume that it was
customary to smooth the surface of wooden
shapes by turning on a lathe. Such shapes prob-
ably included spheres, some of which may have
been intended for use as globes.
  In Cicero's tract De republica,^^ written in 51
BC, there are mentioned two celestial globes
made in the third century BC by the great math-
ematician of Syracuse, Archimedes. Both globes
had been taken from the city of Syracuse when
Marcus Claudius Marcellus captured it in 212 BC
during the Second Punic War. One globe had
been placed in the Temple of Virtus in Rome
and was more widely known than the other globe
because it was more attractive. This particular
globe was said to be solid {sphaera solida), having
no hollow spaces, and according to the opinion
of the astronomer Gaius Sulpicius Gallus, was of
the early type of celestial globe, having the con-
stellations and fixed stars on it. The second ce-
lestial globe made by Archimedes^' had been
taken home by the conqueror of Syracuse as his
only share of the booty and was still in the family
home when Gallus was consul (166 BC). This type
of globe the astronomer Gallus called a newer
class of globe:
Gallus declared that the globe at Marcellus' house, which
showed the motions of sun and moon and of those five
wandering stars or planets as they are called [Mercury,

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

Venus, Mars, Jupiter, Saturn] could not be constructed in
solid form. All the more remarkable, therefore, was Ar-
chimedes' discovery, since he had devised a method of
construction whereby, extremely different though the move-
ments of the planets are, the mere turning of the globe
would keep them all in their unequal and different orbits.
When Gallus rotated the globe, the moon really followed
the sun on the bronze globe by the same number of revolu-
tions as are the days it lags behind in the sky. Thus it
happened that on the globe occurred a solar eclipse just like
the real eclipse; and also that the moon passed into the tract
of space covered by the earth's shadow when the sun [and
the moon were on opposite sides of the earth]. (Text breaks
off suddenly.)"*^
  It is possible that what is being described here
was a celestial globe placed at the center of a
demonstrational armillary sphere, with rings rep-
resenting the courses of the five planets, moon,
and sun. It cannot be a celestial globe mounted
in the two rings (a practice that seems to have
been customary by the first century AD), for this
globe of Archimedes had rings representing the
movements of celestial bodies. It has also been
interpreted as an early form of orrery, though
why it would have been considered a variant of
a celestial globe in that case is not clear. The
description contrasts markedly with the solid
sphere (which has no hollow spaces or rings, and
was probably of the type used by Aratus and
possibly Hipparchus).
  A similar mechanical device for demonstrating
the movement of the planets, sun, and moon
against the background of the fixed stars was said
by Cicero^^ to have been constructed by the
philosopher Posidonius, with whom Cicero stud-
ied in Rhodes. Other references^'' to such con-
trivances can be found in the literature, including
a possible one in Plato's Timaeus,^^ written in the
fourth century BC, and a more detailed one in
the major physiological treatise written in the
second-century AD by the physician Galen.^^
These and other mentions of such demonstra-
tional devices indicate that they were not uncom-
mon in learned circles in Hellenistic and Roman
times. Unfortunately, specific details of the ar-
rangement and its precise relation to a celestial
globe are unknown. It has been conjectured that
even the simplest possible device that could reg-
ulate the movements of the planets, sun, and

moon would require a system of gears meshing
in parallel planes.^' Several means have been
suggested that Archimedes and others may have
employed to turn the device, including a hy-
draulic mechanism^* or a worm-wheel construc-
tion.^^ In any case, no artifact of such a device
remains, though the tradition of a self-moving
celestial sphere (without the planets, however) is
discernable in a treatise written in Persia in the
twelfth century, which describes a celestial globe
half-sunk in a box; the globe is rotated once a
day by an elaborate system of pulleys driven by
a float on a sinking reservoir of sand.'*"
  Writing in the first century BC, Geminus of
Rhodes,^' also a pupil of Posidonius, in his Intro-
duction to the Phaenomena (Eio-a7a)7^ els ra (paiv-
bfieva) mentioned two types of celestial globes, a
"solid" variety {arepea aipalpa) and a "ringed"
style {al KpiKooTal aipatpai).^^ The latter type
must refer to a demonstrational armillary sphere
built about a celestial globe, but without a mech-
anism for indicating the movements of the moon,
sun, and planets as designed by Archimedes.
According to Geminus, the "ringed globe" was
constructed with the "arctic" circle {OLPKTLK'OS)
36° from the north (celestial) pole;"*^ the "arctic"
circle was also 30° from the summer tropic circle
{OepLvos rpo-KiKos), while the latter was 24° from
the equator {larjixepivbs), which in turn was 24°
from the winter tropic circle {xn-p.epLvos rpoir-
i/cos).^"* The latter circle was then 30° from the
"antarctic" circle {avrapKriKos), which was 36°
from the south pole {vbros TTOXOS).
  Geminus goes on to say that both the "solid
globe" and the "ringed globe" were constructed
for only one geographical latitude—that is, 36°
north, which is the latitude of Rhodes. The "arc-
tic" and "antarctic" circles were defined as delim-
iting, respectively, the area of the heavens that
was always visible and the area always invisible.
Since the position of these two celestial circles
would depend upon the location of the observer,
the globe was valid for only one geographical
latitude (in this case 36°), and a horizon ring
surrounding a meridian ring that allows adjust-
ment to different terrestrial latitudes was proba-
bly not employed. Indeed such circles are either
  
NUMBER 46

redundant or incorrect if a meridian-horizon
ring assembly is used, depending on the setting
of the rings. Presumably the "ringed" celestial
globe as described by Geminus differed from the
solid form by having the arctic and antarctic
circles, tropics, and equator (and possible colures)
indicated by rings surrounding the globe, on
which were represented the ecliptic and some
stars and constellations; the solid globe might
have had the circles drawn directly on the surface
of the globe itself.
   The five parallel rings of a "ringed" globe were
possibly held in place by one or two rings placed

at right angles to them and attached to the equa-
torial poles of the globe (see Figure 1). If a third
ring were then attached that touched the arctic
and antarctic circles at one of the points where
the circles intersect the outside perpendicular
ring, this third ring would of course be an indi-
cator of the horizon. There is no mention in the
literature that this was ever done. It is also pos-
sible that the ecliptic was indicated by a ring (as
well as on the sphere itself), in which case the
two perpendicular rings to which the parallel
ones were attached could serve as colures.
We have no information as to how the posi-


FiGURE I.—Conjectured design of "ringed" celestial globe described by Geminus of Rhodes in
the first century BC.

8

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



tions of the stars on a globe of this or earlier
periods were determined. The use of coordinates
seems highly unlikely. The constellations along
the ecliptic were probably drawn first with others
placed proportionately about, and then the stars
marked within and outside the constellation out-
lines as was called for by the treatises discussing
the constellations. From his inaccuracies and
vagueness, one can conclude that whatever
globes Aratus used were doubtless the results of
such imprecise procedures.
  The concepts of the "arctic" and "antarctic"
circles being the "greatest always visible" circle
and the "greatest always invisible" circle were no
longer required on celestial globes shortly after
the beginning of our era, for an horizon ring,
enclosing an adjustable meridian ring, would in-
dicate the always visible and always invisible areas
of the firmament for any latitude desired, al-
though the "greatest always invisible" circle re-
mained important in the design of astrolabes.''^
Consequently on all Islamicate celestial globes
there are no "arctic" and "antarctic" circles in the
sense of circles marking areas of constant visibil-
ity or invisibility, but rather polar circles having
the celestial poles at the center and the ecliptic
pole on the circumference, thus having a radius
equal to the obliquity of the ecliptic (see Figure
30). On Byzantine attempts to reconstruct the
pre-Ptolemaic or Aratean celestial globes (such
as that of Leontius), the "greatest always visible
(invisible)" circle reappears. The very few pre-
served representations of Hellenistic and Roman
celestial globes all represent the pre-Ptolemaic
tradition by having such circles indicated on
them.
  There is no indication in these early descrip-
tions of celestial globes that any of them were
placed in a ring indicating the horizon line so as
to demonstrate the movement and rising and
setting of stars for the given geographical lati-
tude. It appears that all celestial globes prior to
our era were primarily designed to illustrate the
relative positions of the major stars and constel-
lations and to indicate those that were always
visible at a certain latitude, as well as to show in
general terms the course of the sun through the
zodiac.
  
Ptolemy, the great astronomer of Alexandria
(fl. AD 127-148), made two important contribu-
tions to the history of celestial globes: his star
catalog, which formed the basis for all star cata-
logs used by globe makers in the Islamic world,
and his detailed description of a particular type
of celestial globe. In books VII and VIII of the
Almagest {ixeylarrj avvra^ts which, through the
Arabic, became in Latin Almagestumf^ Ptolemy
presents a catalog of 1025 stars.'" Each star is
assigned a magnitude of one to six (some stars
being described as greater or lesser than a given
magnitude) and each star s relation to one of 48
constellation outlines is described by Ptolemy,
who says he has slightly altered the outlines from
those of his predecessors, just as they had
changed the figures passed down by their pred-
ecessors. Five stars are called nebulous
{veipeXoeLbT]s) and nine faint {ap.avpbs), rather than
being assigned magnitudes; each star is given a
specific position in terms of the ecliptic coordi-
nates of longitude and latitude.''* The longitudes
were reckoned along the ecliptic from the posi-
tion of the vernal equinox at the beginning of
the reign of Antoninus Pius (AD 138).''^
  In book VIII, chapter 3 of the Almagest,^^
Ptolemy gives explicit instructions for the design
of a celestial globe that will not become outdated
by the precession of the equinoxes. Ptolemy be-
gins by saying that the maker should take a solid
sphere {arepea aipalpa); no instructions are given
as to how to construct the sphere or what mate-
rial to use. He should then make the sphere a
dark color {xp&tioL ^advrepov) resembling the
night sky. After selecting two antipodal points
on the sphere, a great circle is to be made that is
equally distant from both points; this circle is to
represent the ecliptic. A second great circle is
then td be passed through the two points; it will
then be perpendicular to the first circle and pass
through the two ecliptic poles. Beginning at one
of the two points of intersection of these circles,
the ecliptic is to be divided into 360 parts or
degrees, and then into as many subdivisions of
degrees as desired (such a comment presupposes
a fairly large globe), and then the degrees are to
be numbered. Next, two rings {KVKXOL) of strong
and seasoned wood {e^SjXrjS evrbvov Kal reTap.evr\s)
  
NUMBER 46

are to be made, one whose inner edge can just
pass over the surface of the sphere, while the
second ring has to be large enough to pass over
the first ring. On one of the flat faces of each
ring a semi-circular arc is drawn dividing the
width of the face into two equal parts.^' Then
the semi-circular arc on each ring is graduated
into 180° and labeled (beginning with 90° at
each end of the arc and decreasing to zero at the
center of the half-circle). The smaller of the two
rings is pierced through at the two opposite

points where the graduations begin and attached
by two pins to the ecliptic poles of the globe (see
Figure 2).
  With the ring in place and able to rotate about
the ecliptic poles, the stars can then be placed on
the globe in their proper positions. To avoid the
problem of the slow movement of the precession
of the equinoxes outdating the globe, a constant
and invariable point of reference must be taken,
for which purpose Ptolemy selects the star Sirius
("in the mouth of the Great Dog," a Canis Ma-
  
NORTH  ECLIPTIC
POLE


NORTH  CELESTIAL  POLE


FIGURE 2.—Ptolemaic precession celestial globe.

10

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



joris). For greater accuracy in a later work, the
Handy Tables, Ptolemy selects Regulus {a Leonis),
which is closer to the ecliptic.^^ Sirius is then
placed on the circle at right angles to the ecliptic,
thereby assigning to Sirius a celestial longitude
of 0°. Then the latitudinal position of Sirius on
this great circle is found by lining up the gradu-
ated ring with the great circle and marking the
position on the surface of the globe next to the
graduation on the ring corresponding to the
latitude for Sirius in the star catalog. Yellow or
any other color that would contrast well with the
dark background is suggested for indicating the
stars, with a mark whose size indicates the mag-
nitude (brightness) of the star. All the other stars
in the star catalog are then to be placed on the
globe by comparing their longitudes as given in
the catalog with the longitude of Sirius given in
the catalog, moving the graduated ring to the
point on the ecliptic corresponding to this differ-
ence and then marking off the latitude either
north or south as required. After all the stars are
placed, the constellation outlines should be
drawn in a color different from that of the
ground, with simple, unobtrusive lines. Finally,
the Milky Way is to be indicated on the surface
of the sphere.
  The larger ring is taken as a meridian ring and
two pins are passed through holes bored at either
end of its graduated half-circle. These pins are
attached to the smaller ring so that the topmost
(northern) pin is inserted at the marking of 66 ° 9'
on the upper half of the small ring, thereby
making the angle between the two poles corre-
spond to the obliquity of the ecliptic as given by
Ptolemy (23°51'). Having used the smaller ring
for the purpose of determining the star positions,
it would seem that Ptolemy intended for it then
by some means not specified to be locked into a
fixed position relative to the globe so that it
would form the solstitial colure passing through
the solstices, the equatorial poles, and the ecliptic
poles (to which it is directly attached). This fixed
position would then, according to Ptolemy, be
valid for approximately 100 years, after which it
would need to be reset.
This inner ring is to be set at that point on the

globe's ecliptic that is as many degrees removed
from the circle passing through Sirius as the
latter circle is from the summer solstice for a
particular date. Ptolemy gives as an example
12°20' at the beginning of the reign of Antoni-
nus Pius. The outer (meridian) ring is then placed
within a horizon ring and can be adjusted within
that ring for any given geographical latitude; the
globe with attached inner ring appropriately set
could be rotated east to west to imitate the move-
ments of the heavens. The stars along the equa-
tor could be determined, according to Ptolemy,
by observing which stars on the globe pass under
the zero-point of the graduated half of the me-
ridian (outer) ring; in a similar way the regions
defined by the tropic circles can be seen under
the points 23°51' north and south of the zero-
point of the meridian ring. The equatorial co-
ordinate (declination) of a star, Ptolemy adds,
can be read by rotating the globe so the star is
alongside the graduated side of the meridian
ring.
  Such a description indicates that it was custom-
ary by the time of Ptolemy to place a meridian
ring about a globe, which in turn was set into a
horizon ring (see Chapter 2 of this study for
further details on the use of meridian and hori-
zon rings). Ptolemy does not provide any infor-
mation on the processes involved in actually man-
ufacturing such a globe, nor does he indicate
what mechanism is to be used to allow the inner
ring to rotate freely enough about the sphere so
that it is useful for determining the positions of
the stars, and yet later be rigidly fixed so as to
serve as solstitial colure.
  Ptolemy's design^^ for a celestial globe is
unique in that the sphere has stars and the ecliptic
but not the celestial equator and parallel circles;
furthermore, there is an arrangement of rings
that can be adjusted to any time period as well
as any geographical latitude. As will be seen, such
a design for a precession globe is not known to
have been followed in the Islamic world even
though the Almagest circulated widely and was
for many centuries the fundamental astronomi-
cal treatise in Islamdom as it was in the West.
Unfortunately, there are few surviving Greco-

NUMBER 46

11



Roman celestial globes with which we might com-
pare the later Islamicate ones. It is probable that
most celestial globes produced in antiquity were
made of wood'''* and thus have not survived the
deterioration of centuries. All preserved globes,
or fragments of globes, from Hellenistic and
Roman times are non-functional, inaccurate ce-
lestial globes clearly made for decorative pur-
poses and not for the use of an astronomer. All
reflect the pre-Ptolemaic (or Aratean) design of
a globe made for only one geographical latitude
(bearing circles indicating the always visible and
always invisible areas of the sky) and not placed
in a meridian and horizon ring.^^ No remains are
known of any armillary sphere or device for
indicating the movements of the sun, moon, and
planets,^^ nor of any attempt at a precession
globe of the type described by Ptolemy. The
Islamicate globes clearly represent the tradition
of globe making current by the time of Ptolemy,
which placed the sphere within a meridian and
horizon ring.
  The most famous example from the Greco-
Roman world is the marble Farnese globe held
by a kneeling figure of Atlas.^' The globe itself
is a Roman copy of a Greek original, while the
figure of Atlas is a Renaissance addition. The
marble sphere, 650 mm in diameter, shows 42
constellations (five additional ones, including
Ursa Major and Ursa Minor, have been obliter-
ated) carved in relief and viewed from behind
(i.e., the human figures face into the globe). The
globe bears five parallel circles: the celestial equa-
tor, the northern and southern tropic circles, and
the two circles marking the always visible and
always invisible areas. The ecliptic consists of
three parallel circles, two about 6° to either side
of the great middle circle. The two colures are
also indicated. In these particulars the design
follows Aratus's description. No star positions
seem to be marked, though it is possible that they
were painted on and have worn off with time.
The position of the constellations with respect to
the equinoxes suggests a date from the second
or third century BC (roughly contemporary with
the time of Hipparchus), while the execution of
the piece seems to be no earlier than the first

century AD, for which reason it is thought to be
a copy. This globe, which is not a scientific in-
strument at all but a monumental decoration,
was transferred to the Museo Nazionale, Naples,
from the Palazzo Farnese in Rome, hence the
common name Atlante Farnese.
  Other artifacts from late antiquity, though not
strictly celestial globes, are closely related in de-
sign if not in function. These remains include
vessels for holding liquids, and zodiacal or as-
trological globes. Two hollow globes that served
as vessels are known, one seen in AD 1888 outside
a school building at Larissa,^^ and the other a
fragment of a blue marble Roman vessel^^ with
stars placed arbitrarily and with constellation fig-
ures chiseled out and originally inlaid with an-
other material that has since fallen out. This
fragment is especially interesting in that the
Milky Way is indicated on it, but the placement
of the constellations does not follow completely
either Aratus nor Ptolemy. Four globes having
only the 12 zodiacal figures decorating them are
extant, including a marble one at Arolsen (160
mm in diameter),^^ which had an eagle perched
on top, and a marble one (600 mm in diameter)
at the Vatican,^' which in addition has a few
decorative stars dispersed over it.^^ The marble
"astrological" globe from the third century AD
found at the Dionysius theater in Athens is not a
celestial globe at all, but bears various astrologi-
cal and divinatory symbols.*'^
  Celestial globes of the Aratean variety can be
seen as a rather popular motif in frescos, statues,
coins, and other works of art from the Greco-
Roman world, in many cases intended probably
as an allegorical representation of world or uni-
versal power. Such pictorial representations of
pre-Ptolemaic celestial globes are well surveyed
and illustrated by Thiele, by Schlachter, and by
Gundel and Boker.^'' One gem of lapus lazuli is
especially interesting in that it shows an astron-
omer sitting before a globe using a pair of draw-
ing compasses to measure distances on the
globe.*'^ Though the gem carver has been precise
about depicting the set of compasses, the circles
on the globe make no sense; the globe is not set
in horizon or meridian rings, indicating it is
  
12

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



supposed to be pre-Ptolemaic, and has no stars,
which is quite common in these early artistic
representations of celestial globes.
  In the seventh century AD the Byzantine writer
Leontius composed a treatise entitled On the Con-
struction of an Aratean Globe {Hepl KaTaaKevr]S
^Aparelas aipalpas) that is partially preserved
today.^^ It is apparent that in his time globes
were set within a meridian ring, which in turn
was set within a horizon ring, for what he de-
scribes is a globe based on the design of Aratus
but placed within a set of rings that were proba-
bly unknown to Aratus. Globes in Leontius's day
were apparently not precession globes as de-
scribed by Ptolemy, but bore on their surface the
celestial equator as well as the ecliptic and stars—
that is, the globes were basically of a type similar
to the design evident in Islamicate celestial
globes. Leontius constructed a globe for a man
named Elpidas; its constellations and circles cor-
responded to those in the poem of Aratus. Leon-
tius later set down a description of the globe for
his friend Theodorus. He states that most of the
globes of his day that he had observed agreed
with neither Aratus nor Ptolemy. "For as you
know all the spheres now produced agree reason-
ably well in most ways neither with Ptolemy nor
with Aratus."^' This statement may refer to there
being no precession globes nor pre-Ptolemaic
globes (those with star positions not based on
coordinates) in his day, or perhaps refers to the
inaccuracy of the constellation and star positions,
which did not correspond to those given by either
writer.
  In speaking of the celestial globe described by
Aratus, Leontius says:
One must understand that the statements concerning the
stars made by Aratus are not entirely correct, as can be seen
from the expositions by Hipparchus and Ptolemy. The rea-
son is that, first of all, Aratus for the most part followed the
statements of Eudoxus, which are not very correct. Secondly,
he delineated these things, not aiming at precise accuracy,
as Sporos the conmientator*'" says, but rather at usefulness
for the navigators {vavTiWofitvoL), and quite reasonably he
treated them in a very general way. For indeed they navigate
{OL irXwi^'otiivoL) not by means of ingenious mechanical de-
vices {bia nrjxcuviKwi/ bpyavwv) and exact precision, but by
means of unaided eyesight and observing in general terms
the arrangement of the stars. Consequently the fabricated

sphere is in no way useful for [determining] absolute fact,
but rather very serviceable for understanding Aratean ideas.
It is easy to learn at once the Aratean ideas {ra 'Apartia)
which are not correct as well as to understand the truth of
some of the statements.**^
  Leontius then reminds the reader that Aratus
described the constellations not by coordinates,
but by their spatial relation to one another, by
their positions relative to the tropics, equator,
and ecliptic, and by their rising or setting when
various zodiacal signs rise or set. As an example
of this he takes the constellation of Ophiuchus,
the Serpent Charmer (see Figure 59). Aratus,
according to Leontius, describes it in three dif-
ferent places—first, as having his head toward
the head of "the kneeling man" (Hercules) with
his feet on the chest and eye of the Scorpion,
that he holds the serpent by the middle with the
smaller part in the right hand and the larger part
in the left hand (hence the figure is described as
if facing into the globe), that the jaws of the
serpent are close to the Crown, and that it has
some bright stars at the shoulders and less bright
ones at the hands; the Claws (Libra) are below
the twists of the serpent. In a second place Aratus
says it is cut at the shoulders by the winter tropic
and at the knees by the equator. In yet a third
place he states that at the rising of Cancer, Ophi-
uchus is set from the knees to the shoulders while
dragging off the Serpent up to the neck and that
when Scorpio is rising the head of the Serpent
and Ophiuchus rise together, along with the
hand of Ophiuchus and the first curve of the
serpent; at the rising of Sagittarius the twists of
the serpent rise with the body of Ophiuchus.
After gathering similar information for each of
the constellations from the poem by Aratus, the
maker should note it down and set it aside for a
later stage in the globe making.
  Leontius then notes that Aratus described six
circles that will have to be placed on the globe:
the zodiac (o ^(joibcaKbs, or ecliptic) and five par-
allel circles which consist of the celestial equator
(o lar^jxepivbs), two tropic circles {depLvbs/xitp,e-
pivos rpo-KLKos), each 24° either side of the equa-
tor, and two circumpolar circles each 41 ° from
the celestial poles. The latter two circles are the
  
NUMBER 46

13



"largest of the always visible" and "largest of the
always invisible," which he notes are called the
"arctic" and "antarctic" circles {apKTiKos/
avrapKTLKOs), denoting the areas of the sky that
are always visible and never visible, respectively.
Since these circumpolar circles are 41 ° from the
poles, the globe Leontius is describing corre-
sponded to the geographical latitude of Constan-
tinople, which is probably where Leontius lived.
In contrast to Leontius's interpretation, modern
estimates of the area of constant visibility de-
scribed by Aratus correspond to a terrestrial
latitude between approximately 33° and 38°,
that is, from Tyrus in Phoenicia to Athens. Leon-
tius states that although Aratus does not specifi-
cally name these two circles, he does in effect
describe them by delineating the portions of the
constellations that are always visible, and which
Leontius then concludes correspond to a 41°
circumpolar circle. Leontius felt that the Milky
Way {f] TOV yaXaKTOs ^oivq), on the other hand,
though described by Aratus, ought not to be
placed on a globe, for it is "neither uniform nor
well-placed, but very varied in size, color, extent,
and location."^^
  Proceeding to the actual execution of the
globe, Leontius says to stain (? xP^(^<^vTes) the
sphere, if it is of wood {^vXlvr}), and smooth over
it with plaster {yv\pos), or wax {Krjpbs), thus mend-
ing the cracks in it if there should be any, and
when quite dry to paint it {eiraXelxl/avTes) a deep
color such as that called Xa^ovplos (? azure) and
set it aside till dry. He does not give any specific
indication as to how the sphere is to be made nor
why it would have cracks that need filling. He
does not indicate whether it would be a solid or
hollow sphere, whether in one or more pieces,
nor whether it would be turned on a lathe for
final smoothing and shaping. He does seem to
imply that the sphere could be made of some-
thing other than wood.
  After the sphere has been thus fashioned,
there is to be set on supports a ring that will
serve as the horizon {opC^cjov), and which will
enclose a second ring equal in size (and set at
right angles to it in notches, cut in the first ring)
which will serve as the meridian {nearjfx^ptvbs)
(see Figure 3). The inside radius of both rings is

to be equally distant from the convex surface of
the sphere and each ring is to cut the sphere into
two halves. Furthermore, one half ("the semi-
circle above ground, as is customary") of the
meridian ring should be divided by straight lines
into 180°. Though he does not specify it, the
meridian ring is to be attached to the sphere at
two opposite points between which runs the 180°
graduated half-circle.
  Once the sphere is mounted inside the merid-
ian ring, which in turn is nested in the horizon
ring held horizontal by supports, then alongside
the 41° mark on the graduated half-circle of the
meridian ring, a sharp needle {^eXbvri) is held
perpendicular to the surface of the sphere and is
made to penetrate into the colors, that is, the
surface paint and plaster. Then as the sphere is
turned one full revolution, the sharp point will
trace on the sphere a circle which will be equiv-
alent to the "arctic" circle. At a point 90° distant
from both the north and south poles, another
circle can be traced, which will be the (celestial)
equator, while another circle 41 ° from the south
pole is etched in a similar manner. Then at 24°
from the equator the maker traces by the same
method the summer tropic on the northern hem-
isphere and the winter tropic on the southern
hemisphere.
  Leontius then says to find a point on the sur-
face of the sphere 24° from the north pole (equal
to the distances of the tropics from the celestial
equator and hence the obliquity of the ecliptic).
This point will serve as the pole of the zodiacal
circle. Leontius does not give details for tracing
the great circle of the ecliptic, but one may
assume that the sphere would be realigned and
attached to the meridian ring at the new poles
24° distant from the first set, and the circle
traced alongside the 90° mark of the ring in a
manner similar to the preceding tracings. If the
sphere were left to rotate in the meridian ring at
the celestial equatorial poles, the simple rotation
and use of the graduations on the meridian ring
could not be used to trace the ecliptic as with the
previous circles. Leontius states that the zodiac
will touch the tropics and be divided into two
equal parts by the equator. Once the circle is
drawn, it is to be divided into 12 parts, with the
  
14

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 3.—Basic design of an "Aratean" celestial globe constructed by Leontius in the seventh
century AD.

points of contact with the tropics and the equator
being points of division.
  Each of the six circles is to be distinguished by
a color different from that of the base color—
that is, the incised circles are to be painted over
with a contrasting color. Then the names of the
twelve zodiacal signs are to be painted using a
light color {aeplif} xP^y-otn) that can be easily
removed, beginning with Cancer at the point
where the zodiac touches the summer tropic and
working eastward with Leo, Virgo, and so on.
Although it is not stated in the tract, it is assumed
that at this point the constellations and stars
would be indicated on the globe using the infor-

mation gleaned from the Aratus text.
  That done, if the meridian ring with the en-
closed sphere is placed in the horizon ring so
that, when revolved, the "arctic" and "antarctic"
circles touch the horizon ring on the upper
northern side for the "arctic" and on the lower
southern side for the "antarctic" (set for a geo-
graphical latitude of 41° north), it will be evi-
dent, Leontius says, that the two tropics will be
cut by the faces of the horizon ring in a ratio of
15:9. That is to say, he continues, that since one
day and one night {TO vvxOrffxepav) consist of 24
hours (wpai), the summer tropic will show 15
hours above the horizon on the longest day and
  
NUMBER 46

15



9 below for the shortest night; this will be the
inverse for the winter tropic, which will have 9
above for the shortest day and 15 below for the
longest night. This relation of 15:9, Leontius
notes, is the same as the 5:3 division of the
summer tropic given by Aratus.
  In the last section of the treatise Leontius states
that the sphere should be placed in the stand
"along the pole {rpbiros) of the equator"—that is,
with the globe attached at the celestial poles to
the meridian ring and the ring rotated within the
horizon ring so that the axis of the globe is
horizontal. Then the sphere is turned until the
beginning of the division of the ecliptic belonging
to Cancer appears at the eastern edge of the
horizon ring. The sphere is to be made secure
and immobile "with wax or some such thing" and
a circle inscribed with the sharp needle on the
surface of the sphere alongside the entire circum-
ference of the horizon ring—thus producing on
the globe the solstitial colure, which, however,
remains unnamed. Leontius instructs the maker
to label the semicircle (of the ecliptic) to the east
of this new great circle "East of Cancer" using
small letters, and that to the west, "West of
Cancer."
  It is curious that Leontius would have placed
an Aratean globe, which was constructed for only
one geographical latitude, into a meridian-hori-
zon ring assembly. He must have realized from
Ptolemy's Almagest, cited several times in his
treatise, that the rings allow for adjustment to
different latitudes. It is implied in Leontius's
treatise that the globe is to remain set for only
the one given latitude, in this case that of Con-
stantinople, and not adjusted for other latitudes.
Even held at that one position, however, the
horizon and meridian rings make the arctic-ant-
arctic circles redundant. This treatise represents
an interesting compounding of pre-Ptolemaic
principles with a design that was apparently com-
mon by the time of Ptolemy.''
  A Greek celestial globe was described in the
twelfth century by the astronomer Ibn al-Salah,
who said he saw it, probably in Baghdad, where
he worked for most of his life. This globe, ac-
cording to Ibn al-Salah, had the stars positioned

so that their longitudes were increased 6° over
those in the Almagest of Ptolemy, in which case
(taking a value for precession common in antiq-
uity of 1 ° per 100 years) the globe would be
datable to around AD 738.'^ Unfortunately, Ibn
al-Salah does not provide us with further details
concerning the design of the globe, or whether
it had rings.
  In a ninth-century copy^'^ of the astronomical
poem by the Augustan poet Hyginus, there oc-
curs a schematic drawing of a celestial globe
mounted in a meridian ring by a pin at the
northern equatorial pole. The meridian ring is
supported at the nadir by a small decorated col-
umn and rests in a horizon ring drawn to resem-
ble a Greek building having six Corinthian col-
umns. The globe itself has a wide zodiacal band
with the zodiacal constellations drawn in it (Gem-
ini, Taurus, and Aries showing), with three ad-
ditional northern constellations and three south-
ern drawn on it; no stars are indicated. The
human figures are still drawn facing into the
globe. The illustration is labeled Involutio
spherae, which is the name of a late Greek school
poem appended to the poem by Hyginus.
  By the ninth century AD it is evident that in
the Islamic world'"* all celestial globes are to be
mounted in two graduated rings and made ad-
justable to different geographical latitudes—a
basic design apparently common by the first cen-
tury AD. On all globes described in Islamicate
treatises, as well as all the ones preserved today,
both ecliptic and equator, nearly always gradu-
ated, are indicated directly on the surface of the
globe, so that none are precession globes as de-
scribed by Ptolemy. On all extant Islamicate
globes that have the constellation outlines on
them, the human figures face outward toward
the person using the globe rather than in toward
the globe with their backs to the observer as was
apparently, from our very fragmentary evidence,
commonly done in the Greco-Roman and Byzan-
tine worlds. On every Islamicate globe preserved
today there is a set of six great circles at right
angles to the ecliptic. When and where this first
became customary is unknown. Meridians, with
the occasional exception of the two colures, were
  
16

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



not usually indicated on Islamicate celestial
globes (the exceptions being nineteenth-century
products), just as apparently they generally were
not on Greco-Roman globes.
  The earliest extant Islamicate constellation im-
ages are to be found in a cupola of the palace of
Qusayr ^Amr located in the desert about 50
miles east of the north end of the Dead Sea. The
small palace was built between AD 711 and AD
715 by al-WalTd I to commemorate his victo-
ries,'^ and had a bath consisting of three rooms,
one tunnel-vaulted, one cross-vaulted, and a
third covered by a dome. The dome of this
calidarium was decorated to resemble the vault
of heaven, with the northern and zodiacal con-
stellations plus some southern ones depicted on
it, and the northern celestial pole directly over-
head. Consequently, it displayed a larger portion
of the sky than could be observed at any one
time from one location on the earth.'^
  The design of this fresco presents an interest-
ing blending of late classical and Byzantine tra-
ditions with some early Islamicate influences,"
and is of interest to us as evidence of a transitional
period in celestial globe design. The iconography
of the majority of the constellation figures (many
of which have been badly defaced) appears to be
very closely related to illustrations of constella-
tions found in two late Greek manuscripts, one
of the ninth century and one of the fifteenth
century,'^ as well as to some features of those on
the Farnese globe, indicating a continuous Hel-
lenistic and Byzantine tradition.
  The maker also reflects some Hellenistic influ-
ences in drawing the human figures slightly
turned into the globe (or in this case into the
surface of the dome), and in representing the
ecliptic as a wide band with a center circle and
two parallel circles about 5° or 6° to either side.
Yet three figures (Cepheus, Bootes, and Orion)
are depicted in a manner not seen in Greek
sources, but which is found on some of the later
Islamicate celestial globes.'^ Moreover, there are
a set of six great circles that pass through the
ecliptic poles and divide the ecliptic into twelve
parts (little more than the northern semi-circles
of each great circle is actually indicated on the

dome). These ecliptic latitude circles^^ are not to
be found on any of the few Greco-Roman arti-
facts or drawings of celestial globes or planis-
pheric star maps, and of course are not in man-
uscript illustrations of the individual constella-
tions. They are, however, to be found on all
extant Islamicate globes.
  Of particular interest is the fact that the
painter of this fresco has not designed it so that
you view the constellations as you stand under
the dome in the same way that you would see
them overhead in the sky. Instead he has re-
versed left to right each constellation figure and
the entire order of the constellations—that is, he
has drawn the constellations as you would see
them looking down on a celestial globe rather
than up into the sky. Thus the possibility arises
that he was using an early celestial globe as a
guide to the constellations rather than manu-
script illustrations. Yet the model of a celestial
globe does not completely account for the spac-
ing of the constellation figures. It is also evident
that the artist did not fully understand the tech-
nical significance of certain details, for the circle
of the ecliptic was not made to pass through the
northern solstitial point, possibly to avoid a win-
dow in the dome; some of the constellations are
crowded and poorly positioned.
  In addition to the ecliptic latitude circles inter-
secting at the north ecliptic pole, there are also
a series of concentric circles having the celestial
(equatorial) pole as the center which, in the
dome, is directly overhead. There are six prom-
inent circles painted dark brown. The smallest in
the center is the equatorial polar circle passing
through the ecliptic pole about 23'/2° distant—
that is, what we today call the arctic circle. There
is a circle representing the equator, with one
circle inside it representing the northern tropic
and one representing the southern tropic outside
it, and between the polar circle and the northern
tropic are two nearly equidistant circles. Then
there are three additional concentric circles
evenly spaced between the northern tropic and
the second circle from the center; these circles
are very pale and appear to have been painted
over and were possibly preliminary attempts of
  
NUMBER 46

17



the artist to get the spacing on the dome as he
wanted it. The edge of the domed ceiling is
another of the concentric circles, equivalent to
about 40° south of that representing the equa-
tor. The polar circle is a common feature of
extant Islamicate celestial globes, while there is
no mention of it in Hellenistic, Roman, or Byzan-
tine sources.
  Such concentric circles are to be found in two
ninth- or tenth-century Latin drawings of plan-
ispheric star maps,®' which show the constella-
tions with great distortion in five concentric cir-
cles representing in the center the area bounded
by the equatorial polar circle, then the northern
tropic, the celestial equator, the southern tropic
and the southern polar circle forming the outside
border of the planispheric map. The ecliptic is
represented by a wide circular band apparently
off-center around the ecliptic pole; there is also
a second apparently off-center circle, which is
the Milky Way.
  As has been suggested,"^^ such a stereographic
projection of the heavens to form a planispheric
star map must have been similar to the model
the artist at Qusayr *^Amr employed, for at first
sight these two maps appear to be very much like
the dome. There are, however, differences. The
order of the constellations on the manuscript
planispheres is that which would be seen in the
sky—that is, they are not reversed as would be
seen on a globe or as the painter of the dome
depicted them. Secondly they lack the ecliptic
latitude circles. Furthermore, they depict the
Milky Way, which is not painted on the dome.
  If we assume, however, that what the painter
of the Qusayr 'Amr dome employed as a guide
was not a stereographic projection of the heavens
quite like those in the two extant manuscript
planispheric maps, but rather was a stereographic
projection of a celestial globe, then all the fea-
tures of the dome are quite easily accounted for.
The arrangement of concentric circles about the
overhead celestial pole and the spacing of the
constellation figures would be accounted for by
the method of stereographic projection using
equidistant representation of the circles of dec-
lination; the ecliptic latitude circles are an inte-

gral part of later Islamicate globe design and
were probably common in this period as well.
We have no evidence that the Milky Way was
ever indicated on a globe (except for Ptolemy's
precession globe, which appears to have had little
or no influence on subsequent globe design), so
its omission would be consistent with a globe
having been the basis of the stereographic pro-
jection.
  If we accept such a conjecture, then we can
conclude that in the eighth century in Syria ce-
lestial globes were being drawn with the ecliptic
latitude circles so characteristic of the later ex-
tant globes, and that planispheric star maps
drawn in stereographic projection were, if not
common, at least available. If this were true, then
this dome would provide evidence for the sur-
vival of techniques of stereographic projection in
the provinces of the Roman and Byzantine em-
pires before the earliest extant astrolabes and
before the translation of Greek texts into Arabic.
In this context it is worth noting that the palace
of Qusayr "Amr was built in the same region
where some six to seven decades earlier, the
Syriac scholar Severos Sebokht wrote a Syriac
astronomical text based on Greek sources and a
treatise on the astrolabe.®^
  At the end of the eighth century two Persian
astronomers working in Baghdad were responsi-
ble for introducing into Islamicate astronomical
literature the Sanskrit astronomical Siddhanta
texts which dated from no later than the first
half of the fifth century AD. The two astronomers
were Ya'qub ibn Tariq, who died about AD 796,
and Muhammad ibn IbrahTm al-FazarT (died ca.
AD 801-802), who was assigned by the Caliph al-
Mansur the task of translating the texts into
Arabic. Five Siddhanta texts, which incorporated
many Hellenistic astronomical practices, were
summarized by Varahamihira in the sixth cen-
tury AD. In this extant Sanskrit summary by
Varahamihira there is a brief and rather obscure
mention of a celestial globe, which the older
translation says was to be made simply "of some
material," while the more recent translation says
it is to be "of wood."®'' The globe has the equator
and ecliptic indicated, and apparently the ecliptic
  
1

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



(or possibly the equator) is graduated. No rings
are mentioned, but it is stated that the globe
should be tilted to the north by an amount equal
to the terrestrial latitude. Then a technique is
mentioned that is apparently similar to that dis-
cussed in Chapter 2 for finding how much time
had elapsed on a certain day, given the position
of the sun in ecliptic, the ascendant, and the
latitude of the town. The short reference in the
Sanskrit text is quite obscure, however, and can
be interpreted differently.
  About one hundred years after the building of
Qusayr 'Amr the Arabic translation of Ptolemy's
Almagest was completed in Baghdad by al-Hajjaj
ibn Yusuf ibn Matar, who may well have used an
earlier Syriac version by SarjCin ibn HalTyu al-
RumT. A more authoritative version of this im-
portant treatise was prepared in AD 827-828 by
Ishaq ibn Hunayn ibn Ishaq for the Caliph al-
Ma^miin and corrected by Thabit ibn Qurra.®^
The Arabic version of the Almagest was an im-
portant source for the star positions and constel-
lation shapes for Islamic globe makers, for it
contained tables of latitudes and longitudes for
each individual star and described its position in
relation to 48 constellation outlines. The descrip-
tion of the stars in terms of the constellations was
intended as an aide-memoire to the general loca-
tion of the stars and not a substitute for the
precise charts of coordinates. Some of the extant
Islamicate globes specifically state that they were
based on the coordinates in Ptolemy's Almagest
(Nos. 2 and 3 of the catalog). The celestial globe
described by Ptolemy, however, seems to have
had little influence on later globe design.
  Early in the ninth century the well-known in-
strument maker "AIT ibn Tsa made a large celes-
tial globe.^*' Though he worked in Baghdad and
Damascus, it seems that an even more important
Islamicate center of scientific instrument making
was the city of Harran^' 'yii^g between the north-
ern reaches of the Euphrates and Tigris rivers
southeast of Edessa. The tenth-century author of
the Book of the Constellations of the Fixed Stars,
"Abd al-Rahman al-SufT, speaks of having seen
many celestial globes executed by people from
Harran.®^ Qurra ibn QamTta al-Harram is said®^

to have made a model of the world {sifat al-
dunyd, a terrestrial globe?), which the scholar and
translator Thabit ibn Qurra al-Harram copied.
It was made of DubayqT cloth (named after a
town in Egypt), which was well-known through-
out North Africa and Syria. The cloth was of
unbleached fiber with waxed dyes. The twelfth-
century astronomer Ibn al-Salah who worked in
Baghdad mentions using a celestial globe of Har-
ranian origin.^*'
  In Harran in the latter part of the ninth cen-
tury, the influential astronomer al-BattanT
(known in the Latin world as Albategni or Alba-
tenius)^' wrote a comprehensive astronomical
treatise which included star catalogs and plane-
tary tables. In this extensive tract al-BattanT de-
scribes in detail a celestial globe that is suspended
in five rings.^^ To make this instrument, which
he calls al-baydah (the egg), take "a sphere of
copper {nuhds), adjudged quite round, well-exe-
cuted in every direction, smooth of surface,
turned on the lathe {makhrdtah fi al-shihr) and
of whatever size you wish."^^ He does not supply
any information on how this sphere of copper is
to be constructed, nor if it is solid or hollow.
Once the sphere is obtained, the maker is to
determine two points diametrically opposite on
it, which will serve as poles and inscribe a circle
having these points as poles and cutting the
sphere in half. This circle is then divided into
four equal parts, which are marked by a dot.
Taking one of the dots as the center, a second
great circle is inscribed, which will pass through
the first two poles and cut the first circle into two
opposite halves (this second great circle will be-
come the solstitial colure). Then the maker is to
divide one of the quadrants of the first circle
(which will eventually serve as the celestial equa-
tor) into 90 equal parts.
   Using this graduated quadrant as a measure, a
drawing compass {midwdr) is set to equal the
obliquity of the ecliptic, which al-BattanT specifies
as 23°35'. The maker is then to set one end of
the compass on one of the two poles and inscribe
with the other end a circle around the pole; the
procedure is repeated with the second pole.
(Thus two equatorial polar circles are drawn that
   
NUMBER 46

19



pass through the ecliptic poles; the obliquity of
the ecliptic can, of course, only be approximated
to the nearest degree, since the quadrant of the
equator used as a scale has only 90 divisions.)
These two lesser circles will intersect the second
great circle at four places; two of these points of
intersections diametrically opposite each other
are then taken as a set of poles for a third great
circle, which will be the ecliptic. The circle that
passes through all four poles, al-Battanl notes, is
called the Circle of Cancer and Capricorn (the
solstitial colure, which was the second great circle
to be inscribed on the sphere). The point where
this circle intersects the ecliptic above and to the
north of the equator is labeled the "Point of the
Start of Cancer" and the point below the equator
the "Point of the Start of Capricorn." The two
points where the ecliptic and equator meet are
labeled the "Point of the Start of Aries" and the

"Point of the Start of Libra" (see Figure 4). Then
the maker writes the names of the zodiacal
houses along the ecliptic, allotting three houses
to each quadrant. Each house has six sections of
5 ° each, with each section or interval numbered
with the abjad letters {bi hurHf al-jummal)—that
is, instead of standard numerals, letters of the
alphabet with recognized numerical values were
employed. Al-BattanT notes that the ecliptic is to
be numbered, with 5° through 30° indicated for
each of the twelve houses while the equator is to
have 72 sections of 5°, each section numbered
from 5° through 360°, beginning at the "Point
of the start of Aries."
  Al-BattanT continues by giving instructions for
determining star positions on the globe. The
maker is to take from the graduated scale of the
equator with the drawing compass the amount
equal to the stated latitude of the star. Then with
  

  

n t   o<   t h e
of    Copr'corn


FIGURE 4.—The celestial globe described by al-BattanT in the ninth century AD, showing the
determination of a star position.

20

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



the compass thus set, one end of the compass is
set "on the degree in which the star is" (i.e., its
longitude reading along the ecliptic), and the
other end used to trace an arc in the general area
of the latitude "with an obscure line which will
leave no trace on the sphere." Then the drawing
compass is opened and its two heads set at an
amount equal to 90° of the equator. With it thus
adjusted, one end of the compass is placed at a
point along the ecliptic, which is 90° distant from
the longitude, and the other end used to inscribe
an arc that will necessarily intersect the first arc
at a point that will be the position of the star (see
Figure 4).
  Al-Battani does not cite a specific example.
Figure 4, however, demonstrates his technique
with the star a Herculis, whose coordinates might
be given as latitude 37°30'N and longitude Sag-
ittarius 5°22'. Since the smallest graduations on
the globe are single degrees, the positions can
only be approximated to the nearest degree or
half-degree. First, with the compass set at 37°,
based on the graduations of the ecliptic, one end
would be placed at 5 ° of the House of Sagittarius
and an arc lightly inscribed toward the north.
Then with the compass set at 90°, a second arc
would be inscribed with one end of the compass
set on 5° of the House of Virgo (that is, 90°
distance from the longitude of Sagittarius 5°).
The point of intersection is a Herculis. The man-
ner of marking the stars on the sphere, whether
with incised circles or Xs, or with inlaid silver
points, is not specified by al-BattanT.
  Following the placement of as many stars as
one wishes on the sphere, the maker, according
to al-Battani, should then incise on the globe a
great circle passing between each zodiacal house
and through the ecliptic poles (resulting in a set
of six great circles serving as ecliptic latitude
circles).
  This sphere, according to al-BattanT, is then to
be set within a set of five copper {nuhds) rings
(see Figure 5 for an illustration). One ring serves
as a horizon ring; two rings nested one within
the other together serve as the meridian ring,
the inner ring to which the sphere is attached at
the celestial poles being movable and other ring
stationary. A stationary ring at right angles to

both the meridian and horizon rings marks the
zenith and nadir and east-west points of horizon.
These rings with enclosed sphere rotate within a
larger fifth ring, by which the assemby can be
suspended. This outside ring has a slot cut
through the thickness of one of the upper two
quadrants, through which a pointed gnomon can
be inserted and adjusted to different positions.^''
The details of the design and construction of
these rings described by al-BattanT and the use of
the gnomon will be discussed in Chapter 2.
  With this detailed account of a celestial globe
by al-BattanT we have moved completely away
from the pre-Ptolemaic designs of celestial
globes, which were characterized by the five par-
allel equatorial circles and an emphasis on con-
stellation outlines rather than precision of star
positions. This globe of al-BattanT's has the eclip-
tic latitude circles characteristic of all Islamicate
celestial globes and the star positions determined
by coordinates; in fact, he does not even mention
the depiction of the constellation outlines. The
majority of extant Islamicate celestial globes are
missing the rings, but among those that do pos-
sess them still, all but one have the horizon ring
fixed to a stand and consequently do not have
the outside fifth ring nor the adjustable gnomon.
The one exception is globe No. 5 of the catalog,
made at the end of the thirteenth century. Its
horizon ring, to which the meridian ring and
zenith rings are attached, is suspended rather
than resting on a stand, much as al-BattanT de-
scribed, but without the outside fifth ring. The
zenith ring appears to rotate about the globe,
however, and to have a slit running through the
thickness of the upper quadrant parallel to the
face of the ring, through which perhaps a needle-
like gnomon could have been placed, although
the ring is perhaps a bit closer to the surface of
the sphere than al-BattanT intended his outside
ring to be. Nonetheless these unusual features of
a slit in a rotatable zenith ring and a horizon ring
that was intended to be suspended make this
globe the closest extant artifact to the design
detailed in the ninth century by al-BattanT.
  Toward the end of the ninth century a Chris-
tian from Baalbek in Syria, Qusta ibn Luqa al-
Ba'^labakkT, lived and worked in Baghdad and
  
NUMBER 46

21






FIGURE 5.—The celestial globe of al-BattanT, mounted with five rings and a gnomon. Horizon
ring = I, zenith ring = 2, moveable meridian ring = 3, stationary meridian ring = 4, and
outside ring carrying gnomon = 5.

Armenia.®^ In addition to translations from
Greek and Syriac into Arabic, Qusta ibn Liiqa
composed several medical and astronomical trea-
tises, including one On the Use of the Celestial
Globe {Kitdb ft al-'^amal bi-l-kurah al-nujumiyah).

which is still preserved today.^^ This treatise is of
considerable importance to our knowledge of the
purpose and practical value of such globes to
astronomers (see Chapter 2), and it was trans-
lated into Latin by Stephanus Arnaldus as De

22

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Sphaera solida,^^ into Hebrew by Prophatius Ju-
daeus in the thirteenth century,^® and into Italian
in 1341 by Maestro Bernardo Arabico ouero
Saracino.^^ A Spanish translation was completed
by February of 1259 under the collaboration of
Jehuda ben Moses Cohen and Maestre Johan
Daspa, clerigo del rey. This version can be found
in the translations and compilations from Arabic
astronomical studies known as Los libros del saber
de astronomia prepared for Alfonso el Sabio, who
in 1251 had been crowned Alfonso X of Cas-
tile."^" A treatise on the use of the spherical
astrolabe has been attributed to Qusta ibn Liiqa,
although his authorship is questioned today.'°'
  The first four chapters of the Spanish transla-
tion of the treatise on celestial globes, titled Libro
del alcora, are concerned with the methods of
constructing a globe—a topic Qusta ibn Luqa
had ignored in his original treatise (see Chapter
2). These four chapters were presumably written
by the translators, and a final chapter on some
astrological uses was composed by an unidenti-
fied "Don Xosse al-faquin" (? Don Jose the
learned [faqTh] or Don Jose the fakir). The celes-
tial globe as described by Qusta ibn Liiqa is to
be set inside a meridian ring, which in turn rests
in a horizon ring on a stand {kursi). Both of the
rings are graduated into 360°. The material of
the globe and rings is never specified, though at
one point there is a mention of placing the stone
into the horizon ring. The meridian ring is at-
tached by two pins to the equatorial poles of the
globe. The globe itself has an ecliptic and equator
drawn on it, both graduated into 360°. At right
angles to the ecliptic are six great circles that
divide the ecliptic into the 12 zodiacal houses,
and the names of the zodiacal signs are then
written along the ecliptic, one at each 30° divi-
sion. "Circles of the two poles of the ecliptic" and
"circles of the two poles of the equator" are then
drawn; thus there are four polar circles: two
having the equatorial poles at the center and the
circumference passing through the ecliptic poles,
and two centered at the ecliptic poles and passing
through the equatorial poles. All four have a
radius equal to the obliquity of the ecliptic whose
value is not specified by Qusta. Tropic circles are

not indicated, and constellation outlines are not
drawn on the globe.
  Qusta ibn LQqa specified that only the bright
stars such as those used on astrolabes are to be
indicated on the globe, and these stars are to be
represented by small circles, each of which is to
be labeled with the name of the star it represents.
This description corresponds with a common
type of extant Islamicate celestial globe bearing
only the major stars without constellation out-
lines. Qusta also directs the maker to place the
names of 28 lunar mansions inside small circles
alongside the ecliptic. Lunar mansions are a char-
acteristic of early Arab, pre-Islamic astronomy
and astrology and are of very ancient, possibly
Mesopotamian or Indian, origin (see "Lunar
Mansions" in Chapter 5). The lunar mansions are
occasionally but not commonly indicated on ex-
tant Islamicate globes.
  In the tenth century "Abd al-Rahman al-Siifi
(AD 903-986) was court astronomer to 'Adud al-
Dawla in Isfahan in Persia.'"^ He composed a
most important treatise consisting of a commen-
tary on the star catalog presented by Ptolemy, in
which al-SiifT gives information on pre-Islamic
star names and on some stars not listed by Ptol-
lemy. Accompanying the commentary were illus-
trations of each of the 48 constellations (with
stars indicated and numbered in each) as they
are seen in the sky by an observer on earth {fi
al-samd") and again as seen on a celestial globe
{fi al-kurah) which is to say reversed right to
left, and also a slight revision of Ptolemy's star
catalog in chart form (see Classical Greek and
Pre-Islamic Sources in Chapter 5). This work,
entitled Kitdb suwar al-kawdkib al-thdbitah {Book
of the Constellations of the Fixed Stars), was also
translated into Persian in the thirteenth century
by the astronomer NasTr al-DTn al-TusT and at
least twice again in the seventeenth century.'°^
A considerable number of illustrated copies, both
Arabic and Persian, are preserved today, the
earliest being one copied by his son in AD 1009-
1010; there is also an Arabic copy and a Persian
version, both having the autograph of the fif-
teenth-century astonomer Ulugh Beg.'"4
  This treatise by al-SiifT was no doubt an ex-
tremely important source for the design of con-
  
NUMBER 46

23



Stellation images for globe makers, and the illus-
trations in the manuscript copies reflect the dress
and artistic conventions of different locations in
the Islamic world and time periods, just as do the
depictions of the constellation figures on extant
Islamicate celestial globes. The star catalog pre-
sented by al-Sijf T, moreover, was a direct source
for at least four of the extant globes (Nos. 6, 7,
8, and 62), and was no doubt the source of
coordinates for many other makers who did not
choose to put such an acknowledgement on their
globe.
  Al-SufT speaks of having seen a book on con-
stellations by "^Utarid.'""^ He claims that it con-
tained errors,'*^'' and also attacks the earlier as-
tronomical work of al-BattanT.'°' He notes errors
on the large celestial globe made by "AIT ibn Tsa
a century earlier, and says he has seen many
globes made by makers from Harran. Al-SufT
criticises those celestial globes, which he calls
kurdt musaxvwarah (decorated globes) made by
makers who did not know the true positions or
magnitudes of the stars. He goes on to say, how-
ever,'"** that many globe makers used star cata-
logs, such as the Almagest of Ptolemy, for the
longitudes and latitudes of the stars, changing
the longitudes to correct for the intervening time
interval, but did not realize that when compared
with the actual observations of the stars in the
sky the coordinates were incorrect. By this re-
mark al-SufT seems to be criticizing the reliability
of the available star catalogs.
  According to the scholar al-BTrunT,'°^ al-SufT
told the geometer Abii Sa'^Td Ahmad ibn AbT al-
JalTl al-SijzT (ca. AD 951-1024) that he, al-SQfT,
laid very thin paper on a celestial globe and fitted
it carefully over the surface of the sphere and
then traced on the paper the constellation out-
lines and individual stars as precisely as the trans-
parency of the paper would allow. Such a pro-
cedure was probably used by al-SufT only to
obtain the drawings of the constellations that
accompany his book, and these illustrations of
the constellations and stars give only approxi-
mate relationships and positions; the precise po-
sitions are obviously to be taken from the accom-
panying star catalog, where they are given in
angular measurements. Al-BTrunT, however, goes

on to criticize al-SiifT for thinking that you could
accurately transfer spherical measurements onto
a plane surface by simply tracing the positions on
a piece of curved paper and then straightening
it out. Al-BTrunT then uses this criticism as a
starting point for his own treatise on stereo-
graphic projection, by which one can with math-
ematical methods accurately project spherical co-
ordinates onto a plane, as is done in designing an
astrolabe.
  All such statements by al-Sufi imply that celes-
tial globes were not uncommon in the ninth and
tenth centuries, though none from that period
are known to be extant. It may well be that some
were of a considerably larger size than any from
the Islamic world still preserved today. In the
same treatise on stereographic projection, al-BT-
runT says that celestial globes are preferable to
illustrated texts because the constellations are
positioned around each other and in correct pro-
portion, but "that it does not work well on small
globes, but only on large ones; but these are rare
and costly and too large for carrying and trans-
porting on journeys, so that just as their useful-
ness surpasses [that of the book], the difficulty
involved in the use of these instruments counter-
balances these advantages.""°
  Al-SCifT prepared a treatise on the use of celes-
tial globes'" and is reported to have constructed
a celestial globe himself. A maker of astrolabes
and celestial globes in Fatimid Egypt said that in
the year 435 H/AD 1043 he saw in the library in
Cairo {Khizdnat al-Kutub bi-l-Qdhira) "a celestial
globe {kurah) of silver ifiddah) made by al-Sufi
for the ruler "Adud al-Dawla, and its weight was
3000 dirhams [silver coins weighing about one-
eighth of an ounce] and so it had been sold for
3000 dinars [gold coins].""^
  The renowned late ninth- and early tenth-
century scholar al-BirunT is of interest to the
history of celestial globes not only because of his
comments concerning globes in his treatise on
stereographic projection, but also because he
presented a star catalog that may have influenced
makers of celestial globes. Furthermore, in some
of his writings he not only discussed the constel-
lations in general, but also lunar mansions that
occur on some globes. These treatises are impor-
  
24

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



tant sources of information for the history of star
names and lunar mansions."^ One manuscript of
such a treatise is illustrated with drawings of
constellations having stars placed in them similar
to those illustrating al-Sufi's treatise."''
  No Islamicate globes predating the eleventh
century are known to exist today. The earliest
preserved Islamicate globe (No. 1 of the catalog)
was made in Valencia in Spain in AD 1080 (or
1085) by a leading astrolabe maker IbrahTm ibn
SaTd al-SahiT al-Wazzan with the collaboration
of his son Muhammad. The inscription in Magh-
ribT (western) Kufic script informs us that the
globe with stand {al-kurah dhdt al-kursi) was
made for the holder of the dual office of wazir
{dhu al-wizdratayn, or wazTr of war and wazTr of
peace), the supreme commander-in-chief Abu
Tsa ibn Labbiin, and that the fixed stars were
placed on it in proportion to their magnitudes
(each indicated by a dot made in an engraved
circle). The globe is a hollow metal globe made
in two hemispheres with the seam along the
equator. It has 1015 stars with 47 labeled
constellations"'^ conforming to the basic designs
as described by al-Sufi, but drawn in a style rather
different from that appearing on most of the
other preserved Islamicate globes. Most of these
globes come from the eastern rather than the
western part of the Islamic world.
  The globe has a graduated ecliptic and equator
and six great circles passing through the ecliptic
poles (ecliptic latitude circles), but no polar or
tropic circles. The stand and rings are not the
original ones. The date of completion given on
the globe has been read as the first of Safar 473
H, (or 22 July AD 1080) and as the first of Safar
478 H, (or 28 May AD 1085). The dot under the
final abjad numeral is not sufficiently defined to
make absolutely certain whether the letter is a
hd'' (=8) or ajim (=3). IbrahTm ibn Sa^Td al-SahIT
is known to have been a prominent instrument
maker working in Valencia and Toledo, for a
mention of him among the mathematical scholars
in Andalusia occurs in The Book of the Categories
of Nations written by Sa^id al-AndalusT in AD
1068."^ At least four astrolabes dated from AD
1067-1086 are still known today.'"
  
A very similar globe having no date or maker s
name on it is now at the Bibliotheque Nationale
in Paris (No. 34 of the catalog). It is written in a
western (MaghribT) form of Kiific script. In de-
sign and star positions it closely resembles the
one signed by IbrahTm ibn SaTd al-SahlT.
  From the twelfth century there remain today
two Islamicate globes. The earlier of the two
(No. 59) is of the style in which only the major
stars or small star groups are represented, in this
case about 50 in all, including the 28 lunar man-
sions. Inlaid silver points are used to mark the
stars, but there are no constellation outlines. The
globe is metal with a seam along the ecliptic and
is inscribed in the eastern form of Kiific script.
The maker is stated to be Badr ibn ^Abdallah
Mawla BadT^ al-Zaman, who constructed the
globe in 535 H (or AD 1140-1141) making it the
earliest extant globe from the eastern part of the
Islamic world. The globe is curious in that the
only way of mounting it is apparently at the
ecliptic rather than equatorial poles. The merid-
ian ring is a recent replacement, but the horizon
ring may be contemporary with the globe.
  The expression badi"" al-zamdn at the end of
the maker's name means "marvel of the age,"
and was an epithet given Hibat Allah (died AD
1139-1140). An outstanding scholar of Baghdad
who spent his last years in Isfahan, he was consid-
ered the best astrolabe maker of his day."^ The
word mawla can mean an assistant or an appren-
tice, so that we might conjecture that Badr ibn
■^Abdallah had been an apprentice to Hibat Allah,
which was such a great distinction that he wished
to retain the title even after the death of Hibat
Allah. There is one astrolabe made in AD 1130
signed by Badr Mawla BadT*^ al-Zaman Hibat
Allah ibn Husayn al-Asturlabi, which must surely
be by the same maker.'"^
  The second celestial globe from the twelfth
century was made in 539 H/AD 1144-1145 by
Yiinus ibn al-Husayn al-AsturlabT (No. 2 of
the catalog). The globe is from the eastern part
of the Islamic world and is made of two metal
hemispheres joined more or less along what
would be the course of the Milky Way, though
the Milky Way is not designated on this globe
  
NUMBER 46

25



nor on any known Islamicate globe. It has en-
graved on it the 48 constellation figures and
approximately 1025 stars, represented by inlaid
silver dots that vary in size according to the
magnitude of the star. The inscription on the
globe states that the star catalog of Ptolemy was
used for the coordinates of the stars after adding
15° 18' to correct the longitudes for the inter-
vening time period. The stars are numbered
within each constellation and 72 of the stars are
named, as well as the 28 lunar mansions whose
names are inscribed alongside the ecliptic. As is
common, the meridian ring is missing and the
horizon ring is a recent replacement. Nothing is
known of this instrument maker and no other
examples of his work have been identified, but
he seems to have been a skilled craftsman and
was probably a professional instrument maker,
for he produced a precise globe as well as an
attractive one, which from the iconography sug-
gests a workshop in Persia, possibly Isfahan.
  From early in the twelfth century, probably
before AD 1120, and also from the eastern part
of the Islamic world, we have a fascinating trea-
tise on "The Sphere that Rotates by Itself" by
'"Abd al-Rahman al-KhazinT, who later dedicated
some other astronomical writings to the Saljiiq
ruler of eastern Persia from 1097 to 1157, Sanjar
ibn Malikshah. This treatise, which has been
recently edited and translated,'^° describes a ce-
lestial globe which, instead of being placed in the
usual set of rings, is half sunk in a box and
propelled so as to rotate once a day by a mecha-
nism of pulleys driven by a weight resting on top
a reservoir of sinking sand.
  This sphere, which al-KhazinT designed for
someone named Abii al-Husayn '"AIT ibn Muham-
mad ibn Tsa and executed with the help of a
carpenter {najjdr) named "^AIT al-SarakhsT, is set
halfway into a box whose upper surface serves to
mark the horizon. Over the sphere, along the
north-south line, was affixed a half-circle of brass
{nuhds), which served as a meridian ring, with
the axis of the globe set at the south pole inside
the box and the north pole at a hole in the
meridian ring corresponding to the geographical
latitude of Marw. The rest of the box covered

the top of the mechanism, which automatically
rotated the sphere one rotation per day. The
equator was inscribed on the globe and divided
into 360 equal parts. Taking as a pole a point
23°35' measured along the meridian ring from
the pole of the equator, the ecliptic was inscribed
and divided into zodiacal houses and degrees,
Al-KhazinT then indicated a circle on the surface
of the box around the globe, which he also di-
vided into 360°, and on which he named the
four points of the compass, so as to serve as the
horizon ring.
  Al-KhazinT furnishes no information as to the
material out of which the sphere was fashioned
nor the material from which he made a quadrant
{rub^). This quadrant, which he also called a
mistarah (rule), was equal to the quadrant of a
great circle on the globe. Much of the treatise is
devoted to using the "rule" and the globe in
much the same ways that Qusta ibn Luqa had
earlier suggested for a celestial globe set in the
customary rings (see Chapter 2 below). This
rather ingenious conception of an automatically
rotating celestial globe does not in fact appear to
have been very practical because of the enormous
size of the mechanism required and the difficulty
of precisely regulating the movement. It does
not seem to have influenced the development of
celestial globes. There are, of course, many ce-
lestial globes extant that do not have rings, and
it is always possible that some of them were
mounted in something other than the usual me-
ridian-horizon ring assembly. Consequently one
cannot dismiss entirely the possibility that some
globes were used in such an arrangement, al-
though there is no corroborating evidence to
suggest this for any extant globe.
  From the thirteenth century several signed
globes are preserved today. The earliest and in
many ways the loveliest was made in AD 1225-
1226 for a nephew of the great Saladin (Salah al-
DTn) in Egypt. The globe (No. 3) has the full set
of 48 constellation figures engraved and damas-
cened with copper, with approximately 1025
stars indicated by six different sizes of inlaid
silver points corresponding to the various mag-
nitudes. The globe also has a scale showing the
  
26

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



sizes of silver points used for the first five mag-
nitudes. The inscriptions, which are in naskhi
script rather than Kufic script that would be
more usual for this period, are damascened with
silver or inlaid with silver wire. According to the
inscription, the globe was made for al-Malik al-
Kamil, the Ayytibid sultan of Egypt from AD
1218 to 1238, who had interests in irrigation
works and other agricultural matters and, follow-
ing the conclusion of the fifth crusade, corre-
sponded with Emperor Frederick II of Sicily on
scientific subjects.'"^' The maker of the globe was
C^Alam al-DTn) Qaysar ibn Abi al-Qasim ibn Mu-
safir al-Ashrafi al-Hanafi, who was born in upper
Egypt in AD 1178-1179 and later resided in
Syria, where he died in 1251 in Damascus. He
was a renowned mathematician and architect,'^^
and it is reported that the historian QadTJamal
al-DTn ibn WasTl assisted Qaysar in constructing
another celestial globe of wood and gilt.'^^
  The globe, made by Qaysar in AD 1225-1226,
is unusual in that the horizon ring and stand,
which are probably the ones originally made for
this globe, have incorporated into them two gno-
mons and graduated arcs making them elevation
dials (see Chapter 2). The sphere itself is unusual
in having Latin zodiacal names and Latin numer-
als engraved on it, but these may have been
added later.
  Another example of a thirteenth-century Is-
lamicate globe comes from the famous school of
metalwork that flourished in Mosul on the Tigris
River north of Baghdad. The globe (No. 4) is
made by Muhammad ibn Hilal al-Munajjim al-
MawsilT (the astronomer of Mosul) in 674 H/AD
1275-1276. The globe is again made of two
metal hemispheres with 48 engraved constella-
tions and about 1025 stars indicated by inlaid
silver points with the inscriptions in Kiific script.
Nothing is known of this maker except this one
product, but judging by his name, he was an
astonomer from Mosul who was probably trained
in metalworking at the great center there. The
globe is a handsome example of Islamicate metal-
work as well as fine instrument-making, and com-
pares favorably with other products of the Mosul
school.''-'''
  
An undated but signed globe of the later thir-
teenth century (No. 5) was constructed by Mu-
hammad ibn Mu^ayyad al-^UrdT, the son of the
famous astronomer Mu^'ayyad al-DTn al-^UrdT al-
DimishqT, who made instruments for the observ-
atory at Maragha, about 50 miles south of Tabriz
in Azerbaijan.'-^"^ The observatory had been es-
tablished by order of HQlagu Khan (the grandson
of the founder of the Mongol Ilkhanid dynasty
of Persia, Genghis Khan) under the direction of
the astronomer NasTr al-DTn al-JusT (born AD
1201).'^*^ The observatory was built outside of
the city and the foundations remain visible today.
  The globe was clearly made at the court of
Hulagu Khan in Maragha and may well have
been, even though rather small, one of the sev-
eral celestial globes no doubt used at the observ-
atory, which continued to operate for a period
even after the death of Hulagu in AD 1265. It is
probably good evidence of the type of celestial
globe carried from Persia, along with six other
astronomical instruments, by Cha-ma-lu-ting
(i.e., Jamal al-DTn) to HQlagQ Khan's brother
Kublay Khan, the first Yuan Emperor of
China.'^' This particular undated globe, made
of metal hemispheres, has 48 constellations and
about 1025 stars indicated by inlaid silver points.
The ecliptic latitude circles are damascened in
silver at the zodiacal divisions, while the two
circles forming the graduated band representing
the equator are inlaid in gold wire. The names
of the zodiacal signs are inlaid alternately in gold
and silver; the script is Kufic. The meridian,
zenith, and horizon rings are probably original,
and are unusual in that the horizon ring has no
stand but rather a suspensory device by which it
can be hung. The zenith ring appears to pivot
about the sphere and to have a slit parallel to the
face of the ring along the upper quadrants. It is
possible that a needle-like gnomon was at one
time inserted through this slit. The basic design
of the rings is reminiscent of that described by
al-BattanT in the ninth century, and in fact it is
the only extant example of a ring assembly closely
resembling that described by al-BattanT.
  Another globe (No. 35) very similar in style to
this globe by Muhammad ibn Mu^ayyad al-'^UrdT
  
NUMBER 46

27



is unsigned and undated. This very precisely
executed globe, which unfortunately lacks the
rings, was probably made about five years later
possibly in Maragha or Tabriz, and is a fine
example of Ilkhanid metalwork.'^**
  Further evidence of the widespread distribu-
tion of celestial globes in the thirteenth century
is seen in the report''^^ that when Hulagu Khan
captured AlamQt (the mountain stronghold
south of the Caspian Sea) in AD 1256, he found
several astronomical instruments including a ce-
lestial globe with stand {dhdt al-kursi). From the
western part of the Islamic world in the thir-
teenth century there comes a treatise on astro-
nomical instruments by the Morrocan astrono-
mer al-MarrakushT, which is an important source
of information on the construction of celestial
globes (see Chapter 2).
  Another globe of the thirteenth century (No.
6) dated 684 H/AD 1285-1286 is slightly smaller
even than the one made in Maragha. It has the
full set of constellation outlines and about 1024
stars. It is engraved in KQfic script and bears the
maker's name, Muhammad ibn Mahmud al-Ta-
barT. Nothing is known about him, though the
name suggests an origin in Tabaristan, a province
in northern Persia on the southern shore of the
Caspian Sea.
  This globe presents some questionable or cu-
rious features. It is a hollow metal sphere having
no visible seam and with two plugs (see Figure 6,
where the plugs are marked with arrows), imply-
ing that it might have been cast by the lost wax
method. This method is characteristic of Indo-
Persian globes of a much later period (see Chap-
ters 2 and 3). It is always possible that the tech-
nique for constructing globes in this manner was
developed and perfected much earlier than the
other remaining globes would lead one to sus-
pect; and this particular globe was clearly made
by a metalworker experienced and skilled in
making seamless globes. The unusual features,
however, make this supposition less likely than it
would otherwise be. The KQfic script on this
globe is at times not very angular and tending
towards naskhi {generaWy, but not without excep-
tion, characteristic of later products), as in form-

ing the letters sdd, ddd and "ain. The lines rep-
resenting the celestial equator and ecliptic wob-
ble slightly and the graduations are uneven. At
the equinoxes the crossing of the equator and
ecliptic are quite poorly executed compared with
other globes, and in fact, some graduations are
lacking. The stars themselves are so poorly posi-
tioned that in some instances it is difficult to align
a given star on the globe with a star in illustra-
tions and catalogs such as those given by al-Sufi.
On the basis of the stars near the ecliptic that
can be identified, their positions when compared
with well-made dated globes, indicate an epoch
varying from the tenth to the sixteenth century.
The stand and rings, which from the general
appearance of the alloy, finish, and engraving
appear to be contemporary with the globe, are
graduated in a different scale than that used on
the globe (an incongruous and unusual feature)
and the directions in which the numbering runs
on both the meridian and horizon rings are the
opposite from those found on nearly every other
extant set of rings. The horizon ring and stand
seem to have been intended to serve as an ele-
vation dial like those on the globe by Qaysar in
the thirteenth century. The gnomons, however,
have not been properly aligned over the gradu-
ated arcs so that they cannot function; this misa-
lignment could, of course, possibly be due to
more recent incorrect repairs and restoration.
  The globe has a second inscription stating that
the stars were placed on it according to the Book
of the Constellations of the Fixed Stars by al-SQfi
after increasing their longitude by five degrees
(which gives an unusually low value for the rate
of precession), followed by the interesting caveat
"while the correction of whatever errors occur
and the production of variant readings [is left] to
those who are expert in language." It appears
that the stars were actually placed on the globe
after the constellations outlines were incised, and
therefore that the coordinates from a star cata-
logue were not really used to determine the star
positions despite the statement on the globe.
Hence, one can conclude that while the globe
was fashioned by a skillful metalworker, it was
inscribed by someone who was apparently unfa-
  
28

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 6.—Globe No. 6, dated 684 H/AD 1285-1286, by Muhammad ibn Mahmud al-TabarT.
Musee du Louvre, Section Islamique. Arrows indicate plugs in holes resulting from casting
process. (Photo: Courtesy of the Musee du Louvre)

miliar with the techniques for precisely executing
the ecliptic and equator at the equinoxes, for
holding steady the engraving of the great circles,
for marking equal graduations, or techniques for
accurately placing the stars on a globe.
  The lack of precision evident on this globe
indicates that it was not made by a professional
instrument maker for the use of an astronomer,
but was made by a skilled artisan for its artistic

appeal. It is curious, however, that the maker,
even specifying in the second inscription on the
globe that he was the engraver and hence not
just the caster of the globe, would sign himself
as al-Asturldbi (an astrolabe maker), when he did
not make an astronomically correct instrument.
Such cases of imprecision resulting in a techni-
cally inaccurate and non-functional globe present
the possibility that it is a much later attempt by

NUMBER 46

29



an artisan to copy an earlier thirteenth-century
globe for sale to someone interested in the artistic
merit rather than the actual use of the globe. In
addition, the fact that someone has signed and
dated a non-functional globe—even though an
attractive one—perhaps is further evidence that
it is a relatively recent intentional reproduction
of a thirteenth-century globe rather than simply
a globe produced in the thirteenth century by a
metalworker who did not know how to make
accurate celestial globes but did know the com-
plicated process for casting seamless ones. On the

other hand, it might be concluded from "^Abd al-
Rahman al-SQfi's statement about "decorated ce-
lestial globes" that there is a long tradition going
back to at least the ninth century AD in the
Islamic world of producing quite inaccurate ce-
lestial globes.
  Many of the criticisms made of the globe by
al-TabarT can also be made of the globe that, by
its date, is the next oldest extant globe (No. 60).
This globe (see Figure 7) is also a hollow metal
globe with no observable seam, several plugs
visible, and no seam detectable by probing the
  

FIGURE 7.—Globe No. 60, dated 718 H/AD 1318-1319, by "Abd al-Rahman ibn Burhan al-
MawsilT. Oxford, Museum of the History of Science. (Photo: M.B. Smith)

30

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



interior surface. It is therefore likely to have also
been cast by the cire perdue process. The globe
has only about 100 stars indicated by inlaid silver
points, and has no constellation outlines. The
sometimes poorly executed script is naskhi, which
is unusual for globes prior to the sixteenth cen-
tury.
  One inscription on the globe reads: "the work
of "Abd al-Rahman ibn Burhan al-MawsilT in the
months [sic] of the year seven hundred and eigh-
teen [AD 1318-1319]." A second inscription is
somewhat obscure, but seems to read, "Under-
taken at the instigation of the poor, the weak
Ghiyath known as al-MansQr in the months of
the year [lacuna] and seven hundred." The dates
in  both  these inscriptions are written out  in

words, which is extremely unusual. A third in-
scription appears to read "By order of the trea-
suries of the Sultan al-Matt ["the proud," which
with some modification could be read as al-malik
"the ruler"] al-"Adil Ulugh Beg." If the date
given by the maker Ibn Burhan of Mosul is
considered correct, then the last inscription must
have been added later, for Ulugh Beg founded
the observatory at Samarqand in the fifteenth
century, not fourteenth. All three inscriptions
are quite similar in style and appearance, how-
ever, and do not give the immediate impression
of having been done by different engravers (see
Figure 8).
  All the great circles on the globe are somewhat
wavy, the graduations uneven, and the crossing
  

FIGURE 8.—Detail of inscriptions on globe No. 60, dated 718 H/AD 1318-1319, by "Abd al-
Rahm5n ibn Burhan al-MawsilT. Oxford, Museum of the History of Science. (Photo: M.B.
Smith)

NUMBER 46

31



at the equinoxes poorly executed. The globe also
has the lunar mansions along the ecliptic indi-
cated by patterns of dots that are not to be found
in the strictly astronomical literature, such as the
writings of al-Sufi or al-BTrQnT, but rather in
treatises such as the Marvels of Things Created
and Miraculous Aspects of Things Existing written
by al-QazwTnT in the thirteenth century and in
writings concerned with lunar mansions and pre-
Islamic poetry and calendars (see section on Lu-
nar Mansions, Chapter 5). These patterns of dots
representing the lunar mansions are a unique
feature among extant Islamicate globes and rep-
resent a different tradition used by the maker.
The word for "the work of" {''amal) used in the
signature occurs on only one other globe before
the seventeenth century, and both globes are
inaccurate and non-functional. The pattern for
graduating the equator is nearly unique and is

not as useful as the common pattern.
  The evidence of the circles, graduations, and
star positions and the conflict inherent in the
inscriptions make rather dubious the authenticity
of this globe as an example of early fourteenth-
century globe making. In any case it was certainly
not made by or for an astronomer (and Ulugh
Beg, for whom it was supposedly made, was very
well versed in astronomy). It is inaccurate and of
value only as an example of artistic metalwork
and not as a functioning celestial globe.
  The fourteenth and fifteenth centuries saw
one workshop of instrument makers who pro-
duced four celestial globes still preserved today.
In Kirman in southeast Persia a well-known as-
trolabe maker Ja"far ibn "Umar ibn Dawlatshah
al-KirmanT made two celestial globes, one in 764
H/AD 1362-1363 (No. 7 of the catalog; see Fig-
ure 9), and one in 785 H/AD 1383-1384 (No. 8
  

FIGURE 9.—Globe No. 7, dated 764 H/AD 1362-1363, by Ja"far ibn "Umar ibn Dawlatshah al-
KirmanT. Oxford, Museum of the History of Science. (Photo: M.B. Smith)

32

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



of the catalog). Both globes are made of metal
hemispheres and have the full set of consetalla-
tion figures engraved with approximately 1024
stars. Both are engraved in KQfic script and bear
inscriptions stating that the star positions were
taken from the Book of Constellations by AbQ al-
Husayn "Abd al-Rahman al-SQfi, with an increase
in the longitude of 6°3' for the earlier globe.
The later globe has a lacuna in the inscription
where the minutes are given, but using the same
value he used earlier (and which his son used
later), there would be an increase of 6° 23' on
the second globe. On the earlier globe the in-
scription giving the maker is engraved over an
earlier one saying that Ja"far had made it for a
certain Muhammad ibn AsTl, but for unknown
reasons the maker attempted to obliterate the
name of this patron. This Ja"far ibn "Umar is
also known from four astrolabes made between
AD 1354 and 1388.'^"
  Muhammad, the son of Ja"far, was also an
instrument maker who is known to have made
two celestial globes and two astrolabes'^' that are
preserved today. Both globes by the son are of
the style having only major or astrolabe stars and
no constellation outlines and are made of metal
hemispheres. One globe, made in 813 H/AD
1410-1411 is signed simply Muhammad ibn
Ja"far al-AsturlabT (No. 61). The other globe
(No. 62; see Figure 10) is dated 834 H/AD 1430-
1431, and signed Muhammad ibn Ja^far ibn
"Umar al-AsturlabT, known as {al-mulaqqab)
Hilal. Some have read the last name as Jalal
rather than Hilal. The original rings and stand
are still preserved with the latter globe. The
father's globes are precisely made, clearly exe-
cuted in the tradition of a professional scientific
instrument maker. The son's globes, on the other
hand, tend not to be as accurately inscribed as
those by his father, with slight wobbles in the
great circles and some irregularity in the gradu-
ation. The latter globes display a certain care-
lessness in execution, but are by no means as
technically inaccurate and non-functional as
many of the extant globes, such as the two pre-
viously mentioned.
About the same time that Muhammad ibn

Ja"far was producing instruments in Kirman, an
observatory was flourishing at Samarqand to the
northeast in Transoxiana. The observatory had
been founded by Ulugh Beg ibn Shah-Rukh'^^
in AD 1420, where a staff of astronomers headed
by JamshTd Ghiyath al-DTn al-KashT carried out
observations and compiled astronomical tables
and a star catalog titled Zij-i Sultdni (AD 1440).
This star catalog,'^^ which was essentially that of
al-SQfi with improvements in the coordinates,
was an important source for star positions for
subsequent globe makers in the eastern part of
the Islamic world. In a letter to his father Ghiyath
al-DTn al-KashT described several of the instru-
ments at the observatory, in the course of which
he made an indirect reference to a celestial
globe.'^"
  The sixteenth century is represented by only
two globes, both of them hollow metal globes
that appear to have no seam but show plugs that
are probably from the lost wax casting process.
The earliest of the two is, unfortunately, an
inaccurate and non-functional globe (No. 9). The
inscription giving the maker has been marred by
a large rectangular hole and a plug of lead, but
informs us that the maker was one Jamal al-DTn
Muhammad [ibn] . . . al-DTn Muhammad, who
says he was a Hashimid'^^ by ancestry {al-Hdshimi
nasaban) and a Meccan by birth {al-Makki maw-
lidan). The date is given as 981 H and Yazdijird
952 (AD 1573-1574). Despite the fact that the
maker specifies that he was born in Mecca and a
member of the Hashimid family, which ruled
Mecca from the tenth century to the first part of
the present century, it is unlikely that the globe
was made in Arabia. There is no further evidence
that seamless metal globes were ever made other
than in the eastern part of the Islamic world and
the date is given in the Yazdijird era, which is a
Persian calendar of Sassanian origin and unlikely
to have been used in Arabia even in the sixteenth
century under Ottoman rule.
FIGURE 10.—Globe No. 62, dated 834 H/AD 1430-1431,
by Muhammad ibn Ja"far ibn "Umar al-Ast^urlabi. London,
British Museum, Department of Oriental Antiquities.
(Photo: M.B. Smith)

NUMBER 46

33






34

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



  Nothing is known of the maker, nor are any
other examples of his work known. The globe
itself has poorly drawn constellation figures with
disproportionately large stars added after the
constellation outlines were engraved and very
inaccurately positioned. The graduations are
very irregular and there are no holes at the
celestial poles, which would be necessary for it to
function properly. The maker has also used the
word "'amal for "the work of in the signature,
which does not occur before the seventeenth
century except on globe No. 60 (mentioned
above), which is also non-functional and has
many questionable features. The imprecise na-
ture of this design suggests that the maker was
not a professional instrument maker, but rather
a metal worker desiring to produce an object of
decorative appeal but not one of use to an as-
tronomer.
  It is most unfortunate that the three earliest
extant examples of hollow seamless metal globes
(Nos. 6, 9, and 60) should all be of such an
inaccurate and non-functional form with dubious
characteristics about them, produced by metal-
workers who are otherwise unknown. As with all
such globes, they give rise to speculation that
they might be later attempts by metalworkers
who were not professional instrument makers to
make a reproduction of an older celestial globe
for a market that does not require technical
accuracy and would not recognize errors or in-
consistencies. Such assertions cannot be defi-
nitely substantiated, however, and it is always
possible that in the thirteenth, fourteenth, and
fifteenth centuries there was also a market for
technically inaccurate but artistically pleasing ce-
lestial globes.
  When we turn to the second globe of the
sixteenth century (No. 10, illustrated in Figures
11 and 69) we see a very fine example of a hollow
metal seamless globe produced, judging from the
maker s name and the basic style of the globe, in
KashmTr in the western Himalayas in northern
India. The globe was made by "AIT KashmTrT ibn
Luqman and is dated 998 H, the thirty-fourth
year of the reign of the Mughal Emperor Akbar,
which is equivalent to AD 1589-1590. The globe

was made just one year after Akbar entered the
valley and KashmTr became a suba (province) of
the Mughal Empire.'^^ The maker, who is not
known by any other work, adds that he made the
globe as an aide-memoire to astrolabe makers. The
globe is a well-constructed piece of metalwork as
well as precisely executed instrument, with con-
stellation outlines depicted in a style typical of
early Mughal art. While the present owner did
not permit examination of the globe, it can be
asserted with some certainty, based on the avail-
able photographs, that the globe is a seamless,
probably cast, globe. Some of the plugs (probably
from the casting process) are indicated by arrows
in Figure 11. Consequently, we can be fairly
certain that the complicated technique for mak-
ing seamless metal globes was well established in
northwestern India by the last quarter of the
sixteenth century. (See Chapter 2 for a discussion
of the method of constructing such a globe.)
  In the sixteenth and seventeenth centuries in
northwest India a family at Lahore maintained a
remarkable workshop that, through four gener-
ations, produced numerous well-made scientific
instruments, in particular planispheric astrolabes
and celestial globes. Lahore, on the upper course
of the Indus river, was the capital of a Mughal
province or suba of the same name, later called
the Punjab. The activity of these metalworkers
covered the reigns of the second through the
ninth Mughal rulers of India, who spoke the
vernacular Turki but maintained Persian as the
official language of the court.
  This particular family of metalworkers from
Lahore excelled in certain metallurgical tech-
niques, in particular the production of hollow
cast globes. The variety in the techniques and
designs of the engraved images on these globes
suggests that the family maintained a workshop
of metalworkers and apprentices, in which sev-
eral people would perhaps be involved in the
production of an instrument which nevertheless
bore only the name of the family member super-
vising that production.
  The earliest extant instrument by this family
is an astrolabe made in 975 H/AD 1567-1568 by
the apparent founder of the workshop, Allahdad.
  
NUMBER 46

35







FIGURE 11.—Detail showing signature of maker on globe No. 10, dated 998 H/AD 1589-1590,
by "AIT KashmTrT ibn Luqman. London, Private Collection. Arrows indicate plugs. (Photo: Alain
Brieux)

36

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



He called himself simply Ustadh Allahdad Astur-
labT LahurT, that is. Master-craftsman Allahdad,
the Astrolabist from Lahore. Three extant astro-
labes were made by him, only one of which is
dated.''^^ The name Allahdad is a compound of
Alldh (God) and ddd (gift).''^^ It is only from other
members of the family—his grandsons and great-
grandsons—that further information about Al-
lahdad can be gathered. In the name as it is
written by later family members, Shaykh Allah-
dad AsturlabT HumayQnT LahQrT, it is likely that
HumayunT was intended to indicate the fact that
the founder of the workshop had lived at the
timeof HumayQn,'^^ who ruled India from 1530
to 1556 as the son and successor of Babur, the
TTmQrid conqueror who had come from Kabul
in the Afghan mountains into the Indus plain to
found the Mughal dynasty in India.
  There is a possibility, however, that Allahdad
was also an instrument maker at the court of
HumayQn and that the nisbah (the name usually
derived from the place of origin or trade) was
intended here to be an honorific indicating royal
patronage, though no mention is made of him in
the records of Humayun's reign. If Allahdad was
an astrolabist at the court of Humayun he must
have been quite young, for his son was flourish-
ing as a craftsman 52 years after the death of
Humayun, and his grandsons were productive in
the second quarter and middle of the seven-
teenth century.
  HumayQn's interest in astronomy and astrol-
ogy was well-known. In the Akbarndma^"*^ written
by AbQ al-Fadl, minister to his son and successor,
he was mentioned as having "an extraordinary
interest and proficiency in the astrolabe, the ce-
lestial globe {kurah), and other instruments of
the observatory." His grandson, JahangTr, men-
tions in his memoirs''" that he was presented
"with a compendium in the handwriting of the
late king HumayQn . . . containing some prayers,
an introduction to the science of astronomy, and
other marvelous things, most of which he had
studied and carried into practice." HumayQn was
also responsible for attracting from Persia var-
ious artisans with whom he had become familiar
during his temporary exile in Persia following

his defeat by his brother Sher Khan, after which
he was able to regain control only with the aid
of Shah Tahmasp, the Safavid ruler of Persia,
himself a patron of the arts. The resulting blend-
ing in northwest India of the skills and styles of
Persian artists with those of native Indians sub-
sequently trained by these Persian artisans re-
sulted in a style commonly referred to as Indo-
Persian.
  The one dated instrument bearing the name
of Allahdad was made during the reign of the
successor to HumayQn, Akbar, who ruled from
AD 1556-1605. Akbar had moved the Mughal
court to Lahore, and by 1596 the empire was
consolidated and the borders set to include an
area from the Arabian Sea to the Bay of Bengal
and from the Himalayas to the Narbada river, as
well as the Kabul province (Afghanistan) west of
the Indus. Akbar also entertained European con-
tacts at his court through the Jesuit missions and
embassies sent for promoting trading interests.
  The son of Allahdad, Mulla "Tsa ibn Allahdad,
also worked during the reign of Akbar. The five
extant astrolabes under his name date from AD
1600 to 1604.'''^ There are no celestial globes
known to have been made by him.
  "Isa ibn Allahdad had two sons who were also
instrument makers: Muhammad MuqTm and
Qaim Muhammad. The two brothers signed two
astrolabes as co-makers. One is dated 1017 H/AD
1609; the other is estimated to be made about
1635.'"*^ Muhammad MuqTm produced under his
name alone 33 instruments, of which fourteen
astrolabes can be dated between AD 1624 and
1660.'''^ He is known to have made one celestial
globe (No. 15), dated 1049 H/AD 1639-1640,
which appears to be a seamless metal globe with
plugs.
  His brother Qa^im Muhammad signed two
astrolabes,'^"^ in addition to the two on which the
two brothers collaborated, and four celestial
globes. The earliest globe (No. 11 of the catalog)
was constructed for a certain Nawab I"tiqad
Khan in 1032 H/AD 1622-1623 (the eighteenth
year of the reign of JahangTr, who ruled from
1605 to 1626; see Figures 12 and 13). The
memoirs of JahangTr mention two men by the
  
NUMBER 46

37






FIGURE 12.—Detail showing signature of maker, globe No. 11, dated eighteenth year of the
reign of JahangTr (AD 1622-1623), by Qa^im Muhammad ibn "Isa ibn Allahdad AsturlabT
LahurT HumayiinT. Lancashire, Stonyhurst College. Arrow indicates plug. (Photo: M.B. Smith)

name of I"tiqad Khan, either one of whom might
have been the patron requesting the celestial
globe. One was given the title of Asaf Khan and
was the brother of NQr Jahan, the wife of Jahan-
gTr, and a leading figure at the court.'''^ The
other was a younger brother who was known
originally as ShapQr and during the eighteenth
year of JahangTr's reign was the governor of
KashmTr, at that time part of the Mughal prov-
ince of Kabul.'^' Unfortunately, however, there
is no mention of this family of instrument makers

in the memoirs of JahangTr or in any other writ-
ten source from Mughal India—an omission
which tends to indicate that the name HumayQnT
was used to indicate contemporaneity of Allah-
dad with HumayQn, and that even if the founder
of the workshop had enjoyed royal favor, his
descendants probably did not.'"** While perhaps
not actually enjoying royal favor, any good ar-
tisan in the city where the court was located
would have an adequate supply of wealthy cus-
tomers among the retainers.

38

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY




FIGURE 13.—Detail showing Virgo. Globe No. II, dated
eighteenth year of the reign of JahangTr (AD 1622-1623),
by Qa^im Muhammad of the Lahore workshop. Lancashire,
Stonyhurst College. (Photo: M.B. Smith)
  Of the other three celestial globes by Qa^im
Muhammad (Nos. 12, 13, and 14) the first two
were made in the twenty-first and twenty-second
years of the reign of JahangTr (AD 1625-1626
and 1626-1627), while the third was produced
in 1047 H/AD 1637-1638. This was during the
sumptuous rule of Shah Jahan, who built the
lovely sepulchre for his wife Mumtaz Mahal in
Agra, known as the Taj Mahal. (For illustrations
of the first two globes see Figures 36, 38, and
42.) On the latter globe Qa^'im Muhammad sup-
plied in a Persian inscription some information
on the design of the globe, saying that 1022 stars
in 48 constellations were placed on it based upon
the positions determined at the observatory of
Ulugh Beg and that three degrees were added
to their celestial longitudes so as to have the
positions correspond to the date  1047 H. It is

also likely, for reasons which are explained in
Chapters 3 and 4 of this study, that the anony-
mous globe now at the Smithsonian Institution
which forms the focus of the study, may_also be
attributed to Qa^im Muhammad ibn Tsa ibn
Allahdad AsturlabT LahurT HumayQnT, All of the
globes made by Qa^im Muhammad have a full
set of stars and the 48 constellation outlines. It
is likely that the catalog of Ulugh Beg was used
as a source for the star positions not only for the
globe made in AD 1637-1638 but also for all the
globes of the Lahore workshop.
  Qa^'im Muhammad and Muhammad MuqTm
both had sons who continued the workshop for
a fourth generation. Muhammad MuqTm had two
sons, Jamal al-DTn and Hamid. Only three instru-
ments are known to be made by Jamal al-DTn, all
of them astrolabes dating from_1681 to 1692
during the reign of AurangzTb "AlamgTr I.'^
  Hamid attempted one celestial globe that we
know of (No. 68) dated 21 Shawwal 1065 (24
August AD 1655), in addition to seven extant
astrolabes ranging from AD 1628 to 1691, span-
ning the reigns of Shah Jahan (AD 1628-1657),
Murad Bakhsh and Shah Shuja'' (who both ruled
in AD 1657), and AurangzTb "AlamgTr I (AD
1658-1707).''^*^ An eighth astrolabe made in ap-
proximately AD 1654 bears the signature of both
Hamid and his cousin Diya^ al-DTn Muham-
mad.'"^' The only globe known to be made by
Hamid (see Figure 14) is the type of globe having
only the major or astrolabe stars, in this case 67,
and no constellation outlines. It too appears to
be a seamless cast globe.
  Diya^ al-DTn Muhammad was the son of Qa^im
Muhammad and clearly the most prolific pro-
ducer of instruments from this workshop. His
instruments exhibit great variety in the quality
of workmanship, as might be expected from such
a large number. The range cannot be explained
in terms of the evolution of a technique, but is
probably due to the various assistants involved in
their production. Yet certain artistic and techni-
cal features characterize and differentiate the
products of Qa^im Muhammad and his son Diya''
al-DTn Muhammad and their respective assist-
ants. In Chapter 4 Andrea Belloli discusses the
  
NUMBER 46

39






FIGURE 14.—Detail showing signature of maker. Globe No. 68, dated 21 Shawwal 1065 H/AD
1665 by Hamid ibn Muhammad MuqTm ibn "Isa ibn Allahdad AsturlabT LahurT HumayunT.
Cambridge, Whipple Museum of the History of Science. (Photo: Whipple Museum of the
History of Science)

artistic aspects of their celestial globes, while the
distinguishing features of their designs from the
astronomical standpoint are treated in Chapter
3.
   Piya" al-DTn Muhammad produced the largest
number of Islamicate celestial globes attributable

to any single maker, and the production of these
globes was spread fairly evenly over his working
life, which was at least 44 years. In addition to
the astrolabe made in collaboration with his cou-
sin Hamid, 25 astrolabes are signed by Diya^ al-
DTn Muhammad, of which 20 are dated from AD

40

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 15.—Globe No. 18, dated 1055 H/AD 1645-1646, by Diya^ al-DTn Muhammad of the
Lahore workshop. The meridian ring is not properly mounted. Columbia University, David
Eugene Smith Collection. (Photo: Columbia University)

1637 to 1680, from the reigns of Shahjahan to
that of AurangzTb "AlamgTr.'^^ The 15 celestial
globes (with a single exception they are hollow
seamless globes cast by the cire perdue process)

date from 1645 to 1680 (Nos. 18-24, 26-28,
30, 66, 69, and 71 of the catalog; see Figures 15
and 16). Eleven of the globes have the 48 con-
stellation figures with approximately 1018 stars

NUMBER 46

41


FIGURE 16.—Globe No. 27, dated 1074 H/AD 1663-1664, by Diya"" al-DTn Muhammad of the
Lahore workshop. Royal Scottish Museum, Department of Technology. (Photo: Royal Scottish
Museum)

42

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



indicated, while three have no constellation out-
lines and only major stars indicated.
  The one celestial globe signed by Diya^ al-DTn
Muhammad that is not a seamless globe (No. 30)
is in fact not really a globe, but an open, cut-out
sphere consisting of two raised or hammered
hemispheres, in which the spaces between the
constellations and the great circles have been cut

out, or worked a jour. The stars are indicated by
holes perforating the sphere, some of which seem
to be filled with glass or mica. The metal hemi-
spheres, which have been covered with gilt paint
and varnish, are easily separated and meet along
the line of the equator (see Figure 17). The
inscription on the sphere states that it was made
at the order of the  Mughal ruler AurangzTb


FIGURE 17.—Globe No. 30, dated 1090 H/AD I679-I680, by Diya" al-DTn Muhammad of the
Lahore workshop. Paris, Private Collection. (Photo: Alain Brieux)

NUMBER 46

43



"AlamgTr, and calls the sphere kurah ikhtird^i
ardwi samd^i (a specially invented terrestrial-ce-
lestial sphere). This statement implies that since
it was both a terrestrial and celestial sphere, there
was once a small terrestrial globe placed in it. By
having the stars as well as the spaces between the
constellations and circles cut out, sufficient light
could enter the sphere to allow inspection of the
enclosed terrestrial globe. Thus this sphere, es-
pecially designed for the Mughal ruler "AlamgTr
I, the third son of Shah Jahan, seems to have
been part of a tradition of demonstrational ar-
millary spheres. It is of unique design among
extant Islamicate instruments, for no other ar-
millary spheres are extant. Armillary spheres are
seldom illustrated or mentioned in Islamicate
astronomical literature, and when they are, ex-
cept where there was European influence, they
are nearly all observational armillary spheres in
the Ptolemaic tradition rather than demonstra-
tional armillary spheres as this sphere would have
been.'^3
  The last instrument known to have been exe-
cuted by piya^ al-DTn Muhammad is a type of
astrolabe called a safihah zarqaliyah, different
from the conventional planispheric kind.'^'' It is
dated the twenty-second year of the reign of
AurangzTb "AlamgTr (AD 1680), and according to
the inscription it was made at Delhi for Nawab
Iftikar Khan who was fawjdar (a police official)
at Jawnpur, a city on the Ganges River in the
state of Allahabad (later Oudh). This patron is
possibly Sultan Husayn, who was given the title
Iftikar Khan in the first year of the reign of
AurangzTb "AlamgTr and ruled Jawnpur at the
time the instrument was made. This large astro-
labe had a diameter of 610 mm and was later
placed at the Jaipur Observatory built by Jai
Singh in AD 1734. The fact that this last item is
stated to have been made in Delhi and the celes-
tial-terrestrial sphere made at just about the same
time was executed for the Mughal emperor him-
self, suggests that the workshop of the family had
by that time moved to Delhi, the seat of the
Mughal court since AD 1648,
  This family of Mughal craftsmen from the
Punjab of northwest India (modern  Pakistan)

were clearly the most important producers of
hollow metal seamless globes. It is evident that
the Lahore family must have prided themselves
on their ability to produce a globe by such a
technique, for using the lost wax method would
involve long hours of tedious work, both in clean-
ing out the inner mold and carefully plugging
the holes caused by the cooling process (see
Chapters 2 and 3). Of course assistants were
employed for much of the tedious work. A globe
in hemispheres could be made in a small fraction
of the time required to cast one by the cire perdue
process, for which reason we must assume that
pride of craftsmanship—quite justified by their
finest examples—rather than expediency must
have determined their constant employment of
this method.
  In terms of the number of extant Islamicate
celestial globes, this family is also important.
Their workshop claims 21 signed globes—the
largest number from a single shop. Moreover, it
is very likely that not only the anonymous globe
now at the Smithsonian Institution (No. 38) but
also a considerable number of other Indo-Persian
seamless metal globes could be attributed to this
workshop, including Nos. 39-44. One of the
globes signed by Diya^ al-DTn Muhammad (No.
28, see Figures 43 and 47) is known to have
inside it three rolled up pieces of paper with
writing on them (possibly amulets?), which have
been sewn together and appear to have at one
time been wrapped around a rather long heavy
metal probe. The diameter of the probe is
greater than that of the holes in the globe. Since
there are no missing plugs or any other way to
gain access to the inside except through the quite
small holes at the poles, these items have not
been examined up to this time. It seems that they
must have been placed in the globe by the maker
before the final plug was set into place. Was this
a common practice in globe manufacturing?
Would attempts to retrieve such items account
for the considerable number of globes that have
missing plugs (see Figure 36)? Another explana-
tion for some of the plugs being purposefully
removed might be that seamless cast globes some-
times have pebbles or bits of grit left in them
  
44

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



from the casting process, and when shaken, they
rattle (for example, Nos. 40 and 42). People may
have hoped to find a treasure in them and forced
an opening by removing a plug.
  We know that there were other scientific in-
strument makers in Mughal India besides this
particular family. One, an astrolabe maker for
Humayun, was Mawlana MaqsQd HirawT, who
"was one of the devotees of the Resident of
Paradise [i.e., HumayQn]. . . . He manufactured
astrolabes, globes and other instruments in such
a manner that observers of his products were
filled with wonder."'^^ Unfortunately, none of
his celestial globes or astrolabes are known to be
extant. During the reigns of Akbar and JahangTr
the craftsman "AIT ibn "Iwad produced two
known astrolabes,'^^ while during Shah Jahan's
rule Ibn Muhibb Haqiqah made an astrolabe.'^'
No globes are known to have been made by these
two Mughal artisans.
  Besides the Lahore family workshop, there was
in the seventheenth century another maker in
northwestern India who was producing globes
that appear to be cast seamless globes. The in-
strument maker is known by three astrolabes'^^
and two globes (Nos. 25 and 29). On the earlier
globe, executed in 1070 H/AD 1659-1660 at the
request of a certain Shaykh "Abd al-Khaliq, the
maker signed himself as Muhammad Salih Tatah-
wT, while on the second globe, made in 1074 H/
AD 1663-1664 he signs as Muhammad Salih
TatawT. The spelling of Tatah-wT, which uses
quite unusual orthography, is probably an at-
tempt on his part to indicate the pronunciation
of the name, for with the second spelling one
might be inclined to pronounce it TatwT. It seems
unlikely that he was actually from Tatta in the
delta of the Indus river as some have sug-
gested,'^^ since the name of the town is written
with different characters and should more accu-
rately be transliterated Thattha.'^'^
  Both globes by TatawT seem to be quite precise
with full sets of constellation figures, though the
available photographs of his earlier globe show
little detail. Of special interest is the fact that the
second globe has the names of the constellations
and the signature written in both Arabic and in

Sanskrit (see Figure 18, which also clearly shows
a plug from the casting process). One might
speculate that this maker perhaps worked in the
KashmTr area, where at the end of the sixteenth
century ^AIT KashmTrT ibn LQqman may have
produced his apparently seamless metal globe.
KashmTr was a region where Sanskrit was the
language until replaced for official purposes by
Persian in the late fifteenth century, and conse-
quently might have been an area where a globe
in both Arabic and Sanskrit would have been
requested.
  One other globe maker of the middle seven-
teenth century was producing globes that are
apparently seamless cast metal. Lutf Allah ibn
"Abd al-Qadir al-Muhibb al-AsturlabT made an
undated globe (No. 67), probably working in
northwestern India, since that is the only area
which with certainty can be shown to have work-
shops employing the specialized technique. Little
is known of this maker, except that his father
produced four astrolabes, one of which is dated
1031 H/AD 1621-1622, which are generally con-
sidered Indo-Persian products,'®' Consequently,
if the father was an active instrument maker
around AD 1620, then it is likely that the son was
working around 1650,
  In the middle of the seventeenth century the
entire course of the Indus river was under
Mughal rule, as was Afghanistan to the west. An
unusual and attractive painted wood globe, (No.
36, illustrated in Figure 19) may well have been
made in Afghanistan. This globe, which appears
by the star positions to have been made in the
middle seventeenth century (although this is not
an example of high-precision work), represents a
slightly different iconography for the 48 constel-
lation figures from any other known celestial
globe or any known copy of "Abd al-Rahman al-
SQfi's Book of the Constellations of the Fixed Stars
or similar illustrated treatises that served as de-
sign books and guides to the placement of the
stars within and around the constellations. The
iconography and calligraphy of this globe suggest
that a now unknown instrument maker was work-
ing in Afghanistan or Turkistan about the same
time the Lahore family was excelling in their
  
NUMBER 46

45






FIGURE 18.—Globe No. 29, dated 1074 H/AD I663-I664, by Muhammad Salih TatawT.
London, private collection. (Photo: Alain Brieux)

production of cast globes.
  In addition to the Mughal empire of India,
two additional large empires dominated the cen-
tral lands of Islam in the sixteenth and seven-
teenth centuries: the Ottoman and the Safavid,
A hallmark of the Safavid rulers was their pa-
tronage of the arts. The Safavid empire was
centered in Persia, extending to the sub-Cauca-

sus and Khurasan south of the Oxus. During the
reign of Shah "Abbas I (AD 1587-1629) the Sa-
favid empire was at its strongest and the artistic
achievements of the court artisans and miniature
painters were the most conspicuous.
  One celestial globe (No. 63, illustrated in Fig-
ure 20) is known from an inscription on it to
have been made for Shah "^Abbas I in 1012 H/AD
  
46

SMITH.SONIAN .STUDIES IN  HISTORY  AND TECHNOLOGY






FiGi Rt   19.—Globe No.  36, unsigned and undated.  London, National  Maritime Mustum,
DejjartmeiU of Navigation. (Photo:  The National Mdiitime Museum, London)

16().S-16()4. Unfortunately the maker of this
globe is unknown. This particular globe and two
otlu-T Persian unsigned and undated globes (Nos.
H2 and 83) are particularK interesting as exam-
ples of a separate tradition in Islamicate celestial

globe design. Hiese globes do not have constel-
lation outlines depicted on them, but do have
some stars that are rather poorly positioned.
Rather, they have only the 12 zodiacal signs
represented  inside  medallions.  These  zodiacal

NUMBER 46

47






FiCLRK 20.—Globe  No.  63,  dated   1012  H/AD   1603-1604 and  made  lor Shah  "Abbas  I.
Chicago,  The Adler Plaiietariinu. (PIK^O: M. B. Smith)

signs are not used as guides to the positions of
the stars near the ecliptic and are not derived
from the illustrations of constellations fotnid in
astronomical texts, but indicate instead an alter-
nate Islamicate tradition of displaying the zodia-
cal signs as emblematic motifs rather than as
constellation diagrams. Taurus is a bull with a
hump on his back and a bell round his neck;
Libra a man sitting cross-legged with scales over
his shoulders like a yoke; etc. (see Chapter 2 for
further discussion of this iconography). All of
these globes are constructed with a seam, indi-
cating that they were made of two hemispheres.
Consequently, whatever workshop in Safavid
Persia produced these globes, it did not employ
the cire perdue process of making a seamless globe
used by the contemporaneous Mughal workshop.
Evidence of cultural and artistic exchanges
between Safavid Persia and Mughal India is seen

in the fact that the Mughal ruler Jahangir, a
contemporary of Shah "Abbas I, prided himself
on designing a series of coins emplo\ ing motifs
similar to those used in the zodiacal medallions
of these three Persian celestial globes.'^"Jahangir
lecorded in his memoirs for the thirteenth year
of his reign (AD 1617):
PifMoLislv to this, the lule ot coinage was that on one lace
of the iiKtal tlie\ stamped in\ name, and on the reverse the
name ol the place, and the month and year of the reign. .\t
this time it entered mv mind that in place of the month they
should substitute the ligure ol the constellation which be-
longed to tlicU month; ioi instance, in the month of Farwar-
cliii the figure ol a lani, and in L rclTbihisht the figure of a
hull. .SiinilarK, in each month that a coin was struck, the
ligiitf ot the constellation was to be on one face, as it the
sun were emerging Irom it. This usage is my own, and has
iu-\ci been practised until now."'*
   Dining the leign of SafT 1, the successor of
.SliiTh "Abbas I, a well-known astrolabe maker.
   
48

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Muhammad Zaman, working in Mashhad in the
northeastern part of the empire made a lovely
globe with the full set of constellation figures
(No. 16, illustrated in Figure 21) in the year
1050 H/AD 1640-1641 and another in 1054 H/
AD 1644-1645 (No. 17). This maker also pro-
duced a third globe that is small and undated
and has only the prominent stars indicated on it
(No. 64). All of these globes are metal and bear
an obvious seam, indicating that this Safavid
maker also did not produce seamless cast globes,
Muhammad Zaman is also known by seven astro-
labes,'^^ which span the years AD 1659-1677.
  Other astrolabe makers are known to have
been active in Safavid Persia, such as Muhammad
MuqTm and Muhammad MahdT, both of Yazd
southeast of Isfahan. The former made an astro-

labe for Shah "Abbas II in 1647-1648, engraved
by Fadl Allah al-SabzawarT under the supervision
of Muhammad Shafi, the astronomer of Jana-
bad,'*^^ while the latter made one in AD 1659-
1660 for Safi QulT Beg, an AmTr at the court of
Shah "Abbas 11.'^*^ They are known to have made
a large number of other astrolabes and to have
occasionally collaborated together on an instru-
ment. Unfortunately, no celestial globes are
known to exist by either of these Safavid metal
workers. It was more common in seventeenth-
century Persia than elsewhere for two or more
astrolabists to collaborate on an instrument as
the maker and engraver or decorator. From the
few known Safavid globes, however, it is impos-
sible to determine if the same custom was fol-
lowed in globe production.


FIGURE 21.—Globe No. 16, dated 1050 H/AD 1640-1641, by Muhammad Zaman. London,
The Victoria and Albert Museum, Department of Metalwork. (Photo: M. B. Smith)

NUMBER 46

49



  The father of Muhammad MahdT, Muhammad
AmTn, may well have been the astrolabe maker
whom Chevalier Jean Cardin met in his travels
through Persia as a French merchand royal at the
court of Shah "Abbas II. In his journals Cardin
provides a detailed description of the methods of
constructing astrolabes in Safavid Persia, but
makes no mention of celestial globes, going so
far even as to say that "the astrolabe is almost the
only astronomical instrument of Persians."'^^ It
is indeed curious that the very prolific and fa-
mous Safavid astrolabists at the end of the sev-
enteenth and beginning of the eighteenth cen-
turies, ^Abd al-A'^imma'^® and KhalTl Muhammad
(of Isfahan) are also not known to have made
celestial globes,
  A most unusual anonymous globe appears to
be a product of early seventeenth-century Sa-
favid Persia, This globe, No, 81, which has only
major stars and no constellation outlines, is
unique in that lines are engraved on it to
demonstrate the three methods of determining
the coordinates of a star or planet (see Chapter
2). These clearly correspond to a geographical
latitude of 32° north. This is the latitude of Yazd
as well as the latitude of Lahore, but since the
sphere is made in two hemispheres, we can con-
clude it is from Yazd rather than Lahore. Al-
though there is some irregularity in the engrav-
ing of the circles and graduations, the globe is a
fine example of metalwork, for the seam is not
visible on the outside. Moreover, no plugs are
observable, but when a probe is inserted a seam
can be felt running along the line of the ecliptic.
There are several visible cracks, which would
suggest that it was formed of two cast hemi-
spheres, put together with great expertise (see
Figure 22 for an illustration).
  While there are a considerable number of Is-
lamicate celestial globes from the Mughal and
Safavid empires, only one comes with certainty
from the contemporary Ottoman empire to the
west, and that one example is from Egypt, which
was a province rather than in the heart of the
Ottoman empire. By the middle of the sixteenth
century the Ottoman empire, centered in Ana-
tolia, extended eastward to the borders of the

Safavid empire and westward to Algeria. There
are only a few astrolabes known to have been
made in the Ottoman empire, very few when
compared with the numerous products of the
Safavid and Mughal realms,'®^ The Ottoman
rulers were not without a serious interest in
astronomy, however, for they built an observa-
tory in Istanbul in AD 1577 under the direction
of TaqT al-DTn Muhammad al-RashTd ibn Ma"ruf.
He had come to the Ottoman capital from Egypt
and at this short-lived observatory had under
him 15 astronomers and was himself the inventor
of some new astronomical instruments.''°
  The globe discussed above bearing the signa-
ture of Jamal al-DTn Muhammad al-HashimT al-
MakkT and dated 981 H/AD 1573-1574 was prob-
ably not actually made in Arabia despite the
maker's name. Quite certainly, however, at the
beginning of the eighteenth century a celestial
globe was produced in Egypt, at that time under
Ottoman rule. An astonomer, instrument maker,
and author by the name of Ridwan ibn "Abdallah,
called Ridwan EfendT al-FalakT (the astonomer),
lived in BQlaq (a district of Cairo), where he died
in 1710."' Egypt had been part of the Ottoman
empire since 1517, when Sultan SalTm I over-
threw the MamlQk dynasty in Egypt, From
March 21 to sometime in July or early August of
1701 (the second of Shawwal to Safar, 1113 H)
Ridwan worked on and completed a celestial
globe (No. 31) having a full set of constellation
figures with stars positioned according to the
tables compiled by Ulugh Beg's observatory, with
longitudes increased to correspond to his own
period. This globe, which is clearly made in
hemispheres and not cast in one piece, was made
for Hasan EfendT al-KhalwatT DamurdashT, at one
time RQznamijT of Cairo, The title RuznamijT was
given to the director of the astronomers of the
Qal"at Ruznamah (citadel of timekeeping) in
Cairo, He made calculations and observations
bearing on the determination of the calender,"^
  According to the Egyptian chronicler al-Ja-
bartT (died AD 1822), Ridwan ibn "Abdallah made
several instruments for Hasan EfendT in the years
1112-1113 H/AD 1700-1702, among them a
small number of brass celestial globes on which
  
50

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 22.—Globe No. 81, unsigned and undated. Oxford, Museum of the History of Science.
(Photo: M. B. Smith)

he engraved the visible stars and the constella-
tions, with their names in Arabic, and the ecliptic
latitude circles and meridians {dawdHr al-''urud
wa al-muyul), and then overlaid the globes with
gold and sold them for an increased price. Rid-
wan employed as an assistant Jamal al-DTn Yusuf,
a mamluk (servant) and kaldrji (keeper of the
cellar) to Hasan EfendT, who studied so diligently
that he went on to compose treatises on shadows
and lunar mansions and other technical subjects;
he died 30 years after Ridwan. Al-JabartT states
that both YQsuf and Ridwan's names were in-
scribed on a number of brass and gilded instru-
ments made for Hasan EfendT in 1701 and before
and after.'^^ Yusuf's name, however, does not
appear on the one globe known to be extant.

although Hasan EfendT is named as the patron
and traces of the gilding are evident. The one
extant globe also does not have meridians indi-
cated on it, although they are specifically men-
tioned as being part of Ridwan's design.
  All other workshops producing Islamicate ce-
lestial globes in the eighteenth century seem to
be from the eastern part of the Islamic world.
There is a seamless cast globe dated 1121 H/AD
1709-1710 dedicated to a now unknown patron,
which is quite certainly an Indo-Persian product
(No. 32 of the catalog). It is a highly inaccurate
globe with the stars arranged in a decorative
manner and the constellations crudely and some-
what fancifully delineated. There are also two
undated globes quite similar to this globe in
  
NUMBER 46

51



design and execution of both the sphere and
rings (Nos. 45-46), The former has an undeci-
pherable inscription around the base, and both
have holes bored on either side of the globe
about 25° from the equinoxes whose purpose is

unknown. The latter globe has some interesting
features to its somewhat whimsical iconography,
including an extra bird in the form of a falling
eagle near Cygnus (see Figure 23 for an overall
view). These globes are such extreme examples


FIGURE 23.—Globe No. 46, unsigned and undated. Bloomfield Hills, Michigan, Cranbrook
Institute of Science. (Photo: The Museum of the History of Science, Oxford)

52

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



of technically inaccurate and non-functional
globes and display so many features in common
that they suggest one workshop of perhaps even
more recent date than the rather garbled inscrip-
tion on No. 32 would indicate. They may be late
nineteenth- or early twentieth-century imitations
of Islamicate globes, perhaps produced by some-
one knowing very little Arabic and Persian, thus
accounting for the apparently nonsensical in-
scription on No. 45, but still versed in the tech-
niques of casting a seamless sphere, or having
access to spheres cast by another artisan.
  In north central India between AD 1728 and
1734 Maharajah Sawa^i Jai Singh II of JaipQr'^^
constructed five observatories at Delhi, Jaipur,
Benares, Mathura, and Ujjain. For the Delhi
observatory, which was completed first, Jai Singh
acquired the zarqaliyah style of universal astro-
labe, which had been made nearly half a century
earlier by Diya^ al-DTn Muhammad of the Lahore
family workshop and which is of unusually large
size (610 mm). Jai Singh tended in his design of
the Indian observatories to rely upon greatly
enlarged instruments to bring about greater pre-
cision rather than the smaller portable astronom-
ical instruments.'^^ The two astrolabes which he
himself made"^ are of enormous proportions,
each having a diameter of 2100 mm.
  Jai Singh also directed the preparation of a set
of astronomical tables named Zij Muhammad
Shdhi after the Mughal ruler Muhammad Shah
(AD 1719-1748) who appointed him governor of
the state of Agra and later Malwa in north central
India. These tables were intended to update and
improve those made by Ulugh Beg in the fif-
teenth century. In a manuscript of these tables
at JaipQr it is said that the ecliptic coordinates
(celestial longitude and latitude) used by Ulugh
Beg were converted to declination and right
ascension (equatorial coordinates) by taking them
from a celestial globe."' That is, presumably a
rather large and well-made globe based on the
positions of Ulugh Beg's catalogue was used to
find the coordinates of declination and right
ascension (see Chapter 2 on the uses of celestial
globes), thus avoiding the involved calculations
requiring spherical trigonometry. Other than
this one mention, however, celestial globes are

not described as being among the astronomical
instruments employed in the Indian observato-
ries. Although Jai Singh himself was a Hindu,
the work of the observatories represents for the
most part the Islamicate tradition, just as the
language employed was Persian, influenced by
some European contacts and works such as the
tables of Philippe de La Hire, which were trans-
lated into Persian for comparison with the Zij
Muhammad Shdhi.
  By a century later, however, these Indian ob-
servatories were neglected and in ruins, "^ Yet in
1842 a Hindu metalworker named Lalah Bal-
hQmal LahQrT (that is, from Lahore) made an
Islamicate celestial globe (No. 33) for a Sikh
patron named Nihal Singh Sahib Bahadur Ah-
luwalTyah, From about AD 1750 the Sikhs had
taken over the Punjab, and during the following
30 years had organized it into 12 loosely-knit
states or misls, one of which was Kapurthalah,
ruled by the AhluwalTyah family. The capital of
this misl was the town of Kapurthalah where this
globe was placed, according to the lengthy Per-
sian inscription on the sphere. Thus it is certain
that the globe was constructed for Nihal Singh
who ruled Kapurthalah, only a few miles south-
east of Lahore, from October of 1837 until his
death on 13 September 1852, Nihal Singh had
intermittent quarrels and disagreements with the
Maharajahs of Lahore, Ranjit Singh and his suc-
cessors, so that during the Sikh war of 1845 Nihal
Singh failed to align himself clearly with either
the Sikh or British side, for which reason some
of his lands were taken by the British. When the
second Sikh war occurred he gave immediate
assistance to the British, after which he was then
made a Rajah and apparently ruled rather pop-
ularly over his subjects until his death. "^
   Lalah Balhumal LahurT, who is also known by
at least five astrolabes,'^^ cast the globe in one
piece, for no seam is visible and the plugs arising
from the cire perdue process can easily be seen.
This fact, in addition to his nisbah (the part of
the name indicating place of origin) being La-
hQrT, strongly suggests that the process of casting
globes in which the sixteenth- and seventeenth-
century Lahore workshop specialized continued
to be practiced for more than 150 years after
   
NUMBER 46

53



Diya^ al-DTn Muhammad, Many of the apparently
seamless cast globes may have been produced by
workshops in the Lahore area over a period of
250 years. Since we know of no written treatises
on the technical construction of a seamless globe
by cire perdue it is likely that a system of appren-

tices was required to maintain sufficient skill in
the technique to produce a globe of first rate
quality such as the one made in 1842. Lalah
BalhQmal exhibits considerable skill in the exe-
cution of this globe (see Figure 24), on which he
placed a long Persian inscription describing the


FIGURE 24.—Globe No. 33, dated 1258 H/AD 1842, by Lalah Balhumal LahurT; present
location unknown. (Photo: From the Files of the Whipple Museum of the History of Science,
Cambridge)

54

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



globe, the patron, and the fee he received as
maker. Though he was probably a Hindu and his
patron a Sikh, the globe is a typical Islamicate
celestial globe with the standard set of 48 con-
stellation outlines and a stated 1022 stars, labeled
in Arabic, thus maintaining the basic design evi-
dent in the earliest of the Islamic globes made
762 years earlier in Spain. A second celestial
globe was made by Lalah BalhQmal in the same
year (see No. 127 in the Catalog, Chapter 6).
  There is a striking similarity between the de-
sign of these globes made by Lalah BalhQmal
LahQrT in 1842 and an anonymous celestial globe
with Sanskrit inscriptions (No. 54, illustrated in
Figure 25). The similarity extends to the stands
and rings, methods of construction, and star po-
sitions, leading to the inevitable conclusion that

both globes were either produced by Lalah Bal-
hQmal or someone trained in his workshop and
familiar with his technique and design. Although
the constellation names and the numerals are
written in DevanagarT characters, it is completely
within the Islamicate tradition in design and ex-
ecution.'^'
  A third example, an anonymous globe with
only the major stars (No. 90), can also with some
certainty be attributed to the Lalah Balhumal
LahQrT workshop. It displays the same precision
of execution, the same basic method of construc-
tion, the same star positions (though only 32 of
the major stars are represented on the globe),
and the identical distinctive arrangement of me-
ridian circles. On all four globes (Nos. 33, 54,
90, and 127) there are a set of six great circles
  

FK;URE 25.—Globe No. 54, unsigned and undated; engraved in Sanskrit, Devanagan charac-
ters. Columbia University, David Eugene Smith Collection. (Photo: Columbia University)

NUMBER 46

55



serving as meridians, which intersect the equator
at 30° intervals; they do not begin at the vernal
equinox but rather are shifted 6° westward, so
there is no meridian representing the equinoctial
colure. For these reasons this anonymous globe
can also be assigned to the nineteenth-century

workshop of instrument makers in Lahore (see
Figure 26),
  A fifth globe (No, 53) has the same unusual
arrangement of meridian circles; however, from
the available photograph it is very inaccurately
constructed, and its lack of precision stands in
  

FIGURE 26.—Globe No. 90, unsigned and undated. Rockford, Illinois, The Time Museum.
Arrows indicate plugs. (Photo: Courtesy of The Time Museum, Rockford, Illinois)

56

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



considerable contrast to the fine execution of the
other four globes. Neither its size nor the method
of construction is known, but its general appear-
ance suggests that it is a later copy of a globe of
the BalhQmal workshop design, executed by a
metal worker who was not a professional instru-
ment maker.
  At about the same time as the instrument
maker Lalah BalhQmal LahurT was working in
the Punjab, another metalworker was making a
celestial globe (No, 92) in Transoxiana for Mu-
hammad "AIT Khan, who ruled as the eighth
Khan the area of Khokand east of Bukhara AD
1822-1842,'^2 The maker did not sign the globe,
which was produced in 1241 H/AD 1825-1826,
It has no stars on it and is made of two metal
hemispheres.
  Of the globes having no stars on them, all but
five of those inspected have a seam. The seam is
not always visible (for example, on Nos. 100,
103, and 105), but there are no plugs visible, and
when a probe is inserted inside, a clear seam or
ridge can be felt running around the sphere.
Since the majority of globes without stars have
seams, this would suggest that most of them were
made outside of northwestern India, which ap-
pears to have specialized in seamless globes cast
by the cire perdue process—perhaps in Safavid
Persia. Most of these globes (Nos. 95 through
110) seem to be related stylistically as well, sug-
gesting that they were produced in the same
workshop. No. 97 has a patron's name, MTrza
Muhammad TaqT Khatarat, who is said to be a
mujtahad, or a ShT"ite authority on religious
jurisprudence. Since five of them appear to be
seamless and exhibit plugs from the casting proc-
ess, some Indo-Persian workshops in northwest
India may well have produced them (an example
of a seamless globe lacking stars can be seen in
Figure 34),
  Of the globes that have no stars or constella-
tions, only two are dated. In addition to the one
made in Transoxiana for Muhammad "AIT Khan,
there is one globe (No. 91) dated slightly earlier,
1238 H/AD 1822-1823. It has no visible seam
but neither has it any detectable plugs, and ex-
amination of the interior was not possible; there-

fore, it cannot at this point be classified as to the
method of construction. The maker's name is
quite difficult to read with certainty, but seems
to be Muhammad Sam^T Tan, and the calligra-
phy and engraving style would suggest a Persian
rather than Indo-Persian origin. Since there are
no star positions indicated on this class of globe,
one cannot estimate an epoch for which the
undated ones were designed. Most of the extant
globes of this class are somewhat carelessly exe-
cuted with irregular graduations and uneven cir-
cles. There are, however, some examples of quite
precisely made globes, such as No, 93 (illustrated
in Figure 27), which is a papier mache globe in
quite poor condition, and No, 98,
  A seamless cast globe made in 1241 H/AD
1825-1826 by Muhammad KarTm might possibly
have been made outside of India (No. 74). Noth-
ing is known of this maker,'^^ but his name, the
use of the Yazdijird era on the globe,'^^ and the
calligraphy suggest that he was from Persia
rather than India. The well-made seamless globe
has about 80 stars and no constellation outlines.
If it was indeed produced in Persia rather than
India, this would be evidence of an exchange of
ideas and techniques between the artisans of
Persia and those of northwestern India,
  In the nineteenth century there were active
three celestial globe makers whose styles differed
considerably from each other and from the Indo-
Persian products of the same period. A very fine
globe exhibiting remarkable precision of design
and execution was made by Muhammad Ashraf
TQqadT Zadah in 1215 H/AD 1800-1801, The
small globe (No. 72) is made of painted wood, or
possibly ivory, and the stand and rings are con-
temporary, for the maker has signed and dated
them as well. The globe has no constellations but
about 50 stars marked by dots. There are a set
of meridian circles crossing the equator at 30°
intervals, but no ecliptic latitude circles (possibly
indicating European influence in design). It also
has drawn in very fine ink a complete set of
parallels at 2° intervals north and south of the
equator, and another complete set of 2 ° intervals
parallel to the ecliptic. It is likely that the maker
of this globe of unusual design was from Persia
  
NUMBER 46

57






FIGURE 27.—Globe No. 93, unsigned and undated. Oxford, Museum of the History of Science.
The covering plaster on this globe has deteriorated in some places. (Photo: M. B. Smith)

or Afghanistan, though we have no other prod-
ucts by him and no information concerning his
work.
  The design, precision, and calligraphy of the
anonymous globe (No. 76) are strikingly similar
to that signed by Muhammad Ashraf TuqadT
Zadah. This globe, illustrated in Figure 28, of
painted paper on wood has both meridians and
ecliptic latitude circles with a most unusual net-
work of half-great circles originating at the inter-
section of the celestial equator and the solstitial
colure at the winter solstice and extending to the
equinoctial colure at 5° intervals, with this hem-
isphere also covered by a series of parallels at 1 °
intervals. The purpose of the network—unique
among extant globes—is unknown. The longi-

tudes of the stars indicate an epoch of the end of
the eighteenth century.
  Another quite different but equally unusual
globe was finished by Muhammad "AIT al-HusaynT
on 23 Jumada I 1221 H/9 August AD 1806. This
is a painted and lacquered papier mache globe
(No. 73) of fairly large size. It is a technically
inaccurate and nonfunctional globe with stars
indicated by disproportionately large gilt dots on
the dark rich brown ground. An incongruous
element on such an inaccurate globe is the state-
ment inscribed at each solstice that the obliquity
of the ecliptic is equal to 23°30'17". Only the
12 zodiacal constellations are represented picto-
rially in red, black, and gilt paints. The zodiacal
figures are not actually constellation outlines, for
  
58

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 28.—Globe No. 76, unsigned and undated. Damascus, Musee Nationale. (Photo: The
History of Science Museum, Oxford)

no stars are indicated in them. They are some-
what reminiscent, though by no means identical
to, those figures of the Persian iconographic
tradition evident on the globe made in the six-
teenth century for Shah "Abbas I (No. 63, illus-
trated in Figure 20). There are, however, also
similarities with the zodiacal figures which are
depicted, along with the other constellation
forms, on an anonymous papier mache globe (No.
56; see Figure 29). On the latter globe, the
constellation figures are not outlines of asterisms,
for there are no stars on the globe, and they are
depicted with an iconography radically different
from that usually found on Islamicate celestial
globes, which is associated with the al-SQfi man-

uscript tradition and forms the basis of most
constellation design on scientific instruments.
  Late in the nineteenth century the only globe
we know to come from Ottoman Turkey (No.
75) was made by "a student of the Normal
School" Husayn Husnu KanqarT on 3 RabT'' I
1299 H/23 January AD 1882. This globe, which
is in two metal hemispheres, has meridians and
ecliptic latitude circles, probably indicating some
European influence. The graduations are labeled
in standard rather than abjad numerals, a very
unusual feature. It is unique among all the globes
in having the names of the zodiacal figures writ-
ten with the top of the writing toward the north
(see Chapter 2 for further discussion of this char-
  
Nl'MBF.R  4tS

.'')9


PIGURK 29.—Gicjbe  No.  36,  unsigned and  undated.  London,  National   Maritime  Museum,
Department of Na\igation. (Photo: The Nation,il Maritime Museum, London)
acteristic of Islamicate globes); this reversal may	This Turkish globe is the most recent Islamicate
again show European influence. The star posi-	globe included in this study.
tions are quite imprecise and varied, although	It is evident that the Indo-Persian metal-vvork-
the graduations and circles are well executed.	ers and astrolabists of northwestern India per-

60

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



fected the difficult process of casting a seamless
hollow metal globe by the cire perdue process,
practiced this technique, and taught it to others
from the late sixteenth century to the middle of
the nineteenth century. It is impossible to say
who first developed this extremely complex
method of construction. Two examples seem to
predate the sixteenth century: that of Muham-
mad ibn MahmQd al-TabarT (No. 6), dated 684
H/AD 1285-1286, and that of "Abd al-Rahman
ibn Burhan al-Mawsili (No. 60), dated 718 H/AD
1318-1319. This implies the existence of a thir-
teenth-century workshop in Persia and an early
fourteenth-century one in Mesopotamia which
could produce such spheres. Yet there are prob-
lems with the validity of these globes. Even if
they were actually executed at the time stated on
the globes, it is strange that no other examples
of the technique occur before the end of the
sixteenth century. Nonetheless the two early ex-
amples display complete mastery and much ex-
perience with the technique.
  During the eight hundred years in which the
extant Islamicate celestial globes were made,
from regions as diverse and distant as southern
Spain and northern India, there was no change
in the basic design of the globe, except for the
apparently late and Persian development of a
type of globe having no stars at all. While the
depiction of the constellation figures could be
adapted (within certain limits) to the local artistic
conventions, the number and form of the con-

stellations and the stars remained constant. The
newly recorded stars and constellations of the
southern hemisphere, resulting from the Euro-
pean explorations of the sixteenth century,'*^
were never represented on an Islamicate globe.
The number of the constellations and stars on
Islamicate globes, including the two known San-
skrit globes that are also to be considered part of
this tradition, is always derived from al-SQfi's or
Ulugh Beg's star catalogs or from the earlier
Ptolemaic one, which had only minor differences
among them. On globes having only a selection
of the most prominent stars, the choice of which
stars to use depended on the designer and could
vary, just as they did on astrolabes. Nonetheless
the basic design remained unchanged. In the
matter of construction, however, considerable
experimentation and development seems to have
occurred. During these eight hundred years
globes were made in metal hemispheres, made
of painted wood or papier mache over a wood
core, or were cast as a hollow seamless metal
globe by the cire perdue process. This last
method, which is not mentioned in any written
discussion of celestial globes, either European or
Islamicate, appears to be a distinguishing feature
of Indo-Persian celestial globe making: that is,
although the technique may possibly have devel-
oped earlier and elsewhere, it became the hall-
mark of the workshops located in the Punjab and
KashmTr areas of northwestern India.

2.    The Nature, Use, and Construction of Islamicate Celestial Globes
That inverted Bowl they call
  the Sky
Whereunder crawling coop'd we
  live and die,
Lift not your hands to It for
  help—for It
Rolls impotently on as You
or I.
Edward Fitzgerald
The Rubdiydt of Omar Khayyam

  Today there are known to be 126 extant Islam-
icate celestial globes spanning a period from the
eleventh to the nineteenth centuries and origi-
nating from regions as far west as Spain and as
far east as the Punjab and western Himalayas.
(See the Catalog, Chapter 6, for detailed descrip-
tions; an Addendum to the Catalog describes six
additional globes not included in the analyses.)
All the globes represent a tradition of instrument
design which can be traced back to the Greek
and Roman worlds. The language used in inscrib-
ing the majority of the globes is Arabic, although
Persian, Indo-Persian, Ottoman Turkish, and
even Sanskrit occur on some globes. The con-
stellation and star names and other astronomical
nomenclature is nearly always in Arabic, except
for the two globes bearing Sanskrit inscriptions
in DevanagarT characters. The other languages,
with only a few exceptions, occur in the imprint
inscriptions—that is, those giving the maker,
date, and sometimes the patron. These extant
Islamicate globes can be grouped into three basic
types that we shall now describe in detail.
  The first type of globe, Class A, displays con-
stellation outlines and depicts approximately the
full set of those visible stars listed in star catalogs,
such as those by Ptolemy, "Abd al-Rahman al-
Sufi, and Ulugh Beg.' These globes have the full
set of 48 constellation figures recognized from
Hellenistic times, but their appearance (e.g,, the
clothing) was modified to conform to the artistic
conventions of the particular period and location
where the globe was produced (see Figures 6, 9,

13, and 15). Of the 58 globes placed in this
category, 25 are unsigned and undated. Three
of the undated globes should perhaps form yet a
fourth category of globes, for while they have
constellation figures, they lack stars (Nos, 55, 56,
and 58, one of which is illustrated in Figure 29),
For the sake of convenience, however, they have
been categorized as Class A globes. The 33 dated
globes range in date from AD 1080 to 1842.
  The second type of globe. Class B, has indi-
cated on it only the major stars, varying in num-
ber from 20 to 150, with no constellation outlines
(see Figures 7, 10, 22, 26, and 28). The choice
of which stars were included on a globe of this
class seems to have depended upon the maker,
though those termed "astrolabe stars" in the star
catalogs were nearly always included. There are
32 Class B globes in the catalog, of which ap-
proximately half are unsigned and undated. The
remaining 17 globes date from AD 1140 to 1882.
Among the Class B globes there have been in-
cluded four globes that have zodiacal figures as
well as the major stars (Nos. 63, 73, 82, and 83;
No. 63 is illustrated in Figure 20).
  The third type of globe, Class C, has neither
stars nor constellations but displays those great
circles^ that are common to all three types of
globes as well as some lesser circles (see Figures
27 and 34). In this class there are 34 globes, only
two of which are dated, both from the first
quarter of the nineteenth century.
  While there are a few Islamicate globes made
of painted wood or papier mache on wooden cores
  
61

62

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



(see Figures 19, 27, 28, and 29), the majority are
metallic. More shall be said later of their methods
of construction.
  The three classes described above do not com-
pletely exhaust the 126 globes described in the
catalog in Chapter 6. That catalog includes at
the end two globes for which it was impossible to
obtain sufficient information to classify them. An
addendum lists six additional globes.
  The celestial globe represents the sphere of
fixed stars (without the planets or "wandering
stars" as they were sometimes called) as seen from
outside the sphere rather than from the earth at
its center.^ A celestial globe in Arabic is called
al-kurah (the sphere) or occasionally al-kurah
dhdt al-kursi (the sphere possessing a stand) or
kurat al-samd" (the globe of the sky).'*
  On every extant globe the great circle of the
ecliptic (the apparent annual path of the sun in
the heavens) is clearly marked. The poles of the
ecliptic are prominently indicated, for through
them pass six great circles that intersect the eclip-
tic at right angles and divide it into 12 sections
of 30° each (see Figure 30). These great circles
are not used today in astronomy and so do not
have a common name, but can be termed ecliptic
latitude circles. They are circles along which the
celestial latitude is measured."^ It should be noted
that meridians (great circles at right angles to the
celestial equator) are rarely found on Islamicate
celestial globes.*" The ecliptic latitude circles were
used on Islamicate globes because in the Arabic-
speaking world, as in the earlier Greek world,
the coordinates of a star were most commonly
measured by the celestial longitude or the dis-
tance measured from the vernal equinox east-
ward along the ecliptic, and by the celestial lati-
tude measured north or south from the ecliptic
along a great circle at right angles to it.^
  The Arabic term occurring on celestial globes
for such circles at right angles to the ecliptic is
dd^irat al-^ar.d, sometimes written dd^irat ^ard,
(latitude circle; plural, dawd^ir al-''urud). This is
the circle along which the celestial latitude is
measured, to be distinguished from the term for
parallels or circles of latitude, maddrdt al-^urud
or maddrdt "urud, which are circles parallel to

the ecliptic.*^ In the discussion that follows and in
the catalog of globes I use the term ecliptic
latitude circle when referring to such great cir-
cles passing through the ecliptic poles. This re-
flects the Arabic terminology, since there is no
commonly used term for them in modern astro-
nomical nomenclature.
  On nearly all Islamicate globes the names of
the 12 zodiacal houses are inscribed along the
ecliptic, each name filling the appropriate 30°
interval.^ For a celestial globe to function
properly for northern geographical latitudes it
must be oriented with the north pole up. On
Islamicate globes the zodiacal names are written
so that they appear to be upside down when the
globe is set in this position,'" The reason they
were always written in this fashion lies in the way
the Arabic language is written and read. For
Latin, Greek, and modern European languages,
when the globe is oriented with the north pole
up, the sequential order of the zodiacal houses
as they appear along the ecliptic has the same
sinistrodextral direction as the writing. The na-
ture of the Arabic alphabet, however, required
that the names of the zodiacal houses be written
in what seems to us an upside-down manner in
order to maintain the flow of the order.
  To give an example using the Latin alphabet,
if the sequential order of the houses were from
right to left along a line, then when labeling the
houses in the Latin alphabet, which is written in
the opposite direction left to right, the words
must be inverted.
iNiiMao saaavx S3mv
<—	
GEMINI   TAURUS   ARIES
Only when the labeling is turned upside down
will the correct order be maintained. Since the
actual order of the houses on a globe is left to
right, and Arabic is written right to left, a similar
situation results in which the names appear to be
upside down. The same is true of the abjad
numerals with which the ecliptic and equator are
usually labeled (numerals that are letters of the
alphabet assigned numerical values). The fact
that the names of the zodiacal houses and the
numerals appear to be right side up only when

NUMBER 46

63


North   celestial
(equatorial)    pol
FIGURE 30.—Basic design of extant Islamicate celestial globes, with attached rotatable meridian
ring.
the globe is oriented with the south pole up does	globe be mounted with the north pole up for use
not mean that Islamicate globes were set in the	at northern geographical latitudes. And there
stands with the south pole up. On the contrary,	would in actuality have been almost no need to
it is specifically stated in all the literature and	use a globe for southern geographical latitudes,
required by labeling on extant rings that the	although some rings do have holes for mounting

64

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



the globe for arbitrarily chosen southern geo-
graphical latitudes. Having the labeling in an
apparently upside-down position probably did
not present as large a problem to the user as we
might think. Most globes would have been placed
on a table lower than eye level and the user
would probably have stood over it. The Arabic
names of the zodiacal signs that appear on the
globes are as follows: al-hamal (Aries), al-thawr
(Taurus), al-jawzd^ {Gem'im), al-saratdn (Cancer),
al-asad (Leo), al-sunbulah (Virgo), al-mizdn (Li-
bra), al-^aqrab (Scorpio), al-qaws (Sagittarius), al-
jadi (Capricorn), al-dalw (Aquarius), and al-hut
(Pisces).
  A few Class A globes also have the names of
the 28 lunar mansions" written along the eclip-
tic, usually at intervals of about 12'/2°, desig-
nated by a small dash intersecting the ecliptic.
The fact that the zodiacal houses are along the
ecliptic gives rise to the term used for the ecliptic
on Islamicate globes, that is, dd^irat mintaqat al-
buruj (the circle of the belt of the houses), while
the two poles of the ecliptic are labeled qutb
mintaqat al-buruj shamdli {jandbi) (the north [or
south] pole of the belt of the houses), or some-
times simply qutb buruj shamdli {janubi) (the
north [or south] pole of the houses).'^
  At an angle of approximately 23V2° to the
ecliptic,'^ the great circle of the celestial equator
or equinoctial, dd^irat mu^addil al-nahdr (the cir-
cle of the equality of day), is indicated crossing
the ecliptic between the Houses of Virgo and
Libra and between the Houses of Aries and
Pisces. At the celestial poles—that is, the poles
of the celestial equator that are immediately over
the terrestrial poles, the two points in the heavens
toward which the earth's axis is directed—there
are nearly always two holes through which the
globe was mounted in a ring. These poles, when
labeled, are called qutb mu^addil al-nahdr shamdli
{janubi) (north [or south] pole of the celestial
equator), or sometimes qutb shamdli {janubi)
(north [or south] pole). On one globe (No. 113)
the label reads qutb "^dlam shamdli (north pole of
the universe). The ecliptic poles, where the eclip-
tic latitude circles converge, occasionally also had
holes.''* In a truely functional globe these would

not have been used for mounting the globe on
its axis in a meridian ring, but rather would be
useful during the construction and finishing of
the globe and for the insertion of a graduated
quadrant for measuring coordinates.
  On nearly all the globes the great circles of the
celestial equator and ecliptic are graduated, most
frequently by single degrees with every fifth de-
gree indicated by a longer line and labeled. The
numerals labeling the graduations are written in
abjad notation, that is, with certain designated
letters of the alphabet used instead of the stan-
dard numerals,'^ There are also globes gradu-
ated by single degrees with every sixth labeled,
and some by single degrees with every tenth
labeled, and with every third degree indicated
by a longer line. Occasionally globes are gradu-
ated by larger intervals, such as 2° intervals with
no labeling, 2° intervals with every sixth degree
labeled, 2° intervals with every tenth labeled, or
by 5° intervals with each one labeled, or by 6°,
or even 10° segments with each labeled. The
Smithsonian globe, which is the focus of the
present study (No. 38 of the catalog) appears to
be one of only two globes graduated by half-
degree intervals, (See the tables in Chapter 7 for
details concerning the incidence of these and
other characteristics of the globes.)
  Those globes graduated by single degrees
range in diameter from 55 mm to 330 mm, with
a median size of 150 mm. As one might expect,
the globes whose smallest unit of graduation is
two or more degrees are generally smaller globes
than those graduated by single degrees. The
globes graduated by 2° intervals, with two ex-
ceptions, vary in size from 60 mm to 100 mm in
diameter, with a median size of 70 mm, while
the two globes graduated by 6° intervals are 50
mm and 60 mm in diameter, and the one globe
using 10° intervals is 97 mm in diameter. Rather
surprisingly, the four globes graduated by 5°
intervals are fairly large, two having a diameter
of 115 mm and two of 150 mm, just as there are
two globes whose smallest unit is two degrees,
which have diameters of 200 and 203 mm,'^ The
unusually large size of these six globes is no doubt
due to the fact that they are all Class A globes.
  
NUMBER 46

65



which tend to be larger than the other two classes
of globes, obviously because they must accom-
modate over 1000 stars and 48 figures. Why the
makers chose such a scale of graduation for these
six globes is not known. The estimated diameter
of one of the two globes using one-half-degree
intervals (No. 53) would make it the largest of
all the extant globes by a considerable margin,
while the diameter of the Smithsonian Institution
globe, the only other globe graduated by such
an interval, is the sixth largest of the extant
globes whose diameters are known.
  Several extant globes have graduations of the
ecliptic and equator but have no numerical la-
beling, while 14 globes have no graduations at
all, but rather simple lines representing the ce-
lestial equator and ecliptic. With only one excep-
tion, the numbers marking the degrees along the
ecliptics repeat in 30° intervals: that is, starting
at the vernal equinox, one finds along the ecliptic
the numbers 5, 10, 15, 20, 25, 30, 5, 10, 15, 20,
, . . or whatever series would be appropriate,
such as 6, 12, 18, 24, 30, in the case of those
graduated by 2 with every sixth labeled. The one
exception to the repetition of 30° intervals on
the ecliptic is again No. 38, the Smithsonian
Institution globe. Its intervals along the ecliptic
are numbered continuously from 5 through 360
beginning at the vernal equinox.'^
  Greater variety is to be found in the gradua-
tions of the equator. For the majority of the
globes the graduations along the equator are
numbered continuously from the vernal equi-
nox. Some, however, are numbered in three
consecutive segments of 100° each and a final
segment of 60°. Some are labeled in two seg-
ments of 180°, and on two, the equator repeats
every 30° as does the ecliptic, while on three the
equator is numbered in four segments of 90°
each. Two globes, both nineteenth-century west-
ern Indian products (No, 33, illustrated in Figure
24, and No, 53), number each 6° segment of the
celestial equator with a single numeral, 1, 2, 3,
. , , 59, 60, beginning with the vernal equinox.
  The two points where the ecliptic crosses the
celestial equator, the equinoxes, are not labeled
as frequently as are those great circles them-

selves, but when their names are inscribed the
terms used are nuqtat i^tiddl kharifi (point of the
autumnal equinox) and nuqtat i^tiddl rabi'^i (the
point of the vernal equinox). Similarly, the sol-
stices are seldom labeled, but on the eight globes
where they are labeled, the terms nuqtat inqildb
sayfi {shatwi) (point of the summer [or winter]
turning) and inqildb shams (turning of the sun)
occur.
  The solstitial colure is the great circle that
passes through the northernmost point of the
ecliptic, the summer solstice, and the southern-
most point of the ecliptic, the winter solstice.
This circle also passes through the ecliptic poles
as well as the celestial (equatorial) poles, and thus
is one of the six ecliptic latitude circles found on
all Islamicate globes. The solstitial colure is rarely
labeled, however. On only six globes does it bear
a label, dd^irah marrah bi-l-aqtdb al-arba^ah (the
circle passing through the four poles).
  The point on the ecliptic that is the greatest
distance from the equator, measured along the
solstitial colure, is frequently labeled on globes.
On 25 globes, all but one of Class B or C, the
words mayl kulli are written along the segment
of the solstitial colure between the equator and
ecliptic. Thus, the term occurs twice on such a
globe on opposite sides near the solstices. The
term mayl kulli means "complete obliquity (or
declination)"—that is, the greatest obliquity or
distance of a point on the ecliptic from the equa-
tor.'*^ When discussing the globes in the Catalog
(Chapter 6), I have used the term mayl kulli,
rather than an English equivalent, to indicate
when these two labels are found on a globe. On
only three globes is the actual value of the obliq-
uity specified.'^
  On many globes a great circle is drawn
through the equinoxes and the celestial poles,
and perpendicular to the solstitial colure. This
great circle, the equinoctial colure, is nearly al-
ways indicated on those globes having no stars
(Class C) and on approximately one-third of the
globes having only the prominent stars and no
constellation outlines (Class B), but it occurs on
only four of the globes exhibiting the constella-
tion figures (Class A), The equinoctial colure is
  
66

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



seldom labeled, but on at least two globes it is
termed dd^irat al-mayl or dd^irat mayl (declination
circle).^"
  The name of the equinoctial colure in Arabic
reflects one of the two terms used for the meas-
urement of coordinates of a celestial body in a
celestial equatorial system. Mayl (declination) is
the star s angular distance in degrees north or
south of the celestial equator measured along a
great circle passing through the equatorial poles
and the star in question. This circle is today called
a meridian or hour circle. On Islamicate celestial
globes any great circle passing through the celes-
tial poles is called a dd^irat al-mayl (plural dawd^ir
muyul; declination circle). This is a circle along
which the declination is measured, as distin-
guished from the maddrdt muyul (circles [paral-
lels] of declination), which are circles parallel
with the celestial equator.*^'
  The Arabic term ba^d (distance) refers to the
measurement along the celestial equator from
the vernal equinox to the point of intersection of
the meridian passing through the star (right as-
cension in modern terminology). This system of
celestial equatorial coordinates was known in the
Islamic world, as well as to earlier Greek astron-
omers. Islamic astronomers made use of such
coordinates as early as the ninth century.^"^ An-
other method, which made use of the ecliptic was
employed, as was a third system of coordinates
based on the altitude and zenith, which will be
described in connection with the rings attached
to the globes (see Figure 31).
  From an examination of Islamicate celestial
globes we can confirm the preference for the
ecliptic coordinate system in which the celestial
latitude (called "ard) was measured north or
south from the ecliptic along a great circle pass-
ing through the ecliptic poles and the object.
The celestial longitude {tul) was measured from
the vernal equinox eastward along the ecliptic.
It is in terms of the ecliptic coordinates that the
positions of stars are given in all the star catalogs
from Ptolemy s in the second century through
that of Ulugh Beg in the sixteenth century.
  In addition to the great circles, several lesser
circles often appear on extant globes, though no

mention is made of such circles in the Islamicate
literature surveyed in Chapter 1. The lesser cir-
cles are called maddrdt (singular, maddr) to dis-
tinguish them from great circles called dawd^ir
(singular, dd^irah) both terms are from the same
root and mean "circle." One globe (No, 94),
however, employs the term dd^irah saghirah
(small circle) for the lesser circles.
  The lesser circles of the tropics, placed about
23'/2° north and south of the celestial equator,
as well as the equatorial polar circles, are fre-
quently found on globes of the first two classes,
and these four lesser circles appear on all the
globes of the third class that were examined.
There are two equatorial polar circles, one north
and one south, with a celestial equatorial pole as
center and passing through the corresponding
ecliptic pole about 23V2° distant. These polar
circles are labeled maddr qutb mu'^addil al-nahdr
shamdli {janubi) (circle of the north [or south]
pole of the equator). The Tropic of Capricorn is
usually called maddr ra^s al-jadi {the circle of the
start [head] of Capricorn), but occasionally the
phrase maddr awwal jadi (circle of the beginning
of Capricorn) is found. The Tropic of Cancer is
termed maddr ra^s al-saratdn (circle of the start
[head] of Cancer) and sometimes maddr awwal
saratdn (circle of the beginning of Cancer). One
globe (No. 94) uses for the Tropics an expression
of mixed Persian and Arabic, dd^irah saghirah az
maddr ra^s saratdn {jadi) (small circle belonging
to the circle of the start of Cancer [or Capri-
corn]), and similarly for the polar circle dd^irah
saghirah az maddr qutb (small circle belonging to
the circle of the pole).
  Several globes of all three classes have addi-
tional circles parallel to the equator at about 12°
and 20° on either side of the celestial equator.
The two lesser circles to the north are occasion-
ally labeled maddr awwal thawr wa sunbulah (the
circle of the beginning of Taurus and Virgo) at
12° and maddr awwal jawzd"" wa asad (the circle
of the beginning of Gemini and Leo) at 20°. To
the south are the circle of the beginning of
Scorpio and Pisces, and the circle of the begin-
ning of Aquarius and Sagittarius, In a few in-
stances only one of the sets of additional lesser
  
NUMBER 46

67



No rth   Celestial
Po I e
Zenith

North Eelipt ic
Pole

FIGURE 31,—Schematic drawing of Islamicate celestial globe set into meridian and horizon
ring assembly, with indications of the three coordinate systems.

68

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



circles is indicated, usually at 12°. On a globe
(No. 63, illustrated in Figure 20) made in 1603
for Shah "Abbas I, the lesser circles at 12° to
either side of the equator are labeled maddrdt
muyul (circles of declination), the general term
for lesser circles parallel to the equator. This
globe does not have the circles at 20°, but does
have a set of comparable lesser circles at 12° on
either side of and parallel to the ecliptic.
  The ecliptic equivalent of the Tropics (circles
parallel to the ecliptic and about 23V2° on either
side of it) were commonly drawn on those globes
having no stars, and occasionally indicated on
those of Class B, but seem never to have been
placed on Class A globes. The ecliptic tropic
circles do not appear to bear a particular term
on the globes, but, when labeled, have only the
generic term for all circles parallel to the ecliptic,
maddrdt ''urud (circles of latitude). This term is
also applied to the circles parallel to and at 12°
from the ecliptic on globe No. 63, mentioned
above. As for the ecliptic polar circles, they are
the two circles, one north and one south, having
the ecliptic poles at their centers and passing
through the celestial (equatorial) poles. They are
actually the paths of the celestial poles as they
circle slowly about the poles of the ecliptic in the
precessional motion. The ecliptic polar circles
(other than the ring on which the signature was
sometimes engraved on the southern hemi-
sphere) appear most frequently on the globes
having no stars (Class C), but are known to occur
occasionally on the other two types as well. On
those globes on which the lesser circles are la-
beled, the ecliptic polar circles are termed maddr
qutb mintaqat al-buruj shamdli {janubi) (the north
[or south] circle of the pole of the belt of the
houses). In the case of globe No. 94, the label
reads dd^irah saghirah az maddr qutb al-buruj
shamdli {janubi) (the small circle of the circle of
the pole of the houses).
  The full set of eight lesser circles, including
the ecliptic and equatorial tropic circles and eclip-
tic and equatorial polar circles, with one excep-
tion are found only on the globes having no stars.
They were probably indicated only in order to
complete the symmetry and not because all of

them had astronomical significance. The one
exception, a Class B globe (No. 72, made in AD
1800-1801 by Muhammad Ashraf TQqadT Za-
dah), has the full set of eight lesser circles and
the circles at 12° and 20° parallel to the equator
and to the ecliptic all prominently indicated, and
in addition has a complete set of parallel circles,
marked by very fine ink lines on the wooden
globe, at 2° intervals both north and south of
the equator and north and south of the ecliptic.
Another globe (No. 81, illustrated in Figure 22),
probably made in Yazd in the seventeenth cen-
tury, has a complete set of parallels at 5 ° intervals
from the ecliptic. A most curious network of
circles occurs on an anonymous Class B globe of
painted paper over a wooden core (No. 76, illus-
trated in Figure 28), This globe has a network
of half drawn great circles originating at the
intersection of the celestial equator and the sol-
stitial colure at the winter solstice and extending
to the equinoctial colure, marked at 5° intervals.
The hemisphere thus defined is also covered by
a series of parallels (parallel to the equinoctial
colure) drawn in red at 1 ° intervals, with every
5° parallel drawn in black. The purpose of the
network is unknown. On one globe of Class C,
No. 112, there are lesser circles about 43° on
either side of and parallel to the equator which
are also tangent to the ecliptic polar circles at the
solstitial colures. The circles are labeled on the
globe but the labels are not entirely readable in
the published photograph. They appear to be
maddr . . . "ard sh[amdli] {j[anubi]) (circle of, , .
[47?] latitude north [or south]), A similar set of
circles appears on globe No, 108.
  For a celestial globe to be functional it must
be placed in a set of rings, usually supported by
a stand. Unfortunately the majority of rings and
stands are missing today, or are modern replace-
ments, but we have a sufficient number to know
the basic forms. Slightly over one-third of the
extant globes have one or more rings that could
reasonably be considered contemporary with the
globe.^^ A close fitting meridian ring {dd^'irat nisf
al-nahdr) was placed about the globe and at-
tached to it at the celestial or equatorial poles.
The name means "circle of midday," a term
  
NUMBER 46

69



properly referring only to the upper half of the
ring, for the lower half indicated midnight {nisf
al-layl). A meridian ring is always at right angles
to the horizon ring and passes through the north-
south celestial poles. Thus it is known as a prime
meridian. The globe can then rotate within the
ring, pivoting on the celestial poles.
  Meridian rings are of two basic types, depend-
ing on whether the ring has only one set of holes
that are attached permanently to the celestial
poles of the globe or whether there are several
sets of holes in the ring in which the axis of the
globe can be set. In the first type of ring, the
meridian ring will rotate with the globe as it is
adjusted for different geographical latitudes.
The meridian ring of the first type was graduated
in a manner that essentially divided it into four
quarter circles bearing abjad numerals indicating
the graduations on each of these 90° arcs. The
numerals were nearly always placed so that they
increased from the celestial pole,^'* which served
as the 0° point, to the point on the ring alongside
the celestial equator marked 90° (see Figure 30).
The meridian ring was graduated and labeled in
the same manner and with the same scale as the
ecliptic and equator of the globe to which the
ring was attached. There are very few exceptions
to this rule. Two globes (Nos, 33 and 54, illus-
trated in Figures 24 and 25), both nineteenth-
century Indo-Persian products probably from
the same workshop, having a sliding 6° gradu-
ated scale that can move along one half of the
meridian ring (from pole to pole) to measure
celestial positions and mark them for comparison
with other positions.
  The second type of meridian ring remained
stationary with the 90° points of the non-consec-
utive numbering of the quadrants at the top
(zenith) and bottom (nadir). The axis of the globe
was then changed to one of the sets of holes for
the required terrestrial latitude. Four extant
rings have holes at regular intervals: the instru-
ment maker Ja"far ibn "Umar ibn Dawlatshah of
Kirman (No. 7, illustrated in Figure 9; the upper
half of the ring was detachable and is now lost)
and his son Muhammad ibn Ja'^far (No. 62, illus-
trated in Figure 10) placed holes at 10° intervals

and 5° intervals, respectively, while Lutf Allah
(No, 67) drilled holes every four degrees, and an
anonymous Class C globe (No, 110) has holes at
10° intervals.
  There are five extant stationary rings from all
three classes (Nos, 18, 40, 86, 89, and 114) that
have holes drilled at certain specified latitudes,^^
On all but one (No. 18) not only are the gradu-
ations labeled alongside the ring, but each hole
bears a special label on the outside face of the
ring giving the latitude in abjad numerals. All of
these rings are with globes that are apparently
seamless cast objects from Indo-Persia, probably
all from the seventeenth-century Lahore work-
shop; on all of them the northern latitude of 32°
(that of Lahore) is indicated. The only signed
globe with one of these rings (No. 18) is by Diya""
al-DTn Muhammad and is illustrated in Figure 15
(the photograph shows the ring incorrectly
mounted on the globe). Figure 32 shows in dia-
gram form two such stationary rings, the more
sophisticated version having a rotating arc that is
attached at the zenith point by a pin around
which it can pivot. The 90° arc is graduated and
can be used to measure altitude. On all three
extant rings of this form (Nos, 18, 86, and 89),
the lower half of the ring is twice as thick as the
upper half to allow this arc to rotate. The merid-
ian ring itself is held stationary by two lips at the
zero points, by which it rests immobile on the
supporting horizon ring. On one globe (No. 89)
the outside radius of the top half of the ring is
greater than that of the lower half. When the
ring is in the correct position, the axis of the
globe could be adjusted by setting the north pole
of the globe at the desired set of holes in the
upper righthand quadrant (as seen in Figure 32)
for the corresponding northern geographical lat-
itude and at a hole in the lower right hand
quadrant for a southern geographical latitude. A
simpler form of stationary ring that is not de-
signed to accommodate a rotating arc is also
illustrated in Figure 32, Three of the extant
stationary rings give settings for southern lati-
tudes of 18°, 24°, 29°, 34° or 35°, and 72°,
while one (No. 18) has a hole at every 5° interval
for the southern latitudes and one (No, 40) has
  
70

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY





FIGURE 32.—Two examples of stationary meridian rings
based on globes No. 40 and No. 86.
no perforations for such settings. There would
be little need for such settings, and in fact in
some of the rings the pins do not move freely
through all these holes. The northern holes show
more variation, but always included the one for
Lahore (32°) and nearly always one labeled 66°
30', which is the colatitude of the northern tropic
and the parallel at which the longest possible day
occurs. This is also the arctic circle and was

generally considered the northern limit of the
inhabited world on account of the intense cold.
Life was thought to disappear at a southern lati-
tude of 21 ° to 24° because of the heat.^^
  The meridian ring in turn rests in another ring
at right angles to it called ddHrat ufq (horizon
ring), which is held horizontally in a stand,
though one globe of Class A (No. 5) was designed
to be suspended rather than attached to a stand.
The horizon ring is also graduated and labeled
in abjad numerals. For Classes A and B the
graduations are usually marked by single degrees
with every fifth labeled, though other patterns
occur just as on the meridian rings and celestial
equators and ecliptics.

FIGURE 33.—Design of the horizon ring as seen from
above.
  The horizon rings in Classes A and B fre-
quently have the four cardinal points marked on
them, and the graduations are numbered in seg-
ments of 90°, each beginning at the east-west
points and ending at the north-south points,
which also contain the notches for the insertion
of the meridian ring (see Figure 33). In Class C
globes the horizon rings usually lack the names
of the directions of the compass, and the se-
quences of the numbering tends to be reversed.
  
NUMBER 46

71



On such rings, the notches holding the meridian
ring serve as the zero points and the numbering
proceeds in both directions from these two points
(see Chapter 7 for the distribution of these char-
acteristics).
  A few horizon rings also align the four cardinal
points with the seasons (east-spring, north-sum-
mer, etc.), and occasionally the zodiacal signs are
named around the horizon ring, placing the di-
vision between Aries and Pisces (vernal equinox)
at the point marked east, the summer solstice
(Gemini-Cancer) at north, and so forth. Three
globes give the names of cities along the outer
edge of the horizon ring. Globe No. 31 gives 104
cities, and No. 76 (illustrated in Figure 28) gives
the names of 72 cities. In the case of the third
globe. No. 40, the semicircular ring attached to
the underside of the horizon ring and supporting
the meridian ring and globe gives the names,
latitudes and longitudes of 42 localities, with
space left on the horizon ring itself with the
indication that it was intended to contain addi-
tional names and positions of cities, but in fact
no further names were given.^' The longitudes
and latitudes of terrestrial points were measured
in antiquity and in the Islamic middle ages by
equatorial coordinates. The longitudes were
measured from the Fortunate Islands, today
called the Canary Islands, continuing a tradition
established by the time of Ptolemy.^®
  A few globes of Classes A and B have horizon
rings with notches at the east-west points that
would have accommodated another ring at right
angles to both the meridian ring and horizon
ring. This ring would serve as a zenith ring, for
it would always mark the point directly overhead
and function as a prime vertical of importance in
the horizon system of coordinates. Five globes
(Nos. 5, 18, 33, 53, and 70) with contemporary
rings have the zenith ring (see Figures 15 and 24
for illustrations of zenith rings).
  Many of the celestial globes extant are set
within rings that are clearly modern replace-
ments. Such modern rings frequently lack grad-
uations and, in the case of horizon rings, do not
always serve to accurately represent the horizon.
There are fewer meridian rings that seem to be

part of the original apparatus than horizon rings,
presumably because of their smaller size.
  The horizon ring is supported by a stand that
can vary in style, but is of two basic types. The
most common stand consists of three or four legs
resting on a ring base (sometimes four or eight-
lobed) or on a circular plate with a central sup-
port to hold the meridian ring and globe (for
example. Figure 16), or with one or two semi-
circular metal arcs attached to the under side of
the horizon ring itself to hold the meridian ring
and globe (as in Figure 34). The second basic
style is a pedestal stand on top of which are one
or two semicircular metal arcs supporting the
horizon ring and holding the meridian ring from

FIGURE 34.—Globe No. 113, unsigned and undated. This
globe has no stars and the unusual feature of a graduated
solstitial colure. Paris, private collection. (Photo: Alain
Brieux)

72

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



below (see Figure 10). Eight-lobed ring bases
(Figure 34) and three-legged stands with the legs
in an S-shaped or leaf-shaped curve (Figures 15
and 24) are typical of Indo-Persian workshops,
while four-legged stands with top, central, and
bottom square or hexagonal knobs on each leg
are associated with light-weight seamed globes of
possibly Persian origin.
  One thirteenth-century globe (No. 3) has a
stand in which two gnomons are incorporated
into the east and west sides of the base, each
above a 90° arc that constitutes a leg of the
stand. Each arc is graduated by single degrees,
with every fifth labeled. The gnomons and scales,
which may well be contemporary with the globe,
enable the altitude of the sun to be determined.
On the stand of globe No. 6 a similar arrange-
ment of gnomons and graduated arcs was prob-
ably intended, but the stand as it now exists, does
not have the gnomons aligned over the gradu-
ated arcs so that they could function as elevation
dials. Similar basic designs are seen in the stands
of globes No. 4 (probably not contemporary with
the globe) and No. 115, though the arcs are
ungraduated. Perhaps they are simply unfinished
or poor copies of an elevation dial stand by an
uncomprehending maker.
  A globe placed within a meridian and horizon
ring assembly becomes truly functional and can
be set at any given geographical latitude by ro-
tating the meridian ring within the horizon ring
or changing the axis of the globe in a stationary
meridian ring. The globe itself could then be
rotated about the celestial equatorial axis to rep-
resent the movements of the heavens. On those
globes with stars, the stars always visible at a
given latitude could easily be determined, as
could the risings and settings of certain stars for
given times of the year, and other astronomical
information.
  The meridian and horizon rings, particularly
when there was a graduated zenith ring or a
meridian ring of the stationary style, could also
serve to provide the star positions in terms of yet
a third system of coordinates, that of altitude and
azimuth (see Figure 31). The altitude {irtifd") of
a star is its distance above or below the horizon

ring measured along a great circle or dd^irat
irtifd" (altitude circle) perpendicular to the hori-
zon and passing through the star and the samt al-
ra^s (the direction of the head, or zenith). The
point of intersection of this great circle and the
horizon was called samt (azimuthal point). The
azimuth of a star is the angular distance meas-
ured from this intersection to either the north
or the south point of the horizon ring, whichever
is closer. It was known that the complement of
the azimuth (which could be read more easily on
a horizon ring than the azimuth itself) equaled
the distance from the samt to the equator, and
that the complement of the altitude was the
distance to the zenith (called zenith distance to-
day). The one altitude or vertical circle that
passes through the two points of intersection of
the celestial equator with the horizon was called
dd^irat awwal al-sumut (the circle of the first of
the directions; in modern terminology the prime
vertical), represented in some globe assemblies
by the zenith ring. The one altitude circle that
also passes through the two celestial equatorial
poles intercepts the horizon at the north-south
points and is called dd'^irat nisf al-nahdr (the circle
of midday) and is represented by the meridian
ring.
  On stationary meridian rings with a rotating
90° arc pivoting around the zenith point (see
Figure 32), the altitude of a star position can
easily be read from the scale. If the meridian
ring is of the stationary type without the arc, the
globe can be rotated until the star is under the
meridian ring and the altitude read directly off
the meridian ring. If there is a graduated zenith
ring, then a similar procedure can be used to
read the altitude directly from the zenith ring,
even when the meridian ring is of the rotating
style.
  One anonymous Class B globe, probably from
seventeenth-century Persia (No. 81, illustrated
in Figure 22) has the meridian, horizon, and
prime vertical circles engraved directly onto the
surface of the globe. In this instance they are
stationary and are valid for only one latitude.
The globe was made for 32°, which is the lati-
tude of Yazd. This same globe also has engraved
  
NUMBER 46

73



and labeled on it a declination circle and a celes-
tial latitude circle passing through the star la-
beled "ayyuq {a Aurigae, Capella). Such circles
marked directly on the surface of a globe are
clearly for didactic purposes (as shown in Figure
31). The circles can be used to demonstrate the
three known systems of coordinates: ecliptical,
equatorial, and horizontal. All three of these
systems were discussed by Ptolemy, and through
the Arabic translations of the Almagest, were well-
known in the Islamic world.
  The most common form of meridian ring was
the movable type—that is, one with only two sets
of holes by which it was attached permanently to
the globe (see Figures 30, 31, and 34). By using
the graduations of the meridian ring when it is
attached to the celestial poles, the globe can be
adjusted (rectified) to correspond to any given
terrestrial latitude, and the declination of a star
can be read by turning the globe so that the star
is alongside the meridian ring, and then reading
off the declination from the graduations. The
right ascension of the star would be read by
finding where the great circle passing through
the star and the celestial poles intersected the
equator. The meridian ring in this arrangement
of the rings is nearly always numbered from the
poles,^^ The distance from the star to the celestial
poles was probably read directly from the ring
and its complement taken for the declination.
This particular ordering of the meridian ring
graduations has the advantage, however, of al-
lowing easy rectification of the globe to corre-
spond to a given terrestrial latitude simply by
placing the corresponding degree of the merid-
ian ring at the north point of the horizon ring,
with the north celestial pole above the horizon
ring for northern latitude and below for south-
ern latitudes.
  While it was customary to mount the meridian
ring to the globe at the celestial equatorial poles
so as to correctly represent the rotation of the
heavens, occasionally holes were drilled at both
the celestial and ecliptic poles so that a moveable
axis through the globe and the meridian ring
could be adjusted to either set of poles. When
the meridian ring was set at the ecliptic poles.

the celestial latitude could more easily be calcu-
lated from the meridian ring and the celestial
longitude read from the ecliptic. It was custom-
ary to read a celestial longitude within a zodiacal
house, such as Taurus 15° rather than 45° from
the vernal equinox, which is why the ecliptic is
nearly always numbered in twelve 30° segments
and always has the names of the zodiacal house
written over each segment.
  When there is a hole at the ecliptic as well as
celestial poles, it would be possible to use a scale
such as that described in a chapter appended to
the thirteenth-century Spanish translation of
Qusta ibn LQqa's tract on the use of celestial
globes.^*' This measuring device consisted of a
90° arc whose inner diameter was concentric
with the surface of the globe. The quadrant was
graduated by single degrees with every fifth la-
beled. At the end of the arc labeled 90° a thin,
well-made nail was attached, which could be eas-
ily inserted into one of the ecliptic poles. The
arc, which was thus always at right angles to the
ecliptic, could be rotated about to read off the
celestial longitude and latitude of any star and
could be easily removed to permit the globe to
turn under the meridian ring. It could be placed
in either ecliptic pole depending on whether
northern or southern coordinates were required.
  When it was not possible to attach the meridian
ring to the ecliptic poles of the globe, or to use a
measure like that just described, drawing com-
passes were used to measure the celestial latitude
and to compare the distance with the graduations
of the equator. (The equator was used rather
than the ecliptic since the former was numbered
sequentially.) Sometimes a wooden or brass 90°
arc, with the same curvature as the globe and
calibrated in the same scale as the equator, eclip-
tic, and rings, was placed over the great circle
representing the celestial latitude of the star and
the angular distance then read directly from the
scale, as al-KhazinT described in his twelfth-cen-
tury treatise.
  Stationary meridian rings had the advantage
of allowing one to easily read the zenith distance
and altitude of any star on the globe, while the
azimuth would be read from the horizon ring.
  
74

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



For the globes where the adjustment for geo-
graphical latitudes was made in the more usual
way of rotating the meridian ring along with the
globe, altitude could be found using divider com-
passes, or the 90 ° graduated moveable arc could
be placed on top of the horizon ring and the
altitude read from it. One of these two instru-
ments (the divider compasses or the graduated
moveable arc) was necessary for reading the dis-
tances between stars on the globe.
  The utility of the three classes of globes in the
Islamic world can be deduced from the treatises
on the use of celestial globes, such as that by
Qusta ibn LQqa written in the ninth century, and
from the form of the extant globes themselves.^'
The globe described by Qusta ibn LQqa differs
in basic design from that described by Ptolemy
in his Almagest (as do all extant globes) even
though the treatise was very well known and
formed the basis of all star catalogs in the Islamic
world.
  It is curious that in the literature as surveyed
in Chapter 1, we find no reference to celestial
globes lacking stars (Class C globes) nor to any
attempted constructions of a precession globe of
the sort described by Ptolemy. All Islamicate
globes having stars, by bearing both an equator
and ecliptic are thus made for only a certain
epoch. The precession of the equinoxes is suffi-
ciently slow that such a celestial globe (just as an
astrolabe) will not be obsolescent for 50-75
years. There is evidence that the globes lacking
stars are quite late, and are possibly a Safavid
Persian development. While Qusta ibn Luqa de-
scribed only a globe bearing the major stars on
it (Class B) and does not even mention constel-
lation outlines, we can easily see that some of the
uses to which he puts celestial globes are readily
applicable to all three classes of globes, even
those of Class C.
  When Class C globes are mounted in meridian
and horizon rings they represent a model of the
celestial system not subject to precessional
change, much as does a simple demonstrational
armillary sphere (one lacking the orbits of the
moon and planets). Following the treatise by
Qusta ibn LQqa (chapters 1-8),^^ it is evident that
a globe of this class, though never mentioned by

Qusta, could be used as a didactic device to
demonstrate the following principles: the differ-
ence in the positions of the sun for different
terrestrial latitudes; the equality of day and night
for any latitude when the sun is at one of the
equinoxes; the longest and shortest day for any
given terrestrial latitude; which latitudes have a
six-month night and a six-month day; the latitude
at which there are no shadows at certain times of
the year; and on what day this last event occurs.
  More importantly, a well-made Class C globe
could be not only an instructional tool, but a
simple means of computing certain information.
It should be kept in mind that the measurement
of time can be based either on the apparent
rotation of a star (sidereal time) or on that of the
sun (solar time). Due to the movement of the sun
along the ecliptic, the solar day is about four
minutes longer than a sidereal day. The proce-
dures for measuring time by means of a celestial
globe which are described in the literature, and
of which we shall give some examples, in fact
measure sidereal rather than solar time. The
intervals measured by these methods, however,
are no greater than one day, and the difference
between the solar and sidereal units could be
consequently negligible.
  A day can be divided into equal or unequal
hours. If the day is considered the length of time
required for the apparent rotation of a star and
that interval divided into 24 equal hours, the
right ascension (angular distance along the celes-
tial equator) could be measured in hours as well
as degrees, allowing one hour for 15°. Equal
hours were employed primarily for astronomical
calculations, while civil and religious time-keep-
ing was based on a system of unequal or temporal
hours {horae temporiae). In this system the period
from sunrise to sunset was considered one-half
the day and allotted 12 hours, while the period
from sunset to sunrise was also divided into 12.
On any day other than the equinoxes the length
of an hour of daylight will not equal the length
of an hour at night.
  By means of a Class C globe having rings,
ecliptic, and equator (all of which must be grad-
uated), it is possible to compute the following
items: the difference between the length of night
  
NUMBER 46

75



and day for any given terrestrial latitude; the
difference in terms of equal hours of the daytime
periods of any two days for any given terrestrial
latitude; the difference between the daytime pe-
riod of any given day in two given towns of
different latitudes; the length of the hours (un-
equal hours) for any given day at any given
latitude; the latitude of the sun at noon for any
terrestrial latitude on any given day; the differ-
ence between the maximum elevation of the sun
on any one day in two different towns; the alti-
tude of the sun above the horizon for any day at
any location given the time of day; the hour of
the day (in equal or unequal hours) given the
altitude of the sun, the day of the year, and the
terrestrial latitude; or the hours of the day given
the ascendant (degree of ecliptic at eastern hori-
zon) at a certain location on a given day. Ob-
viously, if the altitude of the sun is known, then
the ascendant can be read at the eastern part of
the horizon ring, and if the ascendant is known,
the altitude of the sun can be measured on the
globe for a given day.
  The last three items require information such
as the time of day or the altitude of the sun,
which would have to be read from another in-
strument, such as a gnomon or quadrant. The
position of the sun in the ecliptic could be found
from a calendar. To give some specific examples,
to find the length of the unequal daytime hour
for a given day in a given town, the astronomer
would, after rectifying the globe for the given
geographical latitude, place the degree of the
ecliptic where the sun is on that day on the
eastern horizon and then mark the segment of
the celestial equator that was also on the eastern
horizon. After rotating the globe until the degree
of the ecliptic containing the sun is at the western
horizon, he would mark the segment of the ce-
lestial equator that was at that time at the eastern
horizon. By counting the degrees from the first
to the second mark along the equator in a west
to east direction and dividing the result by 12,
the astronomer would then have the length of
the unequal hours, allowing 1 ° per 4 minutes.
  If the astronomer wished to know how much
time had elapsed (in terms of equal hours) on a
certain day given the ascendant and the latitude

of the town, then after rectifying the globe for
that particular latitude and setting the ascendant
(the degree of the ecliptic corresponding with
the east horizon) at the east point of the horizon
ring, he would mark down the degree of the
celestial equator that also falls on the east hori-
zon. Then he turns the degree of the ecliptic
corresponding to the position of the sun for that
particular day to the east horizon. He notes the
degree of the celestial equator that is also now at
the east point of the horizon ring. Taking the
difference of these two readings and dividing by
15, he finds the number and fractions of hours
which have elapsed.
  Given the time elapsed (in either equal or
unequal hours) for a particular day, or the alti-
tude of the sun at a given location, the ascendant
could also be found. This was of considerable
interest to astrologers, for once the ascendant
was known, then other positions along the eclip-
tic that were employed in astrological horoscopes
could be very easily read from a celestial globe.
When the degree of the ascendant was at the
eastern horizon (placed in Locus I of the horo-
scope) the descendant (Locus VII) equaled the
degree of the ecliptic coinciding with the western
horizon. The "middle of the heaven" (Locus X)
is the point of the ecliptic at the upper half of
the meridian ring, and the "peg of the earth"
(Locus IV), the part of the ecliptic coinciding
with the lower half of the meridian ring. After
rotating the globe 30° from the initial position
of ascendant at the east horizon. Locus II can be
read at the east part of the horizon ring. Locus
XI at the northern half of the horizon ring.
Locus VIII at the western horizon, and Locus V
at the lower meridian ring. Then by rotating the
globe eastward 60°, Loci XII, IX, VI, and III
could be read at the same four locations, respec-
tively. In an older and more complex system, the
four initial Loci were determined in the same
manner as just described. The other positions
along the ecliptic for the other eight Loci were
determined by finding the length of the unequal
daytime hours for the particular day and then
rotating the ascendant to the eighth hour
(roughly 120°) and reading Locus II at the
northern culmination of the ecliptic and Locus
  
76

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



VIII at the southern culmination; then rotating
it to the tenth hour (roughly 30° further west-
ward) and reading Loci III and IX in the same
manner; and then rotating the descendant to the
second hour (roughly another 60° westward) for
Loci V and XI, and the descendant to the fourth
hour (about an additional 30°) for Loci VI and
XII,^^ In this manner the celestial globe, like the
astrolabe, allows astrologers to read directly the
degree of the ecliptic for the 12 positions of a
horoscope, thus enabling them to avoid compli-
cated calculations.
  The celestial globes that have stars are able to
do all of the things that Class C globes can do, in
addition to a considerable number of demonstra-
tions and calculations involving the stars. Such
globes, with the meridian and horizon rings, can
be used to demonstrate quite effectively the dif-
ferences in the apparent motion and visibility of
stars, as well as the sun for different geographical
latitudes, and at what latitudes the stars never
rise or set, since the horizon ring divides the
upper (visible) part of the heavens from the lower
(invisible) part. They can also be used to dem-
onstrate the daytime path of any star depicted
on the globe as well as the rising and setting of
each fixed star in any given geographical latitude.
  There is no significant difference in the func-
tion of celestial globes of Classes A and B, the
only difference being that those of Class A pres-
ent a fuller notion of the asterisms, depict the
convenient and familiar pattern of the constella-
tion outlines, and are generally larger and so
allow for greater precision than do those of Class
B. In addition to the computations discussed for
Class C globes, all of which can be done with
Class A and B globes as well, the celestial longi-
tudes and latitudes, the declination and right
ascension, and the altitude, zenith distance, and
azimuth for any star on the globe at any given
geographical latitude can be found. An astrono-
mer may calculate the distance between any two
stars on the globe, the distance for any given
location between two points on the horizon
where any two stars rise, the maximum elevation
of each star on the globe for any given town, the
azimuth of the direction of prayer {qiblahf'^ for

any given town. If the altitude of any star not on
the globe, or planet or moon, is known, as well
as the altitude of a nearby star appearing on the
globe, then an approximation of the other coor-
dinates can be found from the globe. The hour
(in equal hours) that any star on the globe rises
could be computed for any given latitude by
setting the star at the east horizon (once the globe
had been rectified for the given geographical
latitude) and noting the segment of the equator
at the eastern horizon. Then by rotating the sun
to the eastern horizon (or, if the star rises at
night, rotating the globe until the nadir of the
sun is at the east point) and noting the degree of
the celestial equator at the horizon, the hour can
be found by counting the degrees between the
two readings and dividing by 15, The twilight
period was considered to occur when the sun was
between 6° and 18° below the horizon. At the
end of twilight all stars visible to the naked eye
could be seen. In this way one hour and 12
minutes (equal hours) after the sun has set all
stars would be visible that were above the horizon
ring when the degree of the ecliptic containing
the sun was set at 18° below the western edge of
the horizon ring.
  The celestial globe (No. 3) made by Qaysar
ibn AbT al-Qasim ibn Musafir al-Ashrafi al-Hanafi
for the nephew of Saladin in 622 H/AD 1225-
1226 is especially interesting. Incorporated into
the quadruped metal stand are two gnomons on
the east-west sides, each gnomon being over a
graduated 90° arc. By means of these gnomons
and graduated arcs, which are essentially altitude
dials, the altitude of the sun can be read by the
shadow cast by the gnomon on the dial. A some-
what similar arrangement seems to have been
intended on globe No. 6, stated to be by Muham-
mad ibn MahmQd al-TabarT and dated 684 H/AD
1285-1286. After obtaining from a calendar the
position for that day of the sun along the ecliptic,
the astronomer can then rectify the globe for the
geographical latitude of the town, and set the
degree of the ecliptic where the sun is to a
position equal to the elevation of the sun; the
angular distance from the point on the ecliptic
occupied by the sun to the horizon ring equals
  
NUMBER 46

77



the angular distance shown on the gnomon scale,
measured either from the east or west side of the
meridian ring depending upon the position of
the sun as shown by which altitude dial casts a
shadow on the scale if the astronomer has ori-
ented the globe toward the north. This done, the
hour of the day can be calculated, or the ascend-
ant located. A nice feature of this design is that
the gnomons or altitude dials built into the stand
of the globe will work equally well at any geo-
graphical latitude. To align the position of the
sun on the globe with its proper altitude, how-
ever, the latitude of the town must be already
known and the globe itself rectified accordingly.
  Globe No, 5 was made at the end of the
thirteenth century by Muhammad ibn Mu^yyad
al-TJrdT, It may have had the type of sliding
gnomon described by al-BattanT in the ninth cen-
tury (see Chapter 1 and Figure 5), That is, the
zenith ring appears to rotate about the globe and
to have a slit running along the upper half
through which a needle-like gnomon may have
been placed. If the gnomon were kept steady
over the spot along the ecliptic where the sun
was known to be for that day while the sphere
was rotated until the shadow cast by the gnomon
was as small as possible, then the resulting posi-
tion of the globe would indicate the altitude of
the sun at that moment.
  It is in fact theoretically possible to use any
celestial globe as a spherical elevation or altitude
dial if the astronomer has a thin, portable gno-
mon. ^^ After the globe is rectified for the geo-
graphical latitude and the north point of the
horizon ring made to point north, if a small
gnomon or pin is placed at the position of the
sun in the ecliptic for that given day and placed
perpendicular to the surface of the globe at that
point, then the globe is rotated east or west until
the shadow of the gnomon is as narrow as possi-
ble. When the globe is in this position the sun's
point on the ecliptic is at the same altitude with
respect to the horizon ring as the sun is in terms
of the observer's horizon. And of course, once
the altitude is obtained, other information such
as the time of day can be determined. In this way
the entire celestial globe could be used, in con-

junction with a thin portable gnomon and com-
pass (to tell the direction of the earth's axis) as a
spherical elevation dial to determine the altitude
of the sun on a given day at a given latitude.
  To perform any operations with a celestial
globe other than the merely didactic ones of
demonstrating the longest and shortest days,
which stars are visible at a given latitude, or
similar celestial phenomena, the astronomer
needed to know the position of the sun in the
ecliptic, which required the use of a calendar (a
calendar was similarly required for the astrono-
mer to use an astrolabe and was consequently
usually inscribed on the back of the astrolabe).
Drawing compasses, or a curved graduated arc
made especially for the globe, were required to
measure some of the coordinates of the sun or
stars. If observational instruments were em-
ployed together with a celestial globe complete
with rings (such as a quadrant to measure alti-
tudes or a gnomon for telling time of day or a
portable gnomon, as just mentioned), an even
greater amount of astronomical and astrological
information could be calculated.
  Unlike a planispheric astrolabe, a celestial
globe possesses no sighting device. If, however,
the use of a carefully constructed celestial globe
with stars is supplemented with that of a quadrant
and gnomon, then all the astronomical and as-
trological data accessible by calculation and the
use of an astrolabe^^ can also be obtained by
means of a celestial globe. The astrolabe, like the
celestial globe, is not a direct reading instrument.
After making the observation, the astronomer
must dial and calculate the desired information.
The common planispheric astrolabe can be
viewed as a flat representation of a celestial globe,
with star positions indicated for only one latitude,
A separate plate is required for every latitude,^^
  The planispheric astrolabe has the advantage,
however, that the alidade or flat ruler with a
sight, which is used to measure the altitude of
the sun and stars above the horizon, can also be
used to measure the height of mountains or
towers or the depth of wells, which of course
cannot be done by means of a celestial globe.
Moreover, the astrolabe is much more portable
  
78

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



than a celestial globe, and requires fewer supple-
mentary instruments.
   As we have indicated, a celestial globe has the
advantage to the user of being usable at any
geographical latitude, and a large carefully con-
structed globe allows greater precision in the
calculations than does an astrolabe. For the
maker a globe, unlike an astrolabe, does not
require a knowledge of stereographic projection
for its construction,"''^ Even though a celestial
globe is easier to design, it is more difficult to
fabricate from the standpoint of the metal-
worker. The awkwardness of transporting a
globe with stand and the lack of a sighting device
clearly mitigated against it being a serious rival
of the planispheric astrolabe as an astronomical
computing device. Nevertheless, it is clear from
the historical evidence that celestial globes were
an important part of the equipment of an astro-
nomical observatory and were considered of
practical value by astronomers. On one six-
teenth-century globe (No. 10, illustrated in Fig-
ures 11 and 69) the maker informs us that he
made his well-constructed celestial globe with the
full set of constellation figures "in a manner
useful for the knowledge of all the requirements
of astrolabe makers, as an aide-memoire to their
craft." It would appear questionable, however,
how many of the extant globes of any class were
of more than didactic and artistic value to their
owners.
  The comparison of celestial globes with a com-
mon or planispheric astrolabe brings up the ques-
tion of whether any of the extant celestial globes
could also have served as a spherical astrolabe.
From the astronomical literature we know that
there were spherical astrolabes {asturldb kuri)
designed to perform the same function as plan-
ispheric astrolabes but not requiring for their
construction a knowledge of stereographic pro-
jection, and allowing for adjustment to different
latitudes."^^ It is possible that spherical astrolabes
developed along with the planispheric ones, and
thus were known by Hellenistic times.^*^ It is
equally possible, however, from the evidence
studied so far, that this type of astrolabe was an
Islamic development.'*' What is certain is that the

spherical astrolabe was never as popular as the
planispheric variety, and in fact only two are
known to exist today: one made by an otherwise
unknown maker, MQsa, 855 H/AD 1480-1481,
now at the Museum of the History of Science,
Oxford,^^ and the other unsigned and undated
but made for use in Tunis (possibly sixteenth
century), now in a private collection.^^ Qusta ibn
Luqa may possibly have written the earliest
known treatise on the spherical astrolabe, in ad-
dition to his tract on celestial globes.
  The two extant spherical astrolabes and those
described in the treatises, while differing in par-
ticulars of style, consist basically of a sphere
(made in two hemispheres) on which are marked
the circle of the horizon, circles of equal altitudes
(almucantars, at regular intervals from the hori-
zon line), vertical (altitude) circles, hour lines for
reading unequal hours, and a series of holes for
adjustment to different latitudes. Over this
sphere is placed a rotatable cap (the rete) of open
metalwork, which represents the ecliptic, the ce-
lestial equator, some star positions north of the
ecliptic, a graduated vertical quadrant with slid-
ing gnomon for measuring solar altitudes, and
some type of sighting arrangement for stellar
observations.
  Clearly none of the extant celestial globes
could have served as part of such a spherical
astrolabe, for all (including those of Class C) have
the ecliptic drawn directly on them, and none
have almucantars, hour lines, or holes bored in
them for various geographical latitude adjust-
ments.^'* Only one has a horizon line indicated
(No. 81, see Figure 22), but it is drawn over the
ecliptic and the celestial equator, making it sta-
tionary for the latitude 32°. While it has a full
set of parallels, they are at 5° intervals from the
ecliptic rather than from the horizon line. This
one globe was clearly designed only for instruc-
tional purposes.
  P. Tannery"*^ has suggested that the hemi-
spherical sundial could have given rise to the
concept of a celestial globe bearing the major
stars over which there was placed a hemispheric
"spider" or rete carrying horizon and hour lines.
Professor D.J. de Solla Price'**' extends this ar-
  
NUMBER 46

79



gument and suggests that there might have been
two types of spherical astrolabes, just as we know
there were two types of planispheric astrolabes.
The classic form of a planispheric astrolabe con-
sists of a rete depicting the ecliptic and major
stars that turns over a plate, with the stereo-
graphic projection of the almucantars concentric
with the observer's zenith and the vertical circles
above the horizon. The early type (derived from
the astrolabic or anaphoric clock, of which we
have fragments dating from the first century
AD)^^ is composed of a solid disc inscribed with
stars and the ecliptic, which revolved behind an
open rete showing the hour-lines, the horizon and
almucantars, and vertical circles. Parallel with
these two styles of planispheric astrolabes, Price
postulates that in addition to the classical form
of spherical astrolabe described earlier, there
might have been a second type composed of a
sphere on which the major "astrolabe" stars, the
ecliptic, celestial equator, and at least two great
circles passing through the celestial poles (i.e.,
the equinoctial and solstitial colures) were
marked. The rotating cap {rete) would consist of
the horizon lines, a series of almucantars (and
vertical circles?), and hour lines for the unequal
hours. The rotating cap would require a series
of holes or a slot, by which it could be attached
to the celestial poles of the globe and adjusted to
different terrestrial latitudes. It would also need
a sighting device.'*^
  Such a concept of a spherical astrolabe suggests
the use of a Class B celestial globe as a base. If
any of these globes could have been used in this
manner we must assume that all the retes have
been lost from all the extant examples. This is
not an unreasonable assumption, since of the two
classic spherical astrolabes, one has lost the rete.
It is true that such a spherical astrolabe is not
known from the published treatises, but this
omission does not by itself rule out the possibility
that such existed. We have a good many Class C
globes, which are not mentioned in any known
accounts of celestial globes.
  There are certain factors that mitigate against
this otherwise appealing theory for the use of
celestial globes. Although Qusta ibn Luqa de-

scribed a globe having only the major stars rep-
resented, he nowhere mentions using it as a base
for a spherical astrolabe. Secondly, of the 28
globes'*^ in the category having the major stars
and no constellation figures, 15 have one or both
rings that are probably contemporary with the
globe. One (No. 72) has the maker's signature
on the stand and horizon ring as well as the globe
itself (see Figures 10 (No, 62) and 28 (No, 76)
for other examples of contemporary globes and
rings). In these cases it is fairly certain that the
sphere was intended as a celestial globe and not
as a base for a spherical astrolabe. Since the stand
and rings are more easily lost than the sphere
itself, it is not at all surprising that the remaining
13 either have none or have replacements. Fur-
thermore, more than half of the globes of this
class have both the equatorial tropic and polar
circles, with some also having ecliptic tropic cir-
cles or ecliptic polar circles, whose function is
unexplained when considered as a base for a
spherical astrolabe (see the tables following the
catalog for the distribution of characteristics). All
but one (No, 72) of the globes have the six great
circles passing through the ecliptic poles (ecliptic
latitude circles) which include the solstitial col-
ure. Only eight globes have the equinoctial col-
ure, and these eight are all among the group
having tropic and polar circles. One globe (No,
59) has a hole only at the ecliptic poles and could
not take a cap mounted at the equatorial poles.
  It is difficult to see how any of the globes of
this sort could easily have functioned as a base
for a spherical astrolabe. Moreover, on those
examined, there are no visible signs of a cap
having been rotated over the surface of the
sphere. It is significant that in the preface to his
tenth-century tract,'^'^ al-NaynzT draws several
comparisons between the celestial globe with
stand and the spherical astrolabe. He argues for
the superiority of the astrolabe. He never men-
tions using a celestial globe as a base for a spher-
ical astrolabe, as one might expect if the idea
were at all current in his day. Nonetheless, the
theory is intriguing, and warrants further inves-
tigation into the written sources concerning both
instruments. It is certainly possible that some of
  
80

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



the celestial globes were used for purposes other
than those outlined by Qusta ibn LQqa or al-
KhazinT.
  In addition to the matter of stars represented,
there are differences in the design of the three
classes, in such matters as the representation of
lesser circles which occur more frequently on
Class B globes than on Class A globes and are
universally present on those of Class C. Class A
globes, in addition to being generally larger and
for that reason more accurate, are frequently
constructed more precisely than those of Class
B, and certainly more so than those of Class C.
For example, on Class C globes, the graduations
are usually carelessly executed. These differ-
ences between the classes of extant Islamicate
globes no doubt indicate different intended uses
for the globes, some of which are not apparent
from the written sources or extant artifacts.
  There are five globes in the catalog that do
not fit easily into this categorization. Globe No.
63 (Figure 20), made in 1012 H/AD 1603-1604
for Shah "Abbas I, the Safavid ruler of Persia,
and globes No, 82 and No, 83 each has a selection
of stars that technically places it in Class B. Each
also has 12 medallions containing the zodiacal
signs represented by emblematic motifs that are
to be found on other examples of Islamicate
metalwork,^' as well as the coins issued by the
Mughal ruler of India, JahangTr, who was a con-
temporary of Shah "Abbas 1.^^ These illustrations
of the 12 zodiacal signs are completely outside
the tradition of the diagrammatic illustrations of
the zodiacal constellations in treatises such as
"Abd al-Rahman al-Sufi's influential Constella-
tions of the Fixed Stars. None of the designs in
these medallions outline asterisms that serve as
mnemonic devices for star positions. In fact,
there are no stars present in the medallions.
There are no planets depicted with the zodiacal
figures, as are commonly found in other pieces
of metalwork.^^ Leo is a lion standing on all fours
with mouth open and tail curled over his back,
but with no sun; Virgo a bearded man seated
holding a sheaf of wheat and a sickle; Libra a
man sitting cross-legged with scales over his
shoulders like a yoke; Pisces two curved fish, one

above the other; Aries a four-legged animal with
what looks like antlers; and Taurus a bull with a
hump on his back and a bell around his neck.
  On all three globes the stars are inaccurately
positioned and are not labeled, in contrast to
most Class B globes. They all have ungraduated
ecliptic and equator circles (rather rare on celes-
tial globes), and several lesser circles, all of which
are labeled. The globe produced for Shah "Abbas
I is the earliest dated globe to have ecliptic tropics
indicated on it. It is also unique in having a large
crescent moon engraved on it between the north
equatorial pole and the House of Aries; the cres-
cent moon is a common motif in decorative Is-
lamicate art,^'* If these three globes had any
function other than as objets d'art, it would be as
a Class C globe, making use of the greater and
lesser circles indicated but ignoring the stars. In
any case, they represent an artistic style clearly
outside the usual Islamicate globe-making tradi-
tion.
  Globe No. 73 was made in 1221 H/AD 1806
by Muhammad "AIT al-HusaynT. It also has zodia-
cal figures. The star positions are slightly more
accurate than the preceding three globes, and
most of the 50 stars are labeled. Iconographically
there are some similarities between the zodiacal
figures depicted on it and those on the globe
made for Shah "Abbas I and the two closely
related globes. The zodiacal figures of the nine-
teenth-century papier mache globe, which are not
set in medallions although they are still not con-
stellation outlines, also have features in common
with the zodiacal figures on globe No. 56, This
latter anonymous papier mache globe (see Figure
29) also does not fit into the basic three categories
of Islamicate celestial globes,^^ It has a full set of
constellation figures painted on it, but has no
stars at all. The constellation figures are not
asterism outlines and seem to represent yet an-
other iconographic tradition,^®
  Of the 126 Islamicate known celestial globes,
all are hollow metal globes except seven (Nos.
36, 56, 72, 73, 76, 93, 94). These are either
painted wood or wooden cores with papier mache
covering, which is painted and lacquered (see
Figures 19, 27, 28, and 29),
  
NUMBER 46

81



  Textual evidence for the different methods
employed in the construction of the sphere itself
is fairly meagre. In the Almagest Ptolemy speaks
only of a solid sphere, but gives no details as to
the material or methods used to form it.*'"' Leon-
tius (see Chapter 1) in the seventh century says
simply that if the sphere is of wood it is to be
stained, and it is to be smoothed over with plaster
or wax to fill any cracks; when dry it is to be
painted a deep azure blue. He gives no specific
information on how it is to be fashioned, whether
it is solid or hollow, whether in one or more
pieces, or whether it was smoothed and shaped
on a lathe. He does seem to imply that it might
be made of something other than wood. At the
end of the ninth century, al-BattanT supplies little
information on the actual construction of the
sphere, only saying that it should be of copper
{nuhds) and turned on a lathe {makhrdtah f i al-
shihr). Qusta ibn LQqa gives no directions for the
construction of the sphere or the rings, although
when mentioning the placement of the globe in
the horizon ring he does refer to it as "the
stone. "•'^^
  For more detailed information on the process
of construction we must turn to the Spanish
translation of Qusta ibn LQqa's treatise, prepared
for the king of Castile, Alfonso X of Seville.
Because the original treatise contained nothing
on construction, King Alfonso directed that a
section be added on the materials and processes
employed in the construction of celestial globes.
Thus the first four chapters,''^ written between
AD 1276 and 1279, provide detailed information
that may reflect practices prevalent in the thir-
teenth century and earlier in Spain, in other
western parts of the Islamic world, and possibly
elsewhere. The treatise states that the sphere
might be made of any of several materials all of
which, except for wood, are discounted as being
deficient. Gold and silver are too expensive,
heavy, and easily bent; copper, while stronger
than gold or silver, is not as maleable and cannot
be worked as easily; brass, which he calls tinto
cobre, is more maleable than copper and stronger
than gold or silver, but if a sphere were thick
enough not to be bent out of shape easily, it

would be too heavy. The author adds, however,
that of all metals this is best. Iron would be too
heavy and rust too easily, while tin would bend
easily if thin and be too heavy if sufficiently thick.
The same arguments apply to lead, which has the
additional problem of turning black and destroy-
ing the constellation outlines and having to be
cleaned by scraping. The author then mentions
ceni,^^ an alloy used by the Arabs for basins and
other household objects, which is too brittle.
Another alloy, fuslera, requires smelting and is
unserviceable for a sphere. Stone is dismissed as
being too heavy and if of the transparent type
(Pquartz) the constellations could not be easily
seen. Clay globes would break and figures could
not be made on them (a rather curious statement
since pottery was easily decorated), and would be
too ignoble a material for such a noble object.
Leather could not be shaped sufficiently nor
retain its shape for long, and would shrivel in
heat. Cloth and parchment would shrivel or
stretch with the weather and in the end be worth-
less. Thus wood is, in the author's view, the
optimal material.
  According to the compiler, the wood selected
must be of a kind that will warp, split, and rot
the least. For this reason it must be gathered
when the moon is waning in the last days of the
lunar month. Immediately after gathering, it
must be soaked in hot water for two or three
days and then placed in sunlight to test for warp-
ing or splitting. If the wood neither warps nor
splits, the artisan may proceed in one of two
ways, depending on whether a solid or hollow
sphere is desired. If the globe is to be solid, the
artisan then makes the piece as round as possible
while keeping it slightly larger than the intended
finished product. After marking the two points
he intends to use as poles, he places those points
on the two ends of a lathe. Then with a steady
hand and using a wide cutting edge made of
good steel with a thin, very sharp edge, he turns
the lathe and evenly removes from the surface
of the wood the desired amount to produce a
round solid sphere.
  If a hollow wooden globe is to be produced,
he can either make it from a solid one as just
  
82

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



described or use wooden rings. If the maker
chooses to produce the hollow globe from a solid
sphere he draws and cuts out a circle on the
surface of the sphere, through which he can
hollow out the rest of the globe; after that he
replaces with glue the circle of wood as a plug
and smooths the sphere again on a lathe.
  This section of the treatise presents two pro-
cedures that can be used for making a hollow
wooden globe involving wooden rings. By the
first method the carpenter cuts a large number
of wooden rings of the same diameter and thick-
ness. Two of them he places at right angles, one
inside the other, having notched both rings in
such a way that when aligned both the joins are
smooth and both rings have the same circumfer-
ence. Inside these two rings and at right angles
to both is placed a smaller wooden ring, whose
outside radius equals the inside radius of the two
first circles. It is attached to the two original
rings by four well-made wooden nails at the four
points of contact. From other wooden rings of
the same size as the first two, the draftsman then
cuts several arcs of varying sizes and shapes,
which are attached with wooden nails to the
smaller inside ring, thereby providing the re-
maining surface betwen the two initial rings of
the globe. (The author does not describe how
these pieces are to be attached to the two original
rings. Perhaps it is done by shaping their ends
into wedges and glueing together adjacent rings.)
  The second method using wooden rings begins
by making a large number of wooden rings of
diminishing diameters (two of each size). These
rings are then stacked with the largest in the
middle, graduated to the smallest on either side,
thus producing, more or less, a sphere. Both of
these methods would produce a frame whose
sphericity would depend in large part on the
width of the rings.
  After the wooden sphere was constructed, it
was to be covered with parchment using a very
strong glue. After that had set, the sphere was
covered with leather of the sort used for shield
covers, but cut thinner. The leather was then
smoothed until the sphere was very round. Fol-
lowing this, the sphere was covered with a layer

of a rather thick plaster called plastro, and then
plastered again with a thin gris (literally, gray).
The plaster was then smoothed and the globe
tested for sphericity with a pair of dividers.
  In the treatise on the spherical astrolabe {as-
trolabio redondo) in the Libros del Saber, that
draws from Arabic sources, Isaac ben Sid pre-
sents another method of constructing a hollow
wooden sphere, this time by making two hemi-
spheres.'^' According to his description the maker
is to produce four wooden rings of the same size;
their thickness will equal the thickness of the wall
of the globe. One of the rings should have 12
notches cut in it at 30° intervals, each notch
going half-way through the thickness of the ring
and all the way through the width of the ring,
being wide enough to allow for the width of one
of the other rings. A second of the four rings is
then cut in half, placing in each end a notch of
half the thickness of the ring. At the midpoints
of these semi-circles and on opposite sides a notch
should be made equal to the breadth and one-
half the thickness of the other, so that they can
be crossed and meshed over each other. The
four notched ends are then placed in the four
notches (90° apart) of the first ring. The two
remaining rings are each cut into four pieces,
and at one end of each a notch equal to one-half
the thickness is cut. The other end of each 90°
arc is tapered so that they can fit together at the
intersection of the two crossed half-circles. Their
notched ends are then placed in the remaining
eight notches of the first ring, resulting in a
frame for a hemisphere with 12 sections formed
by wooden arcs attached to the first ring. The
intervening spaces are then filled with pieces of
wooden rings of equal radius that are placed in
each segment, at first dividing it into four parts,
then eight, and so on, until eventually the space
is closed. The compiler notes that what might
have been intended was for the maker to use
pieces of rings of decreasing radii and place them
in the open segments parallel to the one complete
ring, thus stacking and glueing them together in
each segment until it was filled.
  When two identical hemispheres are made in
this manner, they are to be joined together to
  
NUMBER 46

83



form a hollow wooden sphere, in which two holes
are then bored at diametrically opposite points.
Through these holes a round wooden rod was
passed to serve as an axis {mihwdr) for the sphere.
Such an axis is not used in a spherical astrolabe,
which indicates that the compiler may have ob-
tained this method from a work on celestial
globes, since they do require an axis about which
the globe can pivot, Isaac ben Sid ends by saying
that in order to produce a smooth surface,
leather from a horse or donkey would be placed
over the wooden sphere, and its surface whitened
with chalk.
  None of the seven extant wooden or papier
mache globes has been analyzed as to the precise
method of construction so as to compare their
construction with these thirteenth-century pro-
cedures. To make such an analysis would involve
serious alteration of the globe.
  The vast majority of extant Islamicate globes
are in fact made of metal and appear to be of
two basic types: those with an evident seam, and
those that appear seamless. Both types are hol-
low. The treatises on celestial globes are silent
on the topic of constructing metal spheres. In
the astronomical literature only one passage has
been found that describes the production of a
metal sphere.^^ This description occurs in the
same treatise on the spherical astrolabe compiled
by Isaac ben Sid in the thirteenth century for the
Libros del Saber, where he states that a metal
sphere is not as desirable as a wooden one.®^
  In this particular procedure the hollow globe
is to be made by joining together two cast hollow
hemispheres. Instructions are given to make a
mold from a block of hard stone, which obviously
must withstand the melting point of alloys,
though this is not explicitly stated in the text.
The stone block should be larger in breadth than
the diameter of the desired globe and have a
height greater than the radius. The maker marks
the surface with a circle equal to the circumfer-
ence of the desired globe and then hollows it out
(again, the method for doing so is not specified),
making it a hollow hemisphere in which one can
fit and turn a "form of iron." This was a device
for determining when the hemispherical cavity

had the correct size and curvature, and it con-
sisted of an iron half-circular plate whose radius
is that of the desired globe and that has a handle
attached to the flat edge by which it can be held.
Then a solid hemispheric mold of clay {tierra) is
made whose radius is somewhat smaller than the
hemispheric cavity in the stone and equal to the
inner radius of the finished globe. The maker
then pours the molten alloy into the space be-
tween the clay and stone molds being careful to
fill it to the upper edge of the stone mold; no
directions are given for devices to keep the clay
mold from falling against the stone mold. When
the alloy has cooled the clay inner mold is re-
moved and the completed metal hemisphere re-
moved from the stone mold. By using the "form
of iron" and a hammer, any roughness can be
removed from the finished casting. The edge of
the completed hemispheric shell should then be
carefully filed until smooth and even. After a
second hemisphere is made in the same manner,
a long, thin metal strip of nearly the same length
as the circumference of the globe is cut and
soldered around the inside edge of one of the
hemispheres so that half of its width protrudes
above the edge of the hemisphere. When the
solder is set, additional solder is applied to the
other half of the metal strip and the other hem-
isphere set down over it. In this way the hemi-
sphere should be held firmly together by the
metal band soldered to their inside edges, with
no space showing between the edges.
  There are roughly 41 extant globes clearly
made of hollow metal hemispheres that have
been soldered together or perhaps joined by an
internal band as suggested in the Libros del Saber.
On some the soldering is obvious and mars the
surface of the globe itself (see No. 7 in Figure 9),
while on others it is so neatly executed that the
seam is scarcely visible (see No. 81, Figure 22).
  The seamless, hollow metal globe is never
mentioned in the literature. At least 49 such
globes are known, and most are products of
Mughal or nineteenth-century India, although
there are three possible examples that are earlier
and from outside of northwestern India. All such
globes are without any apparent seam and have
  
84

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



at least one fairly large and several smaller visible
plugs, although on some globes they are difficult
to locate (see Figures 11, 12, 18, 25 and 26 for
examples of plugs). Suggested methods of con-
structing hollow metal seamless globes will be
discussed below.
  Once the sphere itself was constructed, its
surface was made as even and smooth as possible
by turning on a lathe. In the ninth century al-

BattanT specified that it should be turned on a
lathe. Many of the extant globes show clear signs
of having been polished by this method (see
Figure 35) and display the circularity and spher-
icity characteristic of all carefully turned work.^"*
The extant globes of wood or papier mache over
wooden cores do not seem to have been as fre-
quently shaped on a lathe (see No. 36, Figure
19; and No. 56, Figure 29), though globes No.


FIGURE 35.—Detail showing signature of maker. Globe No. 27, dated 1024 H/AD 1663-1664,
by Diya^ al-DTn Muhammad of the Lahore workshop. Arrow indicates plug. Royal Scottish
Museum, Department of Technology. (Photo: The Royal Scottish Museum)

NUMBER 46

85



72 and 76 (illustrated in Figure 28) display better
sphericity than the other wooden or papier mache
globes.
  The maker then proceeded to place upon it
the necessary circles, graduations, stars, constel-
lations, and labels as required by his design.
According to the procedure outlined by al-Bat-
tanT in the ninth century, the maker probably
began with drawing or incising on the surface
two great circles at right angles to one another,
which would represent the celestial equator and
solstitial colure.®^ The maker probably divided
the equator into four 90 ° quadrants and gradu-
ated and labeled each in order to obtain the scale
necessary for the placement of the remaining
items. The celestial equator was usually repre-
sented by a band which was about 5°-6° wide,
one edge of which would represent the line of
the equator itself. At an angle of approximately
23V2° the great circle representing the ecliptic
would be drawn with a lesser parallel circle about
5°-6° away from the line of the ecliptic. The
ecliptic would then be graduated in the same
scale as that used for the equator. Al-BattanT
specified that the star positions were indicated
next and that only after the stars were positioned
were the five great circles then to be drawn at
right angles to the ecliptic at 30° intervals (the
sixth great circle being already represented by
the solstitial colure). Indeed on some globes (as
on the Smithsonian globe. No. 38), it is evident
from the engraving that the ecliptic latitude cir-
cles were engraved after the stars were posi-
tioned, for the engraving of the lines runs
through some of the stars and other engraving.
Lesser circles could then be indicated if the de-
sign called for it; sometimes the lesser circles
must have been added after the stars and con-
stellations were placed on the globe, as was the
case with the Smithsonian globe where the tropic
is only partially drawn since it was incorrectly
positioned, probably causing the maker to stop
work on a nearly finished globe.
  The treatise on the spherical astrolabe in-
cluded in the thirteenth-century compilation Li-
bros del Saber again provides us with some inter-
esting information on the equipment and tech-

niques used to inscribe the circles on a sphere.®^
The compiler described a round compass {compas
redondo) used for this purpose consisting of a
slender band of iron in the form of a 90° arc
graduated by single degrees, whose inside diam-
eter is concentric with the outside surface of the
sphere. Thus for each globe of the same size
there would be a special round compass. A pivot
is attached at one end and at the other a move-
able marker so as to inscribe all possible circles
on the globe. When set at 90° the round compass
could be pivoted about the stationary point to
inscribe the great circles and when set in other
positions could mark the lesser circles.
  Also writing in the thirteenth century, the
North African astronomer al-MarrakushT gives a
method for inscribing great circles at right angles
to one another on a sphere. A point is chosen on
the sphere, and with it as center a small circle is
inscribed around it with a drawing compass. The
circle is divided into four equal parts. Then a
graduated quadrant of copper {nuhds) especially
made to fit that globe is employed. Al-Marraku-
shT had earlier given a means using drawing
compasses for determining the diameter of a
sphere once the sphere is made, which was then
used to make the meridian and horizon rings and
the graduated quadrant. The quadrant was to be
placed with one end at the center of the circle
on the sphere and should pass through one of
the points marking the equal divisions of the
circle. The other end of the quadrant would
mark one pole. The quadrant was placed in a
similar fashion over the opposite point on the
small circle, to mark a second pole. Then using
these two points as poles, a great circle is in-
scribed passing through the center of the small
circle and the two remaining points on the small
circle. Al-MarrakushT says that to inscribe the
circle "it is best that you draw it with the instru-
ment of the turning-lathe {bi-dlat al-khart)."^^
What is intended by this phrase is not entirely
clear, but possibly the sphere would be attached
to a lathe at the two poles. Then as the sphere
rotated in the lathe, a line could be incised by
holding the incising instrument, beginning at one
of the two points on the small circle.
  
86

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



  After the first great circle is marked on the
sphere, then a second one at right angles to it is
determined in a similar fashion, placing the quad-
rant along the other two points on the small
circle. Then the ecliptic is to be drawn, but al-
MarrakushT gives no value for the obliquity of
the ecliptic nor directions for how to place it
accurately on the sphere. Al-BattanT in his early
treatise supplies more specific information on the
determination of the position of the ecliptic (see
Chapter 1).
  Unfortunately no further information is sup-
plied in written sources about techniques em-
ployed in drawing firmly and evenly on spheres,
even though we have many details about the
practices in Safavid Persia by which the astrolab-
ists were able to incise carefully executed lines
on planispheric astrolabes.^^
  On a globe being prepared by or for an as-
tronomer-astrologer (besides Class C globes), the
star positions would no doubt have been indi-
cated next, after the great circles on the globe,
using dividers to measure the ecliptic coordinates
in terms of the equator's scale and to transfer
these distances to obtain the star positions on the
globe. Al-BattanT again is quite specific about
how to determine the star positions accurately
on a sphere using dividers or a drawing compass
(see Chapter 1 and Figure 4). Some makers prob-
ably employed a graduated and labeled 90° arc
concentric with the surface of the globe whose
base could be placed along the ecliptic and the
appropriate longtidue and latitude marked on
the globe. Others might have used the 90° grad-
uated arc having a long pin at one end which
could be inserted in the ecliptic poles and rotated
about, if holes had been drilled at the ecliptic
poles.^^ The latter method would have allowed
for greater precision in the celestial longitudes
of the stars at a distance from the ecliptic. Al-
MarrakushT stated that the sphere was to be
mounted in the horizon and meridian ring assem-
bly before the star positions were determined.
Holes were to be bored at both the ecliptic and
equatorial poles, so that by mounting the sphere
in the meridian ring at the ecliptic poles the
graduated rings could be used to position the

stars. Then the globe would be mounted at the
celestial poles when ready for use.
  The coordinates themselves were obtained,
according to the inscriptions on some of the
globes, from three star catalogs: that of Ptolemy
(compiled for AD 138); that of ''Abd al-Rahman
al-SQfi prepared for the epoch 364 H/AD 974,
and that of Ulugh Beg made for 841 H/AD 1437-
1438. It is possible, of course, that other star
catalogs were employed by some makers."
  Because of the precession of the equinoxes,
the globe maker needed to increase by a constant
the longitudes given in the star catalog he had
chosen to use. Various constants were employed
for the precession: Ptolemy had used 1 ° for every
100 years. Al-SQfi, when changing Ptolemy's lon-
gitudes, employed a value of 1 ° per 66 years,
while the thirteenth-century astronomer NasTr
al-DTn al-TQsT used 1 ° per 70 years.'^
  On eight of the extant Islamicate globes the
precise increment is given as well as the catalog,
so we know precisely what value that particular
maker was employing. In 539 H/AD 1144-1145
YQnus ibn al-Husayn al-AsturlabT (globe No. 2)
based his star positions on the Ptolemaic catalog,
increasing the longitudes by 15° 18' which gives
a value of 65.8 years per degree. In 622 H/AD
1225-1226 Qaysar ibn Abi al-Qasim ibn Musafir
al-Ashrafi al-Hanafi (globe No. 3) also used the
Ptolemaic catalog, with an increase of 16°46',
corresponding to 64.9 years per degree. Muham-
mad ibn Mahmud al-TabarT based his globe (No.
6), made in 684 H/AD 1285-1286, on the star
catalog of al-SQfi^^ with an increase of 5° in the
longtidues, giving a value of 1 ° every 62.4 years,
or 1 ° per 64 lunar years. This is a surprisingly
high ratio, and is yet another slightly irregular
feature of this particular globe (see catalog entry
and Figure 6). Ja"far ibn "Umar ibn Dawlatshah
al-KirmanT based his two globes (nos. 7 and 8) on
al-SQfi's catalog, as did his son Muhammad ibn
Ja"far (No. 62). JaTar stated on his globe, dated
764 H/AD 1362-1363, that he used an increment
of 6°3'; on the globe dated 785 H/AD 1383-
1384 the number of the minutes following the
6° is not legible. His son's globe is dated 834 H/
AD 1430-1431, with an increment of 7°5'. The
  
NUMBER 46

87



value employed by father and son is clearly the
same, 1 ° per 64.5 years, or 1 ° per 66.4 lunar
years. With this consistency we can determine
the missing amount on the second of Ja"far's
globes—that is, the increment should read
6°23'.
  Three globes are stated to be based on Ulugh
Begs star catalog. The most recent of the three
is that made in 1112-1113 H/AD 1701 by Ridwan
(No. 31), who does not state an increment in the
longitudes over that given by Ulugh Beg. A globe
by Qa^'im Muhammad ibn "Isa ibn Allahdad As-
turlabT LahQrT HumayQnT (No. 14), made in 1047
H/AD 1637-1638, is stated to be increased by 3 °,
giving a value of 1 ° for every 66,6 years and 1 °
per 68.7 lunar years. The third globe, which is
anonymous (No, 65) was made in 1056 H/AD
1646-1647 and bears an inscription saying the
longitudes were increased by 4° over those of
the Samarqand observatory. This increment
gives a value of 1° per 52,25 years (or 53.75
lunar years), which must surely indicate an error
on the part of the maker.
  A globe made for a given epoch could be used
with a fair amount of accuracy for the better part
of a century, which was also the period of use-
fulness of an astrolabe. The astronomical phen-
omonon of precession provides a convenient way
to approximate the epoch for which an undated
globe having stars was made.'^ For this purpose
the longitudes of the stars along or very close to
the ecliptic can be measured and compared with
the same stars on a well-made dated globe or
with the star positions in a star catalog. The stars
close to the ecliptic are easier to position accu-
rately than those further from the ecliptic. It is
indeed difficult to date an object to within less
than 50 years by this method, for the limitations
of such dating are considerable. Many extant
globes have irregular graduations (see Table 3 in
Chapter 7), and since the graduations of the
equator and ecliptic served as the scale for the
positioning of its stars, great accuracy or consist-
ency could not be achieved when the scale itself
was irregular. On globes that are precisely grad-
uated, all but two have one degree as the smallest
interval. The Smithsonian globe (No. 38) has

one-half degree graduations and is so well exe-
cuted that it is one of the finest examples of
precision in this field. One other globe (No, 53)
has the ecliptic but not the equator graduated by
half-degrees, but it is not available for study.
When a maker states on a globe that the increase
in star positions was 16°46' or 6°5', it by no
means implies that he could attain such precision
on his globe. The fact that only two globes have
graduations smaller than 1 ° and none of the
globes is of sufficient size to cause the degree
markings to be widely spaced, makes it quite
impossible for star positions to be approximated
within less than half a degree. A statement on a
globe giving more precise increments in star
positions must be interpreted as indicating that
the star catalog from which the maker obtained
his coordinates was based on such an increase,
for the maker of the instrument would be forced
to round off all coordinates to the nearest half-
degree or, in the case of many globes, to the
nearest degree.
  Even granting that a globe can be accurate to
within half a degree, in the vast majority of cases
we do not know the value for precession a given
maker would employ, or what catalog we had at
his disposal. We cannot work backward from
modern values, for such an approach is unhistor-
ical and completely distorts the interpretation.
One approach that might prove productive is to
compare a series of globes, using some as stan-
dards for comparison. In the case of the seven-
teenth-century Lahore workshop, which gener-
ally executed impressively precise globes, the star
positions are identical on all their products, which
among the dated objects range from AD 1622 to
1680. Furthermore, some of the anonymous
items, such as the Smithsonian globe (No. 38),
have these identical star positions. So we can
conclude that all members of this workshop em-
ployed the same set of star positions no matter
when they were producing the actual object.
  On carefully constructed Class A globes after
the stars had been indicated (on metal globes
most frequently by inlaid silver dots) the constel-
lation outlines would be engraved around the
stars based on illustrations such as those in al-
  
SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

SQfi's treatise on the fixed stars, but with dress
and details adapted to local artistic customs.
There is indication that on some globes the eclip-
tic latitude circles and the lesser circles were
added last. On Class B globes, of course, the
globes were completed once the stars had been
indicated and labeled. It is unlikely that the holes
of the celestial poles (and ecliptic poles when
drilled) were made until after all the circles were
drawn on the surface of the globe.
  On certain Class A globes that have poorly
positioned stars it appears that the constellation
outlines were engraved first and the stars then
placed within and around these outlines, for the
stars sometimes are on top of the lines of engrav-
ing. Such globes were probably made by artisans
who were not themselves astronomers or profes-
sional instrument makers. They may well not
have been familiar with the star catalogs and the
methods for measuring the coordinates on the
surface of the globe, and so executed the con-
stellation figures first and determined the posi-
tion of the stars in relation to these figures rather
than to any coordinate system. Such globes, of
course, are sufficiently inaccurate to render them
useless for astronomical calculations, though oc-
casionally they are exceptionally attractive.
  On the methods of constructing the rings and
stand for a celestial globe the literature has little
to say. All but one extant set of rings are made
of metal, probably cast, and then turned on a
lathe for finishing. The one exception is the set
of rings for No. 76 (see Figure 28), which is
wooden and clearly contemporary with the
globe.
  In the ninth century al-BattanT described an
elaborate set of rings that includes both moveable
and stationary meridian rings and a gnomon (see
Figure 5). He explains'^ that you must take two
rings of copper {nuhds) the same size, of smooth
surface and true circularity. The inside diameter
of each is to be equal to the diameter of the globe
(these rings will be numbers 1 and 2 in Figure
5). Then two rings are to be prepared that fit
inside one another—that is, the inside diameter
of the smaller will be the diameter of the globe
and its width will be one-third the width of the

first two rings; the inside diameter of the larger
ring will equal the outside diameter of the smaller
ring and its width will be two-thirds that of the
first two rings. These two rings meshed together
(numbers 3 and 4 in Figure 5) will be the size of
one of the first two rings. A fifth rings is to be
made whose inside diameter equals the outside
diameter of the first four rings, for this ring will
fit around the other four (number 5 in Figure
5). All the rings are to be divided into four equal
quadrants, with each quadrant divided into 18
divisions of 5 ° each and labeled in abjad numer-
als, numbering each quadrant of 90° so that the
90° labels occur at two opposite places on the
ring (that is, in non-consecutive segments of 90°).
The smaller of the two rings that nest together
(number 3) is to be perforated (at the zero points)
and at these points is to be affixed to the equa-
torial poles of the globe and fastened from out-
side the ring with filed nails. The points of at-
tachment on the ring are to be labeled "pole of
the north" and "pole of the south." The larger
of the two rings to be nested (number 4) is to
have notches cut on the outside end at the zero-
points; the notches are to be one-half its width
and the thickness of the horizon ring. Ring 3, to
which the globe is attached is then placed inside
ring 4 and the two are encompassed by the
horizon ring (number 1), which has notches equal
to two-thirds its width on its inside edge at the
90° labels, marked North-South, There are also
to be notches on the outside edge of the horizon
ring at the zero-points, labeled East-West. The
ring number 2, the zenith ring, is also to have
four notches, two opposite ones equal to the two
outside notches of the horizon ring, and the
other two equal to the height of the small ring
(number 3) and one-half the height of number
4. Al-BattanT does not specify notches having
been made on the outside edge of ring 4, but he
must have intended them. The zenith ring (num-
ber 3) then overlays the horizon ring and the two
meridian rings. The largest ring, number 5, is
apparently to have a slot made in it in the middle
of its thickness along one of its upper quadrants
(the text is slightly obscure here). Al-BattanT says
that "we make a piece of four-sided {murabba"ah)

NUMBER 46

89



copper the size of the thickness of the ring and
the width of the slot. We incise in its middle a
straight line cutting it into two equal halves; we
file along the two sides of this line with a file {bi-
l-mibrad) and make it small at the lower end of
the piece like a rounded file."'^ The lower end
touches the surface of the sphere and the upper
end extends up through the slot in the ring.
  The large ring is then fixed to the poles of the
meridian and zenith rings by pins. The upper
part of the outside rings has a suspensory device
of two rings {"urwah and halqah) like that found
on an astrolabe {dhdt al-safd"'ih). Al-BattanT then
concludes his discussion by telling the reader how
to use this moveable gnomon {muwiri) to deter-
mine the altitude of the sun. His special design
of a two-part meridian ring, the inner part move-
able and the outer stationary, combines the ad-
vantages of both forms of rings. Such a form of
meridian ring is not known to occur among ex-
tant globes, although there may once have been
a similar gnomon ring on globe No. 5.
  The thirteenth-century astronomer Abu "AIT
al-Hasan ibn "AIT "Umar al-MarrakushT (died AD
1261) includes an interesting section on the de-
sign and construction of the rings for a celestial
globe in his treatise on astronomical instru-
ments.^^ He begins by explaining a technique for
determining the diameter of a given globe by
means of a geometrical construction on a plane
surface using only dividers and drawing com-
passes. Once the diameter is determined, he says
to make two metal rings each of inner diameter
equal to that of the globe. One of the two rings
(the one to be the horizon ring) should be twice
the width of the other. One face of the horizon
ring should be divided into four equal parts and
each of these into 90°, which are then labeled in
such a way that there are two zero points dia-
metrically opposite labeled East and West, and
with the numbers converging at two opposite
points of the ring that are labeled North and
South. Indentations are than made at the north-
south points, equal in width and breadth to the
meridian ring,
  A semicircle of metal is then attached firmly
at right angles to the underside of the horizon

ring at the east-west points. In the middle of this
arc a notch is made to hold the meridian ring
and prevent it from wobbling. Three columns of
metal slightly higher than the radius of the globe
are fixed firmly to the underside of the horizon
ring at equal distances from one another. One
face of the meridian ring (that which will face
east when placed in the horizon ring) is then
divided into four parts, each graduated into 90°
by lines extending to the outer edge of the ring.
Holes are then drilled through the ring at every
degree around half of the ring (probably a 180°
segment between the two points where it will be
attached to the poles of the globe). Then the
maker is instructed to produce a mdsikah (tongue)
of metal, which is attached by a pin to the upper
face of the horizon ring near the south point so
that it turns around the nail in order that the
pointed end can be inserted in any one of the
holes of the meridian ring. In this way the merid-
ian ring can be held stationary and secure at any
desired terrestrial latitude. Then al-MarrakushT
directs the maker to file down and bevel the face
of the meridian ring that is not graduated so that
it slants in toward the inner edge; by this wedge
shape al-MarrakushT feels that as much as possible
of the surface of the globe will be exposed to
view.
  Two grooves are to be cut into the meridian
ring, but not going through it, beginning at the
inside edge of the ring. They should be placed
precisely opposite each other and positioned on
the meridian ring so that the middle of each
marks one end of the 180° segment of the me-
ridian ring where holes had earlier been drilled
for each degree. Into each of these grooves there
is to be inserted a metal unbubah, which is pre-
sumably a hollow pin or cap attached to each of
the celestial poles of the globe.
  None of the extant Islamicate globes reflects
al-MarrakushT's particular design of the meridian
ring with its beveled shape and holes placed at
every degree for 180° through which a "tongue"
could be slipped to set it for a particular geo-
graphical latitude, and its grooves in which the
axis of the globe are set. His description reflects
a theoretical rather than a practical interest in
  
90

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



globe making, as witnessed by his mathematical
determination of the diameter of the globe. He
never mentions the tools and processes necessary
for constructing such rings and appears to have
had little practical experience himself, since it
would be virtually impossible to drill a hole
through the meridian ring at every degree unless
he were making a globe and ring of monumental
size.
  Although the extant rings do not correspond
in detail to what could be produced by the meth-
ods described above, it is obvious that procedures
were used that required considerable skill, special
equipment, and techniques for executing the
rings to the desired size and for measuring and
marking accurate graduations. No doubt most
metalworking techniques, even for intricate sci-
entific instruments, were taught in workshops
and through an apprentice system, for which
reason the artifacts themselves must serve as our
major source of information as to their design
and construction. From these artifacts sufficient
information is obtained to warrant considerable
admiration for the makers. This applies particu-
larly to the matter of the spheres themselves, in
which case the entire method of casting hollow
seamless globes is not mentioned in the litera-
ture, but the existence of such a process has been
confirmed beyond doubt by laboratory analysis.
  Since there is no mention or description of
casting hollow seamless spheres in the literature,
we can only conjecture as to the methods em-
ployed, relying on some laboratory analyses and
comparisons with other metal-working tech-
niques. There are no metal-workers today in any
part of the world who are known to cast such
spheres. While lost wax casting is a very ancient
technique that has been used for centuries to
produce elaborate and intricate pieces of metal-
work, the casting of a complete sphere presents
special problems for a metal-worker. These Is-
lamicate celestial globes are the only examples
known to exist of hollow cast spheres.
  Throughout this study and in the following
discussion the author has chosen to employ the
vague terms metal or alloy rather than using
more specific terms such as brass or bronze. The

term brass usually denotes a binary alloy of cop-
per and zinc, generally two parts copper to one
part zinc, with sometimes small amounts of other
metals present, while bronze denotes a binary
alloy of copper and tin with only small propor-
tions of other elements such as zinc or phospho-
rus and is generally harder and stronger than
brass. The actual composition of copper alloys
varied enormously from one area to another and
one maker to another, and it is unlikely either of
the terms adequately covers the composition of
most alloys employed in these globes. Further-
more, the one globe that has been analysed was
made of what has been termed a quaternary
alloy'^ which has a very high proportion of lead,
and if our conjectures are correct more than one-
third of the extant globes were made with such
an alloy. It is fairly certain that none of the extant
globes are made of pure, or nearly pure, copper.
The vast majority of the metal globes have a
golden cast to the alloy, although a few are very
dark brown or nearly black,^*^ and one (No, 111)
appears to be a silvered alloy.
  A metal globe in hemispheres could be made in
one of four ways: (1) In the sheet metal
technique®' the prepared alloy is first heated and
hammered down to sheets of desired thickness
(for the walls of the globe). Then a piece of the
sheet metal would be hammered omnidirection-
ally over hemispherical anvils to take up the form
of the required hemisphere. The two hemi-
spheres would then be soldered together. (2) In
the wrought metal technique®^ the prepared alloy
ingot would be formed into the final shape of a
hemisphere by systematic hammering over an
anvil, using a team of three or more men working
in a rhythmic sequence. The technique essen-
tially eliminates the sheet metal stage. It is a very
uncommon method today, but was commonly
employed until the end of the nineteenth cen-
tury, (3) In box-mould casting each hemisphere
would be cast separately in a fairly simple box
mould and then soldered together. (4) In lost
wax casting each hemisphere could be produced
by using a lost-wax mould, but it is an unlikely
method to have been used in this case. It is
unnecessarily elaborate for what is the relatively
  
NUMBER 46

91



simple form of a hemisphere.
  There is no way to tell from the outside ap-
pearance of a globe whether it was made of cast
or raised (that is, hammered) hemispheres, if the
globe is well done and turned and polished
properly. The walls, however, tend to be thicker
and the globe heavier when made from cast
hemispheres than from raised ones. The deter-
mination of the exact method requires examina-
tion of the inside for stress marks and confirma-
tion by radiographic analysis.
  When we turn to seamless globes, visual inspec-
tion is even less reliable as the means of deter-
mining the construction. A seam is not always
visible on a very fine piece of metalwork, even
under inspection with a strong lens. For example,
globe No. 81 illustrated in Figure 22 does have
a seam, but it is not observable outside even
when examined in great detail. When the globe
is turned over and over, you can detect a very
slight difference in sphericity when compared
with the more uniform sphericity of a seamless
cast one such as the Smithsonian Institution
globe, No. 38. Most importantly, no circular or
square plugs can be detected, such as are visible
on the globes in Figures 11, 12, 26, and 35. The
lack of such plugs should suggest that the globe
may well in fact have a seam, although there are
some globes that were coated by the maker with
different substances which cover most of the
traces of such plugs. If no plugs are to be seen,
one method, of ascertaining whether it has a
seam is to insert a thick wire or probe through
one of the axis holes to see if the ridge from a
seam can be felt inside the globe, as is the case
with No, 81 (Figure 22). In the case of a seamless
globe there are often protruberances or rather
short stumps inside the globe; thus one must be
careful to distinguish these from a continuous
ridge indicating a seam. In some cases it is not
possible to decide definitely even with a pliable
probe, and most globes do not admit much visual
inspection without special equipment. Occasion-
ally a globe will be missing a large plug (see
Figure 36) and a person will be able to see readily
that there is no seam. The plugs are not repairs,
but are part of the construction process. Since

plugs are frequently not visible to the naked eye
or even with the aid of a lens, one must be very
careful when asserting that a globe has a seam or
not.
  In Table 6 (Chapter 7) the 50 globes that have
been listed as hollow metal seamless globes all
have visible plugs from the casting process. The
41 classified as metal globes made in hemispheres
all have either a clearly visible seam on the out-
side or one that can be detected by a probe on
the inside, leaving 26 metal globes that could not
at this point be classified. The most certain
method of confirming the construction is the use
of radiographic analysis.
  Assuming that you can demonstrate that there
is no seam in a hollow metal globe, there remain
two possible methods of construction. The first
is called raising,®^ This involves taking a sheet of
alloy and hammering it omnidirectionally over
various spherical anvils until a nearly complete
sphere was attained. The large plug that is fre-
quently visible on seamless globes (see Figure 12)
and which would have filled the hole evident in
Figure 36, could be interpreted as a plug placed
over the final opening; the smaller plugs such as
those in Figure 35 could be seen as filling other
holes made in the sphere for the insertion of
raising tools, necessary so that the outer surface
could be beaten with a hammer. An argument in
favor of such a method being used is the thin
metal cross-section of the globes, which appears
to modern metal-workers upon first considera-
tion to be much too thin for casting,®^ The inte-
rior surfaces of all the globes so far inspected
have revealed the rough and pitted texture char-
acteristic of casting rather than the stress patterns
visible in a raised object. I know of no hollow
metal seamless globes that have in fact been
raised, but admittedly only a small number have
actually been inspected.*^^
  The second possible method would be casting
by the lost wax method.®^ This method would
also require that the finished product have sev-
eral plugs, one of considerable size, and would
produce relatively thin walls. Such a method
would result in the interior surface being rough
and pitted with stumps projecting inside, and
  
92

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 36.—Detail showing the area where the plug has been removed. Globe No. 12, dated
1035 H/AD 1625-1626, by Qa^im Muhammad of the Lahore workshop. Paris, private collec-
tion. (Photo: Alain Brieux)

NUMBER 46

93



would probably result in noticeably better spher-
icity than would be obtainable by raising. Close
inspection of the inside surface can be important
in determining whether a seamless sphere has
been raised or cast.
87
  The Conservation and Analytic Laboratory of
the National Museum of American History at the
Smithsonian Institution, undertook an analysis of
an unsigned and undated seamless metal globe
that can definitely be attributed to Qa^im Mu-
hammad of the seventeenth-century Lahore
workshop (see Chapters 3 and 4), This globe
would have been the most precise of all the globes
made by that prolific workshop had not the
maker at the last moment made a mistake when
engraving one of the tropics and ceased work on
the object, since it is impossible to correct an
error on metalwork. The laboratory employed
radiography, metal analyses, and hundreds of
photographs of the inside taken with a tiny cam-
era.
  The alloy was found to have copper as the
major component, but it also contained over 30%
lead, plus zinc, tin, and traces of other metals.
The high lead content is in keeping with recent
analyses of Persian items (which did not include
any globes) of the twelfth through the sixteenth
centuries,®® These analyses found that a substan-
tial group of cast items (buckets, bowls, mortars,
etc) were made of an alloy containing from 12%
to over 40% lead. The precise measurement of
the lead is difficult on lead-rich objects because
of the segregation of lead during the cooling of
the object. This is a quaternary alloy, for its
components are, in descending order, copper,
lead, zinc, and tin (Cu > Pb > Zn > Sn), with a
trace of other elements, and with the lead always
over 12%.
  The reasons for employing such a lead-rich
alloy in casting would be the resulting lower
casting temperature and the greater fluidity of
the alloy, which would be quite important in
filling a spherical mould. Moreover, greater
malleability is obtained, which is particularly use-
ful if decoration is to be engraved on the object.^^
  The laboratory found the walls of the sphere
to be unexpectedly thin (0.47-1.0 mm) in rela-

tion to the diameter (217 mm) of the sphere and
in relation to the weight (2976.7 gr). Again, the
lead-rich alloy would account for the unexpected
weight.^"
  On the basis of a recently published anthro-
pological survey of metalworking techniques and
workshops in India^' combined with the results
of the laboratory analysis conducted at the Smith-
sonian, we can conjecture a fairly detailed picture
of the procedure that might have been followed
in manufacturing such a globe by lost-wax cast-
ing.
  The first step would be to form a center spher-
ical mould, which might be either solid or hollow.
A solid center mould would have been quite
heavy, but would hold more firmly on the lathe.
If hollow, it could have been built up by using
ropes or coils of clay and building and shaping it
into a sphere much as is done in making clay
bowls. If solid it might be formed from earth
with cow or goat dung (presoaking the dung in
water and grinding it in a mortar and pestle and
then grinding it with equal amounts of clay); or
with equal amounts of sticky clay, sandy clay, and
old cloth; or with river clay and bran.^^ The
spherical mould thus formed would be allowed
to dry, and a hole would be bored in it by which
it could be placed on the turning-lathe and
turned until quite smooth and spherical.
  A wax wrap would then be applied to the
center mould. Beeswax is formed into wires by
forcing it through a sieve so that each filament
has the same thickness. The wires of wax are
then neatly wrapped around the core and
smoothed with a knife. The center mould with
wax wrap is then turned on the lathe, producing
a wax covering about 0.5-2.0 mm thick, with an
opening where it had been inserted on the lathe.
No design is put into the wax, for in the case of
celestial globes all decoration is engraved later
onto the finished cast product.
  An outside investment of clay is then necessary
to complete the mould, but before that can be
done several preliminary steps must be com-
pleted. Channels must be added to the wax wrap
to serve as runners (also called feeders). To do
this, wax wires could be attached to the wax wrap
  
94

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



(and then coated with the outside clay mould
except for their upper ends). These melt when
heated and leave hollow channels connecting
with the then open wax space where the wax
wrap had been.^^ Bamboo sticks could also be
placed just touching the wax wrap. These would
be coated with the outside clay and when the clay
was dry, removed to leave a hollow tube. The
placement of these channels is very important,
particularly in the production of a sphere.
  Armatures (also called chaplets) must also be
inserted. When the wax runs out of the mould
the two surfaces of the mould would fall together
unless prevented by some supports. To prevent
this the maker must, before the outer clay coat
is applied, insert pins or nails, driving them
through the wax and into the center core. These
will be imbedded in the outer clay mould. If the
worker is skilled, the nails will be of the same
alloy as the main casting. The armatures must be
positioned evenly over the surface, up to twenty
or so in number, depending on the size of the
casting.
  Then the wax wrap, and runners and arma-
tures, must be covered with clay. This must be
done soon after the wax wrap is formed or it will
harden and crack. The outside clay investment
most frequently employed today in India is made
of sandy river soil moistened and ground to a
fine paste. If an especially smooth surface is re-
quired (which would probably be the case with a
globe) the earth is mixed with cow dung. The
paste is applied over the wax wrap until it is
about 10 mm thick. When this has dried, the
mould is then recoated with common earth and
bran (or clay, goat dung, and sand, or a mixture
of ant-hill dirt and rice husk) until the coat is
about 25 mm thick. At this point the bore in the
center core by which it had been turned on the
lathe would need to be filled.
  The completed mould is then inverted and
baked. The wax runs out through the runners
and the main opening in the outside encasement.
The wax is usually recovered to be reused later.
  The actual casting of the object was probably
done by a different craftsman from the one who
made the mould, as is seen in contemporary

workshops in India. The main pouring point
(called a sprue) was probably the opening in the
wax wrap where it was inserted onto the lathe.
Possibly some of the runners were used as addi-
tional pouring points. Other runners would be
allowed to fill up as the mould was poured. The
alloy would continue to be poured until it rose
in the runners. The alloy contracts when cooling
to such an extent that runners containing extra
alloy are needed to fill the spaces. If there are an
insufficient number of runners, a pock-marked
area of porosity will result in the finished prod-
uct, which must then be filled with plugs. This
appears to be a major problem in casting a
sphere, judging from Smithsonian globe, which
has such an area.
  When the metal is cool, the outside mould is
broken off and the runners and armatures cut
away. To remove the inside mould, a circular
plug had to be cut out (usually 40 to 60 mm in
diameter and probably at the sprue). Through
this opening the inside was dug out, a most
laborious and tedious process. And through this
opening all work on the inside took place.
  The armatures in most such castings are
merely broken off and then polished down on
the outside, leaving a stump protruding on the
inside of the mould. This appears to have been
done with many of the armatures on the Smith-
sonian globe, but some also seem to have been
removed and plugs inserted, made of the same
batch of alloy as the body of the globe. In the
Smithsonian globe nearly all the plugs (except
the two largest ones one of which was the last to
be put into place) were actually hammered into
place from the inside and then smoothed and
polished on the outside, using no solder. The
laboratory analysis of the Smithsonian globe dis-
closed 74 plugs in the globe. The large round
plug where the sprue was and the last to be placed
in the globe is indicated by arrows in Figure 77.
It is rather difficult to detect on the globe unless
you already know where to look. Most of the
smaller plugs are equally difficult to see, if not
more so. On many globes this large round plug
is quite obvious (see Figure 12), Occasionally the
final plug has the shape of a figure-eight, as seen
  
NUMBER 46

95



in Figure 18. This final plug would have beveled
edges and be held in place with soft solder. There
are globes in which the large plug has been
removed and the beveled edges of the surround-
ing area of the globe can be seen (see Figure 36,
where the plug was pryed out of the globe). Some
sprue plugs are elongated ovals. There are some
globes (e.,g,, Nos, 23, 44, 85, and 87) that have
two large round plugs placed precisely opposite
each other on opposite sides of the sphere; why
they are so positioned is unknown. Some of the
globes are biased in weight (e.g., No. 23) proba-
bly because of a thick and heavy plug causing the
uneven distribution of weight. The globe illus-
trated in Figure 25 has the final plug made of an
entirely different alloy from the rest of the globe.
  Figures 51 and 54 show a weak area that
probably occurred because of poorly placed run-
ners. In this area there are 17 large plugs, 7 small
square ones, and 6 round ones, some of which
are indicated by arrows. Most of these were
hammered from the inside as is evident from the
mushroom shape of the interior surface of the
plugs. Rectangular or irregularly shaped plugs
are generally the result of patching an area where
the alloy contracted and was not properly filled
with the alloy in the runners. Similar ones are
indicated in Figures 6 and 35, The small circular
plugs usually indicate the location of armatures.
In Figure 11 two can be seen to be of a different
alloy, while the two of the same alloy as the globe
are nearly invisible. The globe illustrated in Fig-
ure 26 has several visible circular plugs, the larg-
est one probably being a patch rather than an
armature. This particular globe is missing its
large round plug at the sprue hole. One of the
circular armatures is visible in Figure 58 (marked
by the arrow) and a large one can be seen in
Figures 61 and 62.
  After the casting and plugging process was
completed the sphere would be turned on a lathe

and polished; after the engraving and placement
of the stars it was polished again (evidence of
polishing is visible in Figure 35). Scrubbing the
product with different compounds could pro-
duce various finishes. For example, a scrubbing
with tamarind water and sand produces a matte
finish. Perhaps some such compound was used to
produce the very dark, nearly black matte finish
on globe No. 28 (Figures 40, 43, and 47). Occa-
sionally globes (Nos. 100 and 115) have a black
or brown surface coloring that is not "fast" (it
rubs off when energetically scoured). It obscures
seam lines, plugs, and other evidence of construc-
tion. One method of obtaining such a surface
coloring is to heat the item and smear it with a
mixture of charred coconut fibre and mustard
oil. As the metal cools, the color sets.^^
  In the workshops there was probably much
division and specialization of labor, just as can be
seen in workshops today,^^ where several artisans
are involved in producing the basic item and
other craftsmen work on designing and engrav-
ing the finished product. The precision and the
uniformity in the sphericity of a seamless globe
cast by cire perdue is impressive. The workshops
in the Punjab and western Himalayas seem to
have specialized in this extraordinary technique
of producing celestial globes from the sixteenth
through the nineteenth centuries. Only pride of
craftsmanship could have justified the enormous
expenditure of time and effort that went into
such a technique, when much simpler methods
using hemispheres had long been available. The
products of these Indian workshops display ad-
mirable skill on the part of the artisans in forming
a nearly perfect sphere and inscribing it accu-
rately and precisely. One of the finest examples
of their products is the Smithsonian globe (No.
38), which is the particular subject of Chapters
3, 4, and 5.
  
     The Smithsonian Globe
3.  Description and Attribution
The Pleiades will mount up in security from the
  aggression of Capella and Aldebaran,
And the Milky Way will show forth as a she-camel
cleaving to her young, the twin stars of
  Ursa Minor drawing near to her Taurus.
The Lord of the heavens and earth will reveal Himself
  to us in His justice and compassion,
And the colonialists will acknowledge defeat, and
the lands will shine forth in prosperity.
Ma"rijf al-RusafT
Diwdn

  The anonymous and undated Islamicate celes-
tial globe now in the collection of the Smithson-
ian Institution, National Museum of American
History (Division of Physical Sciences, Inventory
No. 330,781) was acquired in 1974 from the
collection of Ernst Nagel of San Francisco. (The
globe was on loan to the Museum from 1972.)
This metal seamless globe has a diameter of 217
mm. The thickness of the wall varies from 0.47
mm to 1.00 mm. The globe weighs 2976.7
grams. It is the sixth largest extant Islamicate
globe, and is one of only two celestial globes
preserved today to be graduated by half-degrees.
For overall views see Figures 51, 73, 76, 82, 88,
and the cover.
  The globe has 48 constellation outlines, which
were engraved after the stars (represented by
silver points) were placed upon it. A small part
of the globe in the region of Serpens (see Figure
59), which probably contained three stars, is now
missing. Assuming this to be the case, the total
number of stars represented by the maker is
1019. This total included one star not in the
catalog of Ulugh Beg or al-SQfi, but described by
al-SQfi as the "overlooked star" in the tail of Ursa
Major, by which a man could test his vision.
Clearly this star was purposefully added to the
globe, for it bears the label al-suhd (the over-
looked one); this star occurs occasionally on other
globes as well (see Nos. 4, 5, and 6 of the Cata-
log).
  
In four constellations the maker inadvertently
placed an extra star (in Pisces, Lupus, Libra with
an extra external star, and Taurus with an extra
star in the Pleiades), and in four constellations
he omitted a star (Scorpio, Aquarius, Canis Ma-
jor, and Corona Australis). The positions of the
stars (their celestial longitude measured along
the ecliptic), correspond roughly to what they
would have been in the second quarter of the
seventeenth century allowing for an appropriate
increment in the longitudes. They are identical
to the star positions found on the signed globes
produced during the seventeenth century by the
Lahore workshop.
  The constellations as well as many of the stars
are labeled in Arabic, written in naskhi script in
a manner suggesting an Eastern hand, though it
is not the nasta^'liq used in Persia. The maker
misspells 21 out of 48 times the word for con-
stellation {surah), writing it with a terminal form
of td"' rather than a td"' marbutah. (A similar
mixture of spelling occurs on globes Nos. 11, 13,
18, 20, 22, 23, 27, 28, and 30.) Some elementary
grammatical errors and irregular genitive con-
structions are evident in the Arabic, such as surat
"uqdb for the more acceptable siirat al-"uqdb (the
constellation of the eagle), and surat hUt janubi
instead of surat al-hut al-janubi {the constellation
of the southern fish). These elementary errors
indicate that Arabic was probably not the native
tongue of the maker, while the script itself indi-
  
96

NUMBER 46

97



cates the globe came from the Eastern part of
the Islamic world.
  The style of dress worn by the human constel-
lation figures is clearly Mughal—that is, from
western India of the sixteenth and seventeenth
centuries. This style is shared by globes Nos. 11-
15, 18-24, 26-28, and 30, all of which were
produced by the Lahore workshop in the seven-
teenth century.
  Having made a general attribution on the basis
of the star positions and the general style of the
constellation figures, it is possible to go further
and make a specific attribution to a particular
maker. This can be done on the basis of two
approaches, each of which independently pro-
vides sufficient evidence to conclusively establish
the identity of the maker. The first argument is
based upon technical terminology and astronom-
ical features of the Smithsonial globe, and follows
immediately. The second is based on the icon-
ography, and is presented in Chapter 4 by An-
drea P,A. Belloli,
  In Ursa Major the two stars on the lower front
paw are labeled tatimmat qafzah al-thdlithah (the
complement of the third leap), while in Pegasus
the star in the northernmost hoof bears an un-
deciphered label, mrhlt al-faras. Furthermore, in
the constellation Perseus the star name written
in the sword held overhead is written as mu"tasam
al-thurayyd (the refuge of the Pleiades), which
must be a misspelling of mi"sam al-thurayyd (the
wrist of the Pleiades). These terms are not known
to occur in any star catalog nor on any globes or
astrolabes except the globes No. 11 and No. 13,
made by Qa^im Muhammad ibn "Isa ibn Allah-
dad AsturlabT LahQrT HumayQnT made in 1032
H/AD 1622-1623 and AD 1626-1627 (the
twenty-second year of the reign of JahangTr) and
illustrated in Figures 13,37,38, and 42. On both
of these globes the identical terms are very
clearly inscribed. The two other globes by Qa^im
Muhammad (Nos. 12 and 14) were not available
for detailed study, and it is not known if these
terms occur on them as well. These very unusual
terms appear to be unique characteristics of the
globes bearing the name of Qa^im Muhammad
and do not occur on any other globes, even those
made by or supervised by his son Diya" al-DTn

Muhammad or his nephew Hamid. Unfortu-
nately the globe by Qa^'im Muhammad's brother
Muhammad MuqTm (No. 15) was also unavailable
for close study, but from the available photo-
graph appears to be of a different style from
Qa'^im Muhammad's products.
  The positioning of the sequence of lunar man-
sions along the ecliptic as done on the Smithson-
ian globe is known to occur on only one other
globe, that being the earliest globe made by
Qa^im Muhammad (No. 11), On these two
globes, the first lunar mansion is placed at 13°
House of Aries with the twenty-eighth lunar
mansion at the vernal equinox, in contrast to the
usual arrangement, in which the first mansion is
at the vernal equinox and the second about 13°
House of Aries (see Chapter 5).'
  Moreover, the maker of the Smithsonian globe
has numbered the ecliptic continuously from the
vernal equinox. On all other globes but one the
graduations of the ecliptic are numbered in 30°
intervals. This one exception is again the earliest
globe by Qa^im Muhammad (No. 11) where he
labels the ecliptic continuously from the vernal
equinox as well as giving it a second set of num-
bers which divide it into 30° intervals. Finally,
the star positions on the Smithsonian globe are
identical to those on the globes by Qa^im Mu-
hammad that have been examined.
   In the subsequent chapter Andrea P.A. Belloli
also attributes the globe to Qa^im Muhammad,
based on the iconography of the workshop. For
these reasons this anonymous and undated globe
can with confidence be attributed to Qa^im Mu-
hammad ibn Tsa ibn Allahdad AsturlabT LahurT
HumayunT.
  The construction of the Smithsonian globe
appears to be like the other examples of the
Lahore workshop—hollow, metal, and seamless,
with several plugs visible. The Analytic and Con-
servation Laboratory of the National Museum of
American History of the Smithsonian Institution
undertook an extensive analysis of this globe,
involving radiography, metallography, and metal
analysis. It is to be hoped that this detailed
analysis will be published. For the moment it
must suffice to give some of the general conclu-
sions.
  
98

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



  The radiography proved convincingly that the
globe was not raised from a metal disc, but rather
was cast as a hollow, seamless globe by the cire
perdue method. Seventy-four plugs were found
in the globe, only a few of which are easily visible.
All are made of the same alloy as the rest of the
globe. The alloy was found to have copper as the
primary component, with over 30% lead, plus
zinc, tin, and traces of other metals, A large
round plug, rather difficult to see, is found at
the end of Eridanus (see Figure 77), This plug is
at the place where the sprue was located—that
is, where the alloy was poured into the mold, A
weak area occurred around Draco, Cepheus,
Hercules, and Serpens, where there is a small
part missing (see Figure 59), and there are in this
region 17 large plugs, 7 small square ones, and
6 round ones, the latter being the location of
armatures or chaplets which protrude on the
inside. Some of the plugs are indicated by arrows
in Figures 50, 52, and 54. Many of the plugs
were hammered into place from the inside. This
is evident from the mushroom shape of the in-
terior surface of the plugs. Soft solder was em-
ployed in securing the large plugs.
  The globe was turned on a lathe for smoothing
and polishing. The precision and the uniformity
in the sphericity of the globe is impressive. This
is generally true of seamless cast globes.
  The alloy employed was highly leaded, prob-
ably to increase the fluidity. The circles, gradu-
ations, figures, and writing are engraved, not
etched, and it is evident that the engraving of
the figures was done after the placement of the
stars, for some of the outlines are incised over
the stars. The silver studs used to represent the
stars were soldered in; they do not go all the way

through the wall of the globe.
  Although analytic tests have been conducted
on only this one globe, which we feel certain was
produced by Qa^im Muhammad of the seven-
teenth-century Lahore workshop, it is likely that
the other globes by the workshop were made by
the same cire perdue process, as were probably
other products of northwestern India. The three
apparently seamless globes that appear to be of
an earlier date and from outside of India (Nos.
6, 9 and 60 illustrated in Figures 6, 7, and 8)
were also likely to have been made by the lost-
wax method and of a similar heavily leaded alloy.
  Twenty-one celestial globes were produced by
this workshop and inscribed with the name of
the master craftsman and the date of completion.
The question naturally arises as to why this globe
was not similarly inscribed. It is quite probable
that the maker, after completing the tedious
construction of the sphere and the precise place-
ment of the stars, the drawing and graduation of
the great circles, delineation of the constellation
figures and labeling of the stars, began to draw
the equatorial tropic circles. In the northern
hemisphere of this globe a lesser circle parallel
to the equator but not in the position of the
Tropic of Cancer was partially drawn (see Fig-
ures 73 and 76). It would seem that having
started to draw the tropics, the maker discovered
that he had placed the northern circle too far
from the equator. For this reason the maker may
have stopped incising the circle and ceased work
altogether on the globe, for such an error in
engraving would be impossible to correct. This
may well be the reason the globe bears no mak-
er's signature and date, which were no doubt the
last items to be added to a globe.
  
4.  The Constellation Figures on the Smithsonian Globe
Andrea P.A. Belloli
Over the black dome of the sky
  the veil hangs deep,
Whereon the stars soft-cradled lie.
Like flowers asleep.
Ilyas Abii Shabakah
Night-presence

  The purpose of this chapter' is to examine the
Smithsonian celestial globe (No. 38 in the cata-
log) not as a technical device, but as an art
historical document. The interest of such an
object for the art historian lies most specifically,
of course, in the constellation figures on its sur-
face. Ideally, an analysis of these figures in stylis-
tic and iconographic terms can complement tech-
nical study by providing a way to suggest or verify
attributions of anonymous, undated objects to
particular metal-workers. Thus, the argument
that follows is presented in terms particular to
the discipline of art history, with a methodologi-
cal caveat. An attempt will be made to put aside
the aesthetically based terminology used by
scholars such as Wellesz in favor of a factual
examination of what were, first and foremost,
pure and simple diagrams. Our aim, then, is to
formulate a firm attribution and date for the
Smithsonian globe itself on the basis of stylistic
criteria objectively described and evaluated, just
as one would evaluate information gathered
from a written rather than a visual document.
Factors to be considered include the particular
manner in which the basic constellation dia-
grams, known from earlier globes and the al-SufT
manuscript tradition, were rendered in this in-
stance; the iconographic peculiarities of the fig-
ures and their relation to counterparts on earlier
Andrea P.A. Belloli, 1169 Lida Street, Pasadena, CA 91103.

and contemporary globes; and the significance
of contemporary factors relevant to the globe's
manufacture.
  Each constellation figure on the Smithsonian
globe is presented in simple outline form. The
use of such an outline to delineate each figure is
a feature common both to the illustrations to the
manuscripts of al-SQfi's Kitdb suwar al-kawdkib
(Book of the Constellations of the Stars) and to most
Islamic globes. The garments of the human fig-
ures on the Smithsonian globe have none of the
tiny, nervous folds and bunched undulations seen
in early al-SQfT illustrations such as those in Ox-
ford Bodleian MS Marsh 144 (dated 400 H/AD
1009-1010).^ Rather, the drawing of the figures
on the Smithsonian globe is clear and straight-
forward; groupings of parallel straight or curved
lines or hatchings to indicate surface texture or
pattern are kept to a bare minimum. Such deco-
rated areas include Virgo's wings, a cross-
hatched detail on Cassiopeia's throne, and the
striped or pleated kilt-like skirts worn by Her-
cules and Serpentarius, as well as the scaley bod-
ies of Pisces, Draco, and Serpens. Many of the
human figures, including Auriga, Hercules, Ser-
pentarius, Virgo, Orion, and Aquarius, have
been rendered with two left or two right feet,
while on some figures the hands have too many
or too few fingers (e.g., Virgo, Hercules, Cas-
siopeia, and Aquarius). Both Perseus's and Her-
cules's left arms, which are raised over their
  
99

100

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



heads, end in right hands.
  Almost all of the human figures have roughly
heart-shaped faces with broad, smooth fore-
heads; some have slightly pointed chins, while
others possess squared-off jaws. The one excep-
tion is Virgo, whose head is longer and more
truly rectangular in appearance, perhaps to ac-
commodate the two large stars that fall here on
her cheeks; although these are not labeled on the
globe, al-SQfT refers to them as no. 3 al-shimdli
min al-ithnin al-tdliyin li-humd fi al-wajh (the
northerly of the two consecutive stars on the
face) and no. 4 amyalhumd ila al-janub (the one
more inclined toward the south of the two stars).^
Most of the figures have prominent ears, some
(see Aquarius and Serpentarius) with the pointed
or curling lobes seen also on the animal figures
of Leo and Draco. The eyes of each figure are
almond-shaped, in some cases blank and in others
containing single, short hatch-lines to indicate
pupils. Simple, comma-like eyebrows and noses
formed by slightly curving lines ending in knobby
points and simple curls characterize all the fig-
ures' physiognomies except for those of Orion
and Virgo. The latter figures have eyebrows that
join with the bridges of their noses to form
continuous lines. Mouths are indicated by single
straight or semicircular hatch-marks. Male fig-
ures such as Hercules wear moustaches consisting
of two broader horizontal incisions between nose
and mouth, while the beards of Perseus, Ce-
pheus, and Centaurus are formed of short, par-
allel striations following the chin-line and culmi-
nating in a rough point.'* The hair of all the
figures, both male and female, was incised in
long, narrow, slightly curving, blade-like forms
growing out of the top or back of the head and
fanning out slightly at either side; in the case of
Hercules and Serpentarius, the leaf-like appear-
ance of these coiffures is enhanced by palmette
terminations and small knobs or curls.
  The rendition of human hair styles on the
Smithsonian globe is markedly different from
conventions observed in al-SufT manuscripts such
as Oxford Bodleian MS Marsh 144 and later
versions such as those in Leningrad (dated AD
1606) and Copenhagen (dated AD 1601-1602),^

where we are given a sense of contemporary
tastes in both male and female coiffures. The
maker of the Smithsonian globe rendered human
hair (and also certain details of animal fur) in a
manner seen in one of its earliest manifestations
on globe No. 5 (probably thirteenth to four-
teenth century AD),^ On the latter, many of the
figures' coiffures are represented as long, blade-
like bunchings as are the manes of Leo, Pegasus,
Ursa Major, and other animal figures. This con-
vention for animal figures may be pre-Islamic,
since it can be seen in Oxford Bodleian MS
Marsh 144 on figures such as Pegasus,' though
the hair of the human figures in that manuscript
was painted in large opaque areas of black ink.
  All of the human figures on the Smithsonian
globe wear simple tunics or short skirts, with the
exception of the Gemini, whose naked, rubbery
bodies are unarticulated save for digits on hands
and feet. The representation of the Gemini as
naked figures—usually of indeterminate sex—is
distinctly more common both on globes and in
al-SQfi illustrations than their portrayal as
clothed figures.^ The twins are usually seen from
the front in three-quarter view with arms out-
stretched and overlapping. The human upper
torsos of Sagittarius and Centaurus on the Smith-
sonian globe also appear to be naked, and are
unarticulated as well. Hercules and Serpentarius
are bare-legged, and wear short striped or
pleated kilt-like skirts; the outlines of Hercules's
arms end in half-circles on his chest.^ Both of
these male figures appear to be bare-chested,
though it may be that double or triple bands at
wrists, elbows, and upper arms were intended to
be read as drapery folds or decorative bands on
fabric, rather than as jewelry worn on bare skin.
  The supposition that these bands represent
sketchily rendered drapery folds is based on a
comparison of the garments of relevant figures
on the Smithsonian globe with figures on globes
of different dates whose clothing was finished off
in every detail. On globe No. 5, for example,
figures such as Auriga and Perseus wear ring-
like or bunched cuffs and tiraz bands,'° while
Andromeda on No. 26 (dated 1071 H/AD 1660-
1661) has parallel bands on her cuffs and lower
  
NUMBER 46

101



and upper sleeves." Serpentarius, however, was
often represented as a bare-chested figure; ex-
amples include the figure on globe No. 5 itself,
and that in the Copenhagen al-SufT manuscript.'^
This type of representation appears to have been
the exception rather than the rule. What can be
said with certainty is that the maker of the Smith-
sonian globe did not clearly define the details of
dress necessary for any easy reading of each
figure as draped or undraped; this stylistic clue
will aid us in formulating an attribution of the
globe to a specific metal-worker.
  A number of the other constellation figures
on the Smithsonian globe wear garments with
unpleated short skirts; in the case of Aquarius
and Andromeda, the neckline (or a necklace) is
indicated by a row of small circles around the
shoulders. This sort of beaded or pearl necklet
is an occasional component of Andromeda's cos-
tume, and can be seen on globes Nos, 5 and 11;'^
other female figures such as Cassiopeia often
wear a similar ornament, although the Smithson-
ian Cassiopeia does not. While Aquarius's gar-
ment has a simple band at wrist and hems, An-
dromeda's is sashed below the belly and has
striped borders. Cassiopeia and Perseus wear sim-
ilar sashes and, in contrast to the aforementioned
figures, wear footgear. Perseus wears winged
boots decorated with a scale pattern; he also
wears two small pockets or purses at his belt.'"*
  Four of the male figures (Bootes, Auriga, Or-
ion, and Cepheus) wear long-sleeved, short tunics
belted or sashed at the waist; these tunics have
simple round necklines or wrap around to be
fastened at the side. Virgo wears a similar sashed
tunic, but it reaches her ankles. Both Auriga and
Cepheus have extra bands or sashes at their belts
that curve back over their hips. Bootes wears low
winged shoes'^ and two pockets or purses at his
waist; Cepheus sports knee boots decorated in a
large geometric pattern.'® Cepheus, Auriga, and
Orion each wear particular headgear: Cepheus,
a pointed cap topped by three feather-like forms;
Auriga, a circle-like turban with fluttering end;
and Orion, a low scalloped cap."
  The constellations representing real and
mythological animals are also outlined in single,

roughly continuous lines, with characteristic an-
atomical details receiving a bit more attenion
than in the human diagrams. The contrast be-
tween the rendition of the human constellations
according to a single, standardized type and the
anatomical individualization of the animal dia-
grams is typical of all the extant globes and many
of the al-SQfi manuscript illustrations as well; in
fact, it is typical of Islamic art in general. Occa-
sionally within the al-SQfi cycles, small internal
variations in facial type can be observed. The
shoulders and haunches of the larger and smaller
Canis and Ursa figures on the Smithsonian globe
are indicated by internal, gently curving lines,
and the two dogs have dainty, rounded paws in
contrast to the broader paws with pointed toes
of such figures as Leo. Ursa's wide, flat tail is
distinguished from the pointed tail of Canis Ma-
jor and tufted tails of Canis Minor and Leo.'^
The various creatures have almond-shaped eyes,
some with pupils and some without, crescent
eyebrows in some cases (see Ursa Major and
Delphinus), and, occasionally, bared teeth (see
Ursa Major and Capricorn). Paws are articulated
by rows of short parallel striations to signify the
surface texture of fur or claw; such lines were
also placed around necks (see Cetus), under bel-
lies (see Ursa Major and Canis Major), and across
or along foreheads, as in the case of Cetus and
Taurus.
  The use of such groupings of parallel striations
can be seen in Oxford Bodleian MS Marsh 144
and on globe No. 1; in the manuscript, the lines
are long, fine, and closely spaced, while on the
globe they are short and widely spaced. The
convention was also used on globes such as No.
6 (dated 684 H/AD 1285-1286) and No. 11
(dated 1032 H/AD 1622-1623);'^ its presence in
two variants on globe No. 1 and in the Marsh
manuscript suggest that it would have been char-
acteristic of the earliest copies of al-Sufl's trea-
tise.
  On the Smithsonian globe, hatched lines were
also used to indicate tails and manes, as were the
narrow, pointed, blade-like forms already men-
tioned as the device for representing human hair.
Body-surface textures were indicated in the case
  
102

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



of Pisces, Delphinus, and the reptilian figures of
Hydra and Draco; the latter constellation has a
particularly individualized face with a long,
rolled snout, pointed ears, small curved teeth
and beard, and a narrow forked tongue.^^ The
bodies of these latter beasts are covered with a
scale pattern also used for the wing tops of
Pegasus and the figure of Virgo. Pegasus's lower
wings are hatched in short thinner and thicker
lines as are those of Cygnus, apparently to indi-
cate the overlapping of successive feathers.
  It can be seen from this brief analysis of the
constellation figures on the Smithsonian globe
that the human figures all follow a standard,
simple figural type defined by curving, roughly
continuous outlines. Harsh angles were avoided
although sashes, hair, the corners of hems, and
sometimes hands end in single points. Body con-
tours appear rubbery, sometimes bulbous, and
there is virtually no internal articulation of body
parts except for hands, feet, and basic facial
features. Details of dress and decoration received
minimal attention. It is clear that the individual-
ization of each figure was less important than the
simple outlining of generalized body forms with
the addition of a few attributes. The animal
figures were rendered in the same fashion, with
long, gently curving outlines, a minimum of
pointed or angular forms (except in cases where
they are anatomically necessary to the animal's
recognizability), and a shorthand or cursory in-
dication of skin and fur textures in some in-
stances.
  A stylistic comparison between the Smithson-
ian globe figures and those on the other celestial
globes listed in the catalog indicates an extremely
close affinity between the Smithsonian figures
and those on globes (Nos. 11 and 13) signed by
Qa^im Muhammad ibn Tsa ibn Allahdad Astur-
labT LahurT HumayQnT and dated to the reign of
the Mughal ruler JahangTr.^' The attribution of
the Smithsonian globe to this maker, or to some-
one working closely with him (but not his son
piya" al-DTn Muhammad ibn Qa^im Muhammad
Mulla Tsa ibn shaykh Allahdad AsturlabT Hu-
mayunT LahQrT), can be confirmed by the exam-
ination of individual figures as they appear on

the Smithsonian globe, globes Nos. 11 and 13,
and the group of globes signed by Diya^ al-DTn
(Nos. 18, 20, 22, 23, 26, 27, 28, and 30).^'
  The figure of Auriga on the Smithsonian
globe sits on bulbous, foreshortened buttocks;
his left leg appears to be crossed back over his
right knee (see Figure 58). The somewhat
cramped posture of this figure on this and other
globes may derive from his denomination as
mumsik al-a"innah or "the holder of the reins," a
name that suggests the figure might sometimes
have been imagined as sitting astride an invisible
horse rather than standing in the posture of the
classical charioteer. Since Auriga was repre-
sented in the al-SQfT manuscripts and on globes
either with both knees drawn up or in a standing
posture,^^ we appear to be dealing with two
separate traditions of representing the constella-
tion, one more closely allied to the original Greek
conception of its appearance. This variation in
pose may also have arisen out of the necessity of
incorporating two large stars, placed here on
Auriga's ankles, within the constellation, Al-sufT
calls the star on the western upper ankle, no. 10,
alladhi "aid al-ka"b al-aysar (the one on the left
ankle-bone), while the star on the eastern lower
foot is called no. 11, alladhi "aid al-ka"b al-ayman
wa huwa al-mushtarak lahu wa li-l-qarn al-shimdli
min al-thawr (the one on the right ankle-bone
and it is common to both this position and the
northern horn of Taurus),^'* They were variously
positioned on globes and in the al-SQfl copies on
Auriga's ankles, feet, or legs.
  The eastern shoulder of the Smithsonian Au-
riga figure slopes in a long curve around a large
inlaid silver star, labeled mankib dhi al-"indn (up-
per arm having the rein), to a rounded elbow
held close to his waist. He carries a long upright
staff with five crescents along it and a trefoil
point from which dangles two tasselled ribbons.
This resembles a horse whip and is standard on
all extant globes, though it appears in a number
of different forms,^'"^ Horse whips are also held
by Auriga figures in the al-SQfT manuscripts,
though these tend to have thin thongs instead of
broad bands,^® Auriga's western shoulder forms
an approximately straight line from neck to el-
  
NUMBER 46

103



bow, and he rests his western fist at his belt. It is
this fist that on other globes and in the al-SufT
manuscripts is shown gripping a narrow folded
band, no doubt the reins from which the figure
derives his name.^' Auriga's tunic, of unspecified
length, has a simple V-neck that curves slightly,
and double bands at upper arm, elbow, and wrist.
The figure is bald and wears a fluttering turban
with a large circle of fabric standing straight up
at his forehead to frame two silver stars; al-SQfT
refers to these as no. 1 amyal al-ithnin alladhin
"aid al-ra"'s ila al-janub (the one of the two on
the head which is more inclined toward the
south) and no. 2 amyalhumd ila al-shimdli wa huwa
fawq al-ra"'s (the one inclined more toward the
north, i.e., above the head),^®
  But for three small variations, this figure re-
produces almost exactly the Auriga figures on
globe Nos. 11 and 13 by Qa^im Muhammad (see
Figures 37 and 38). One variation involves the
placement of the constellations right shoulder,
which, on the signed globes, more closely mirrors
the rounded form of the figures' left shoulders.
A second difference can be seen in the rendering
of the neck and front band on the tunics of Qa^im
Muhammad's figures, neither of whom wears a
belt. Instead, Qa^im's Auriga figures grasp their
front tunic plackets, which continue in single
sinuous lines from rounded necklines down
around the figures' seats. Thus the striped band
that hangs from Auriga's belt on the Smithsonian
globe appears on the other globes as the contin-
uation of the tunic's front closure; in all three
cases, Auriga's reins have disappeared entirely,
and he grasps his own clothing instead.
  The final variation that sets the Smithsonian
globe somewhat apart from globes Nos. 11 and
13 involves the facial features apparently com-
mon to figures of Qa^im Muhammad's signed
productions. Most of Qa^im's constellation fig-
ures have almond eyes with enlarged circular
pupils; sometimes the eyes are surrounded by
double lines with eyebrows continuous above
them. Noses are rendered as slightly lopsided
trefoils or double-lobed forms resembling arrow-
heads, and, in the case of the Auriga figures on
globes Nos. 11 and 13, the horizontal incisions


FIGURE 37.—Detail showing Auriga. Globe No. 11, dated
in the eighteenth year of the reign of JahangTr (AD 1622-
1623), by Qa^im Muhammad of the Lahore workshop. Lan-
cashire, Stonyhurst College. (Photo: M.B. Smith)
signifying moustache and mouth respectively are
accompanied by shorter lines below to indicate
chin-clefts. Thus the faces on Qa^im Muham-
mad's figures differ somewhat in their idiosyn-
cratic rendition from those on the Smithsonian
globe, although in all other ways the figures on
the three globes are almost identical.
  The Auriga figure on one of Diya^ al-DTn's
globes (No, 23, dated 1068 H/AD 1657-1658)
follows the seated posture utilized on his father's
  
104

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 38.—View showing Auriga, Globe No. 13, dated in the twenty-second year of the
reign of Jahangir (AD 1626-1627), by Qa^im Muhammad of the Lahore workshop. London,
The Victoria and Albert Museum, Department of Metalwork. (Photo: M.B. Smith)

globes, but shows variations that are character-
istic of Diya^ al-DTn's work (see Figure 39), These
include the rendering of the figure's nose in a
wedge shape and the particularization of his two
attributes. His horse whip has been transformed
into a staff consisting of a row of clusters of leaf-
like forms with two blade-like terminations at the
top, and his soft hat is shaped like an inverted
bowl with a sash around the drooping brim and
a pointed finial. As we shall see, this particulari-
zation and attention to detail, stylistic character-
istics already noted as being conspicuously absent
on the Smithsonian globe and the two globes
signed by Qa^im Muhammad, are the hallmarks
of piya^ al-DTn's mode of working with the tra-
ditional constellation figures.
  
The Auriga on another of Diya^al-Din's globes
(No, 18, dated 1055 H/AD 1645-1646) departs
entirely from the seated posture seen on Qa^im's
globes and represents the figure standing or
walking. Once again, his headdress and whip
have been adjusted; he wears a large, floppy
turban with a triangular pin or feather, and
carries an oversized whip with toothed and tas-
seled bands tied to one end. His short skirt is
unpatterned, and his right hand, rather than
grasping his tunic placket or reins, holds a simple
necklace ending in two large cross-hatched tassels
below his belt.^^ Thus, though his wedge-shaped
nose and other facial features connect the figure
to those on globe No, 23, the posture and outfit-
ting have been changed so as to give this Auriga
  
NUMBER 46

105




FIGURE 39.—Detail showing Auriga. Globe No. 23, dated
1068 H/AD I657-I658, by Diya^ al-DTn Muhammad of the
Lahore workshop. Cardiff, Welsh Industrial and Maritime
Museum. (Photo: M.B. Smith)
a very different personality; he has been trans-
formed from a classical charioteer into a male
figure holding a sort of ceremonial staff (the
original significance of which is all but lost) and
dressed in contemporary clothing.^^
  The theory that suggests itself here is that
while Qa^im Muhammad worked within a rather
limited framework, avoiding experimentation
and giving scant attention to the decorative po-
tential of the figures with which he covered his
globes, his son manipulated the traditional forms
used by his father, embroidering and adjusting
them here and there during his career, which
spanned at least 44 years, Diya^ al-DTn's interest
in decorative detailing, his care in finishing off
his figures, and his attention to variations in
surface textures should be mentioned here as

being fundamental characteristics of his style.^'
  As further proof of this theory, the kinship of
the Smithsonian globe with Qa^im Muhammad's
work as opposed to that of Diya" al-DTn can also
be demonstrated if we turn to the figure of Virgo
(see Figure 68), The figure of Virgo on the
Smithsonian globe is an almost exact replica of
the corresponding figure on one of Qa^im's
globes (No, 11; see Figure 13). Qa^im Muham-
mad's Virgo is a dumpier figure with thicker
legs, and the lines of her gown and its simple
banded edges are once again more sinuous than
those of the Smithsonian figure. Qa^im's Virgo
has a characteristic trefoil nose, as well as the
"inverted" ears^^ seen on other of his figures. On
the other hand, the Virgo on globe No, 28 (dated
1074 H/AD 1663-1664) by Diya^ al-DTn is a more
slender creature with a long, thin neck and small,
egg-shaped head (see Figure 40), In Diya" al-
DTn's hands, her gown has become a more sump-
tuous affair, with wide pearl bands at the cuffs
and upper sleeves, a long, curving sash patterned
with circles and stars, and a delicately scalloped
hem. She wears beaded anklets and necklaces,
and her breasts are visible through her bodice,^^
Her coiffure consists of a row of thick S-shaped
curls across the very top of her head;^^ this fea-
ture and the treatment of her wings (scalloped
only at the top and extending in long, narrow
feathers below) differentiate her from her sisters
on Nos. 11 and 13 and the Smithsonian globe,^^
In addition, the south arm and wing of Diya^ al-
DTn's Virgo are completely visible beneath the
ecliptic down to her hand, while on both the
Smithsonian globe and globe No. 11, Virgo's
southern arm and wing are partially cut by the
ecliptic, A disembodied six-fingered hand, drawn
large to accommodate a silver star labeled simdk
a"zal (an old Bedouin name for the star Spica, a
Virginis), marks the end of her southern arm at
roughly knee level on the Smithsonian figure;
her wing tip is visible just below this hand.
  Comparisons of other human and half-human
figures could be made with their counterparts
on globes by Qa^'im Muhammad and later globes
by his son to flesh out more fully the attribution
of the Smithsonian globe to Qa^im himself or a
  
106

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 40.—View showing Virgo. Globe No. 28, dated 1074 H/AD 1663-1664, by Diya" al-
DTn Muhammad of the Lahore workshop. Oxford, Museum of the History of Science. (Photo:
M.B. Smith)

close associate. From these two examples alone,
however, it should be clear that Diya" al-DTn's
interests or points of visual reference differed in
certain basic ways from those of his father, and
that the Smithsonian globe fits in with Qa'^im's
uncomplicated style of around the 1620s rather
than with his son's more detailed and experimen-
tal way of working. A survey of the animal con-
stellations and those representing inanimate ob-
jects supports this hypothesis; here we will con-
sider only the figures of Leo and Argo Navis,
two of the more prominent of the non-human
constellations.
  The figure of Leo on the Smithsonian globe
consists of a thick-chested feline with extremely

short front paws, each ending in four pointed
toes (see Figure 67). The rest of Leo's body has
a somewhat telescoped appearance due to his
exaggeratedly high rump (placed to encompass a
group of stars, the largest of which is labeled
zahr al-asad [the back of the lion]), and long,
practically straight hind legs ending in narrow
paws with pointed toes. The lion's hind leg joints
were placed too low and his lower legs drawn flat
like long feet, instead of having him stand only
on his paws and setting the upper and lower legs
at angles to one another.^® Leo's thin tail curls in
a circle around a single large star labeled sarfah
(change of weather, fi Leonis), with its tufted tip
radiating outwards in a series of points. The star

NUMBER 46

107



sarfah was usually placed at the tip of Leo's tail,
although on Nos. 1 and 26 the tail curls around
the star as it does on the Smithsonian globe and
other globes by Qa^im Muhammad and Diya^ al-
DTn, Despite the profile rendition of his body,
the lion's head is seen full-face with its features
(almond eyes and a trefoil nose) drawn as if on a
completely flat surface. Representations of Leo
on other globes and in the al-Sufi manuscripts
inevitably show the lion's body in profile, but his
head is often turned in a three-quarter pose so
that both eyes and ears are visible. This is how
he was represented in Oxford Bodleian MS
Marsh 144; in an al-SQfi manuscript in the British
Library, London (MS Or. 5323);^' on globes
Nos. 3, 4, 5; on Qa^im's globes Nos. 11 and 13;
and on at least one globe by Diya^ al-DTn, No.

28, This pose was also used in illustrations accom-
panying astronomical and astrological treatises,^^
  The eyes of the Smithsonian Leo contain
hatched pupils. Below his nostrils are two pointed
horizontal bands for whiskers; below these, two
stars occupy the lion's rounded jowls. The short
fur of his mane stands up like a row of flames
across the top of his head, with four blades of fur
falling down along his neck,^^
  Just as with the figures of Virgo and Auriga,
the Smithsonian Leo reproduces almost exactly
the leonine diagrams on Qa^im Muhammad's
globes Nos, 11 and 13, but for the smallest
variation: on globe No, 11 Leo has slightly longer
mane fur, and on Nos, 11 and 13 he has double-
outlined eyes and what appear to be smiling
mouths instead of whiskers (see Figures 41 and
  

FIGURE 41.—Detail showing Leo. Globe No. 11, dated in the eighteenth year of the reign of
JahangTr (AD 1622-1623), by Qa^im Muhammad of the Lahore workshop. Lancashire, Stony-
hurst College. (Photo: M.B. Smith)

108

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



42). Otherwise all three lions, with their vestigial
front paws and extremely long hind legs, are
practically identical.
  On one of Diya^ al-DTn's globes. No. 28 (dated
1074 H/AD 1663-1664), we observe a similar
Leo, but once again the form has been manipu-
lated and details added (see Figure 43). Although
the general posture and angle of the head were
reproduced, Diya^ al-DTn modified the lion's
front legs by broadening his shoulders and paws,
rounding his toes, and adding a furry ridge along
the bottom of the elbow as well as a ring of short
hatches around the neck. Leo as depicted by
Piya^ al-DTn has more prominent, softly furred
ears, and his hind paws are also broader, with a
furry edge. Thus, the Leo on globe No. 28 is

more particularly leonine in all its parts than
Qa^im Muhammad's figures on globes Nos. 11
and 13, which are recognizable as felines solely
on the basis of their schematized facial forms and
characteristically tufted tails.
  Diya^ al-DTn's further adaptation of the Leo
figure can be seen on two other globes. No. 23
(dated 1068 H/AD 1657-1658) and No. 26 (dated
1071 H/AD 1660-1661; see Figures 44 and 45).
While leaving the general outline of the lion's
body unchanged, Diya^ al-DTn gave him shorter,
thicker hind legs and reduced his hind paws to
their natural size; this reduction is especially ap-
parent on globe No. 23. In addition, the lion's
head is shown in profile like his body, thus avoid-
ing the somewhat awkward twisting necessary to
  

FIGURE 42.—Detail showing Leo. Globe No. 13, dated in the twenty-second year of the reign
of JahangTr (AD 1626-1627), by Qa^im Muhammad of the Lahore workshop. London, The
Victoria and Albert Museum, Department of Metalwork. (Photo: M.B. Smith)

NUMBER 46

109




FIGURE 43.—Detail showing Leo. Globe No. 28, dated 1074
H/AD 1663-1664, by Diya^ al-DTn Muhammad of the Lahore
workshop. Oxford, Museum of History of Science. (Photo:
M.B. Smith)

show him full-face. The head itself comes even
closer to what might be termed a realistic repre-
sentation of a particular animal, in that it is broad
and short with slightly drooping chin, box-like
snout, and small erect ears; the lion's neck has
also been thickened to indicate the presence of a
mane, although the texture of its fur is not
shown. This total profile view of the lion con-
stellation was the alternate type used on globes,
and more especially in al-SufT illustrations; ex-
amples include the Leo figures on globe No. 1
and in the al-SQfT manuscripts in the Metropoli-
tan Museum of Art, New York, and in Lenin-
grad.'*° The total profile lion existed from at

FIGURE 44.—Detail showing Leo. Globe No. 23, dated 1068
H/AD 1657-1658, by Diya^ al-DTn Muhammad of the Lahore
workshop. Cardiff, Welsh Industrial and Maritime Museum.
(Photo: M.B. Smith)

least as early as the eleventh century, as did the
three-quarter version we see in Oxford Bodleian
MS Marsh 144.
  Piya^ al-DTn's version of a Leo figure unlike
that used by his father follows the pattern already
observed in connection with his rendering of the
human constellation figures. If we turn in con-
clusion to Argo Navis, we see that he also ad-
justed this figure. The boat on the Smithsonian
globe is a flat-bottomed vessel with a high vertical
prow surmounted by a leonine figurehead (see
Figure 81). It has open railings along either side,
a pergola at the stern, and two furled sails at-
tached to a central mast.'*' A thin double chain
falls from the mast down to the figurehead, and
three unmarked pennants, one each at the lion's
head, the mast top, and behind the mast on the
  
110

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 45.—Globe No. 26, dated 1071 H/AD 1660-1661, by Diya^al-Din Muhammad of the
Lahore workshop. Berlin, DDR, Staatliche Museum. (Photo: from Bassler-Archiv)

main deck, fly overhead. The candlestick-shaped
oars stick diagonally out of the boat below the
prow. The vessel's decoration consists of a row
of cusped half-cartouches or bracket-like forms
along the forward edge of the prow and a similar
but larger form at the boat's bottom,^^ The per-

gola roof is rimmed with a row of small scallops
and topped by a trefoil finial, a larger version of
which crowns the mast.
  The representations of the Argo constellation
vary from globe to globe perhaps more than any
of the other diagrams; this may have been due
  
NUMBER 46

111



to local and temporal differences in ship design
of a kind more radical than, for example, varia-
tions in basic fashions of dress. Not all of the
boats have sails, pennants, or even masts. Thus
it is all the more significant that the Argo Navis
diagram on Qa^im's globe (No, 11; see Figure
46) is practically the twin of that on the Smith-
sonian globe, piya'' al-DTn's treatment of the
constellation on Nos, 23, 26, and 28 is similar in
certain details (the leonine figurehead and
bracket-like decoration, for example) but vitally
different in others (see Figures 45, 47, and 48),
These variations include the lack of sails, the
modification of the open pergola into a tall, bud-
like form, the elaboration of a central structure
based on vase-shaped supports and with another,
more elaborate pergola atop it, and the modifi-
cation of the oars into large, bell-like forms on
chains. The shape given to the oars by Piya^ al-
DTn is closest to that on an early globe (No. 4)
and the earliest al-SQfi manuscript,'*^ though in-
tervening Argo diagrams show many variations.
Piya^ al-DTn's substitution of chains for the oar
arms is unusual and possibly unique, and suggests

that he interpreted the forms rather as anchors
than as oars."*^ piya^ al-DTn's pennants have a
spotted pattern also used on another of his globes
(No. 18), where the Argo constellation received
a much more complicated and consistent treat-
ment involving the total modification of its char-
acter as conceived by Qa^im, even to the substi-
tution of a bird facing backward as the figure-
head (see Figure 15). This final detail exists in at
least two al-SQfT manuscripts, in the Metropolitan
Museum of Art, New York, and the Leningrad
versions;'*^ while the former Argo diagram shows
a feline figurehead, that in the latter manuscript
has a bird with a curved beak. This peculiarity
of Argo's decoration may ultimately make ref-
erence to Aratus's claim that the ship sailed back-
ward, in other words, stern first,'*® and it is inter-
esting that the reversed figurehead was only
shown on some versions of the constellation dia-
gram.
  The consistency with which the non-human
constellation diagrams on the Smithsonian globe
reproduce those on globes signed by Qa^'im Mu-
hammad, while differing in one significant way
  

FIGURE 46.—Detail showing Argo. Globe No. 11, dated in the eighteenth year of the reign of
JahangTr (AD 1622-1623), by Qa^im Muhammad of the Lahore workshop. Lancashire, Stony-
hurst College. (Photo: M.B. Smith)

112

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 47.—Detail showing Argo. Globe No. 28, dated 1074 H/AD 1663-1664, by Diya^ al-
DTn Muhammad of the Lahore workshop. Oxford, Museum of the History of Science. (Photo:
M.B. Smith)

or another from those on globes signed by his
son, reinforces our original attribution of the
unsigned globe to Qa^im Muhammad or a close
associate. The fact that globe No. 11 by Qa^im
himself is dated AD 1622 (the eighteenth year of
the reign of JahangTr) suggests a date for the
Smithsonian piece somewhere around that year,
and, if we can interpret Qa^im's toponym as an
indication of the area in which he worked as well
as his birthplace, a locale in northwest India for
its production. Certain details of dress point di-
rectly to the globe's manufacture in a Mughal
context. Such a date and provenance help to
explain why many characteristics of Qa^im's (and
Piya" al-DTn's) figures differ from those on ear-
lier globes and in pre-seventeenth-century man-
uscripts of al-SufT's Book of the Constellations of the
Fixed Stars.
Long-   or   short-sleeved   tunics   of   varying
lengths, sometimes worn over trousers, are stan-
	        ^, ,    .,    ^„    .     .	dard garb for the human constellation figures on
FIGURE 48.—Detail showmg Argo. Globe No. 23, dated	",   ,             , .      ,       ,__,_.„            -47        ,
1068 H/AD 1657-1658, by Diya^ al-DTn Muhammad of the	"^^^^ S'^^es and in the al-SuH illustrations,    and
Lahore workshop. Cardiff, Welsh Industrial and Maritime	turbans are the normal headgear for the male
Museum. (Photo: M.B. Smith)	figures."*^ Usually the tunics are tucked in at the

NUMBER 46

113



waist or are belted. Occasionally, figures are
represented clad only in skirts; Andromeda on
Nos. 4 and 5 was engraved in this fashion. On
the Smithsonian globe, with the exception of
Virgo, all the figures wear knee-length tunics
with long sleeves; in some cases the tunics are
pleated. All have a pronounced flare at the waist.
They are tied with long, occasionally tasselled
sashes below the belly; these sashes are looped
up once over the waist. In some cases (see Au-
riga), a part of the sash also dips back over the
wearer's hip.
  Such flared tunics, in various lengths, were
characteristic male dress at the Mughal court, as
can be seen in paintings of the late sixteenth and
seventeenth centuries. Work executed during
the reigns of "Abd al-Fath Jalal al-DTn Akbar (AD
1556-1605) and his successor NQr al-DTn Mu-
hammad JahangTr (AD 1605-1627) show cour-
tiers (and also laborers and soldiers) wearing
knee- or calf-length tunics'*^ tied with a variety
of sashes: long, narrow ones worn singly and in
pairs tied with a simple knot;^^ short, thicker
ones knotted in the same fashion;^' combinations
of both types worn together;^^ or, as on the
Smithsonian globe, single, narrow sashes knotted
in a short loop that hangs over the waistband.

Since few published paintings from this period
are definitely dated, it is impossible to say pre-
cisely when the single-looped sash came into fash-
ion. A survey of paintings done under Akbar and
JahangTr shows single-looped sashes occurring in
works dated or attributed to the years between
AD 1583 and 1610.'^^ This suggests that such
sashes were popular during Akbar's reign and
that the fashion died out completely shortly after
JahangTr gained the throne.
  A preference for the single-looped sash worn
by itself on Qa^im Muhammad's globes suggests
that he must have begun work as a maker of
celestial globes during the reign of Akbar, con-
tinuing under JahangTr without adjusting his
mode of dressing the constellation figures to
mirror minor changes in contemporary male
fashions. Such a theory fits quite well with the
generally conservative picture we have con-
structed of Qa^im's style, especially in compari-
son to that of his son. In any event, Qa^im's
particular modification of the outfit standard for
male constellation figures on earlier globes to
reflect contemporary garb justifies a dating of
the Smithsonian globe to the first quarter of the
seventeenth century or slightly earlier.
  
A History of Star Names, Based on the
Smithsonian Globe
I am the shepherd of the skies.
  Deputed to preserve
The planets as they sink and rise,
The stars that do not swerve.
Those, as they swing their lamps above
  Our earth, by night possessed.
Are like the kindled fires of love
Within my darkling breast. . . .
Were Ptolemy alive today,
  And did he know of me,
"Thou art the maestro," he would say,
"Of all astronomy."
Ibn Hazm
The Ring of the Dove

CLASSICAL GREEK AND PRE-ISLAMIC SOURCES
  The star names used in the classical Islamic
world were derived from two distinct sources:
the names used by pre-Islamic Bedouins, and
those transmitted from the Greek world. As
Greek astronomy and astrology were accepted
and elaborated, primarily through the Arabic
translation of Ptolemy's Almagest, the indigenous
Bedouin star groupings were overlaid with the
Ptolemaic constellations that we recognize today.
  From about AD 1500 navigators in the Indian
Ocean contributed Persian elements to the star
nomenclature.' These later influences, as well as
the nineteenth-century translation of European
star names for the southern constellations not
known to classical antiquity, are not reflected in
the stars and star names on any Islamicate celes-
tial globes, even those made as late as the nine-
teenth century.
  For the early Bedouin star names and constel-
lation images, the sources of information avail-
able today are complex and confusing, although
quite rich, and consist mainly of anwa^'-literature
and writings of astronomers, such as al-SufT and
al-BTrQnT, who concerned themselves with com-
paring the traditional Bedouin terms and concep-
tions with  the later Ptolemaic ones.  (See the

following section on "Lunar Mansions" for fur-
ther details on this type of literature and on the
pre-Islamic star names and images).
  Although "Abd al-Rahman al-SQfT discussed in
his Book of the Constellations of the Fixed Stars the
early Arab star names and stars, some of which
were not described by Ptolemy, the star catalog
that accompanies al-SQfT's treatise is strictly Pto-
lemaic, giving only those stars listed in Ptolemy's
earlier catalog with Arabic translations of the
Greek names and an occasional earlier Arab
name that he could align with a Ptolemaic star.
The constellation outlines illustrating the treatise
by al-SQfT with very few exceptions depict the
Ptolemaic constellations rather than the earlier
Bedouin groupings.
  References on extant Islamicate celestial
globes to the sources used by the makers for the
star positions mention three particular star cata-
logs. Ptolemy's Almagest is named on Nos. 2 and
3, the catalog of al-SQfT on Nos. 6, 7, 8, and 62,
and the fifteenth-century catalog of Ulugh Beg
on globes Nos, 14, 31, 60, and 64. While it is
possible that some of the other Islamic star cata-
logs were employed by globe makers,^ it is prob-
able that most makers used one of these three.
  In his catalog Ptolemy presented a description
of 1025 stars with accompanying ecliptic coor-
  
114

NUMBER 46

115



dinates for the year AD 138 and designations of
magnitude. The stars were not assigned ordinal
numbers by Ptolemy, but were described under
48 constellation headings that would usually con-
tain two groups of stars: those formed {p.bp(poiToi)
and those unformed {ajxbpipooTOi). These terms
refer to whether the stars lie inside or outside
the constellation outlines. The number of stars
making up each of the 48 constellations was then
totaled as so many formed stars and so many
unformed ones, and the total number of stars
given for the three groupings of constellations:
northern, zodiacal, and southern. In the total for
the unformed stars of Leo, Ptolemy did not
include the three stars of Coma Berenices {o irXb-
Kafios), though he listed them individually in the
catalog. Nor did Ptolemy include them in his
total for the zodiacal constellations or the total
for the entire catalog, which he gave as a total of
1022 stars consisting of 15 of the first magnitude,
45 of the second magnitude, 208 of the third
magnitude, 474 of the fourth magnitude, 217 of
the fifth magnitude, 49 of the sixth magnitude,
plus 9 faint, 5 nebulous, and o izXbKapos {Coma
Berenices).^ Modern writers have confused the
issue of how many stars were in Ptolemy's catalog
by consecutively numbering every star listed.
Since Ptolemy repeated the description of three
stars because they each were shared by two con-
stellation outlines (ones shared by Hercules and
Bootes, by Taurus and Auriga, and by Aquarius
and Piscis Austrinus), though he counted each
only once in the summaries of the totals, and
because of his failure to include in the totals the
three stars in Coma, modern writers have con-
cluded there are 1028 stars in the catalog.^
  Al-SufT repeated the catalog of Ptolemy, mak-
ing changes in the longitudes to correspond to
364 H/AD 974, and in some of the magnitudes.
He assigned each star a number within the
formed or unformed stars of a constellation,
including the three stars of Coma Berenices in the
eight unformed stars of Leo, but did not repeat
the description of the three shared stars, so that
they were assigned only one number each, Al-
SufT also omitted the last of the Ptolemaic stars
in Auriga because it could not be observed. In
this way al-SufT arrived at a total of 1024 stars in

his catalog. Of these 1024 stars he noted that the
last six (the unformed stars of Piscis Austrinus)
were not visible, also adding that no. 11 of Lupus
and no. 30 of Centaurus were also not observa-
ble. He chose to leave them all in the catalog,
however, with a brief statement that they had
not been observed. In the text accompanying his
star catalog al-SQfT described many stars which
he did not give in the catalog, one of them being
"the overlooked star" of Ursa Major, which oc-
curs on several celestial globes including the
Smithsonian globe (No. 38) and three early
globes (Nos. 4, 5, and 6).
  When discussing particular star positions al-
SQfT frequently refers to celestial globes he has
seen. For example, he says that the fifth star of
Libra is usually given the northern latitude of
1 °40' when represented on globes following that
given in the Almagest, but that it ought to be
1 °40' south latitude, and he says that he has seen
globes on which the constellation of Libra was
depicted as a man holding a small pair of scales,^
  The star catalog that Ulugh Beg prepared at
Samarqand with coordinates for 841 H/AD 1437-
1438 was probably the major reference for the
later eastern globes.^ The catalog essentially re-
peated the stars and numbering of al-SQfT's cat-
alog of five centuries earlier, with some changes
in coordinates, but with the last six stars of the
Southern Fish omitted. Thus it had a total of
1018 stars. In the thirteenth chapter of the third
section of the preface to this zij, Ulugh Beg
makes the following statements, where he clearly
speaks of placing stars on a celestial globe in
order to see how the positions in a given catalog
compare with those in the sky:
   Before the time of Ptolemy 1,0,22 fixed stars had been
observed. Ptolemy has given them in a catalogue in the
Almagest. The stars are distributed in six magnitudes; the
largest are of the first and the smallest of the sixth magni-
tude. Each magnitude is divided into thirds, and in order to
recognize the stars, 48 figures or constellations have been
imagined, of which 21 are north of the ecliptic, 12 in the
Zodiac, and 15 south of the ecliptic. The larger number of
the stars are within the figures, the others are in the neigh-
bourhood, and are designated as unformed stars of the
constellation.
   "Abd al-Rahman al-SufT composed a treatise on the stars
which all learned men have received with gratitude. Before
  
116

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



determining by our own observations the positions of these
stars, we have laid them down on a sphere according to this
treatise, and we have found that the greater part of them
are situated differently from their appearance in the heav-
ens. This determined us to observe them ourselves with the
assistance of Divine Providence, and we have found that
they were advanced from the epoch at which al-SufT's work
was written, so that on giving them, according to this general
observation, their absolute positions, we no longer found
any difference from their appearance to the eye.
   It is on this principle that we have reobserved all the stars
already determined, with the exception of 27 which are too
far to the south to be visible at the latitude of Samarqand,
namely the 7 stars in the constellation Ara; 8 in Argo Navis,
stars 36 to 41 and 44 and 45; II in Centaurus, from the
27th to the end; and one star, the 10th in the constellation
Lupus; and we have taken these 27 stars from the work of
"Abd al-Rahman al-SufT, taking account of the difference of
epoch.
   Besides these there are 8 stars mentioned by "Abd al-
Rahman al-SiJfT in his book, of which Ptolemy gives the
positions, but which "Abd al-Rahman al-SufT could not find,
and which notwithstanding all our researches, we have been
unable to discover. For that reason we do not indicate those
stars in the present catalogue. These Ptolemy stars are the
14th of Auriga, the 11 th of Lupus and the 6 unformed stars
of Piscis Austrinus.
   In our catalogue we have given the position of the stars
for the beginning of the year 841 of the Hegira, so that at
any time we may be able to find the place of any stars on
the supposition that they advance one degree in seventy
solar years.^
  Despite these statements by Ulugh Beg, how-
ever, al-SQfi did omit the fourteenth star of Au-
riga from his catalog,^ though he left in no. 11
of Lupus and the six external stars of Piscis
Austrinus. Furthermore, all the extant manu-
script copies of Ulugh Beg's star catalog^ indicate
that Ulugh Beg also left the eleventh star of
Lupus in the catalog, despite his statement that
he was omitting it.
  What appears rather curious then is the fre-
quent use of the number 1022 for the total
number of visible fixed stars in the catalog of
Ulugh Beg. For examples, Ghiyath al-DTn al-
KashT in his letter describing the scientific activity
of the observatory of Ulugh Beg says that the
astronomers had attempted to construct an astro-
labe that had the totality of the 1022 observable
fixed stars.'" Qa^im Muhammad ibn "Isa ibn Al-
lahdad AsturlabT LahQrT HumayQnT on his globe

made in AD 1637-1638 (No, 14) states that 1022
stars have been placed on it in accordance with
the catalog prepared by Ulugh Beg. Similarly,
the globe by Ridwan completed in 1701 bears an
inscription stating that the number of fixed stars
represented on it is 1022 and that their positions
were based upon the catalog of Ulugh Beg (globe
No, 31). The globe produced by Lalah Balhumal
LahQrT in 1842 (No. 33) also speaks of having
1022 observable stars indicated on it, although
Ulugh Beg's catalog is not specifically named.
  The total of 1022 could possibly be obtained
by taking the 1018 stars in the catalog, counting
the three shared stars twice, and adding in the
"overlooked" star of the Greater Bear. The latter
star, however, does not occur on the globes that
specifically claim to have 1022 stars. It is most
probable that globe makers and astronomers sim-
ply used the catalog for the coordinates and did
not count the precise number of stars. They
perhaps continued to use the number 1022 that
Ptolemy had given for his catalog and did not
note the minor variations from one catalog to
another and the consequent changes in the total
number of stars.
  All globes up through the fourteenth century
have the six external stars of Piscis Austrinus.
Two of these globes (Nos. 6 and 7) are stated to
be based on the Book of the Constellations by al-
SQfT, and in fact have only 13 stars in Auriga,
just as al-SQfT described the constellation in his
catalog. The earlier globes (Nos. 1-5 and 34)
have 14 stars in Auriga, following Ptolemy's cat-
alog (except for No. 3 which for some reason has
13 stars in Auriga, even though the maker spec-
ifies Ptolemy's catalog as the source for the star
positions). All globes of the sixteenth century
and later were most likely based on Ulugh Beg's
catalog, and on all of them the fourteenth star
of Auriga is omitted, as are the six unformed
stars of the Southern Fish,
  When Ptolemy's Almagest was translated in the
ninth century, the descriptions and names of the
Greek zodiacal constellations were translated
into Arabic, In most cases these newer Arabic
terms were subsequently replaced by pre-Islamic
Arab names.  Both the Greek and indigenous
  
NUMBER 46

117



Arabic zodiacal constellations had been derived
ultimately from a common Sumerian source."
For example, the Arabic translation of the Al-
magest gives al-rdmi as the name for Sagittarius,
but the older term al-qaws (the bow) soon re-
placed it. Similarly al-taw^amdn (Gemini), al-
"adhrd^ {Virgo), al-samakatdn (Pisces), and al-dali
or sdkib al-md" for Aquarius were replaced by the
older traditional terms of al-jawzd"', al-sunbula,
al-hut, and al-dalw. Kunitzsch'^ suggests that the
Arabs were more conservative when adapting
the Sumerian zodiacal images than were the
Greeks, for the Bedouins did not tend to put
people with the objects; for example, they imag-
ined a grain of wheat but without the girl (Virgo)
or a bow but no centaur (Sagittarius).
  In pre-Islamic groupings it seems that al-dalw
(the bucket) covered the large areas of Aquarius,
Pegasus, and part of Pisces; al-hut (the fish) cov-
ered the regions of Andromeda and Pisces; while
al-asad (the lion) was much larger than our Leo.
This extremely large lion had its nose in what we
call Cancer, its eye, forehead, neck, and shoul-
ders in Leo, its one foreleg formed by the heads
of Gemini, the other foreleg formed by Canis
Minor, its hind quarters in Corvus, and its hind
legs formed by Spica in Virgo and Arcturus in
Bootes. The liver of this huge lion was the first
unformed star of Ursa Major.
  Another important image in the Bedouin sky
was that of a large woman called al-thurayyd. Her
head was composed of the open star cluster called
the Pleiades in the constellation Taurus. The
Pleiades were consequently frequently called al-
thurayyd. The woman had one arm passing
through the regions of Perseus and Cassiopeia
where her finger tips were said to be stained with
the red day henna, while her other hand was
seen in the area now occupied by the Ptolemaic
constellation of Cetus. The Hyades, another
open star cluster in Taurus, were called young
female camels {qald"'is) and, according to tradi-
tion, were driven before Aldebaran (the fol-
lower; a Tauri), the most prominent star in the
cluster and sometimes called the faniq (camel-
stallion), as evidence of his wealth when he went
again to court al-thurayyd, who had earlier re-

jected him because of his poverty. Aldebaran's
two dogs were seen in two small stars near one
another in a narrow space between the Pleiades
and Hyades.
  The region of the constellations Orion and
Gemini was seen in pre-Islamic Arabia to contain
a huge giant named al-jawzd"', which was possibly
a feminine figure originally.'^ The stars in the
feet of Gemini, according to some, made up the
bow which al-jawzd"" throws at the huge lion. A
throne for him was formed by the stars in the
Ptolemaic Lupus and another chair was formed
by the stars in Eridanus. Astronomers used the
named al-jawzd"' not only for the Ptolemaic con-
stellation of Orion but also for the zodiacal house
and constellation of Gemini. One legend re-
counts the marriage of al-jawzd'^ (interpreted as
a feminine figure) with suhayl (Canopus), the
second brightest star in the heavens. It was said
that suhayl lived with al-jawzd"" but broke her
vertebrae and back; so he escaped slaughter by
going toward the south, thus avoiding pursuit.
  A common Babylonian source gave rise to both
the early Arab term for the brightest star in the
heavens {a Canis majoris) called shi"rd and the
Greek term aelpios, from which we get the name
Sirius. In the Bedouin tradition there were said
to be two Sirii and both were sisters of Canopus
{suhayl who had married al-jawzd^). The south-
ern Sirius was the star in Canis Major which we
call Sirius today, while fi Canis majoris was its
companion {mirzam). The northern Sirius was
the star Procyon {a Canis minoris) with its com-
panion {mirzam; fi Canis minoris). The southern
Sirius with its companion was called al-shi"rd al-
"abur (the Sirius passing over) because it was said
to cross over the Milky Way southward toward
Canopus when he fled toward the south after
injuring al-jawzd"', while Procyon with its com-
panion was called al-shi"rd al-ghumaysd"' (the Sir-
ius shedding tears) for it had to remain behind.''*
It was also said that the two stars in Canis Major
were called the oath breakers, for when a person
who did not know the skies very well would see
Sirius and its companion rising he would be
willing to take an oath that it was Canopus and
its companion star, but he would have perjured
  
118

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



himself when Canopus and its companion really
did rise.
  There are testimonies from as far back as the
third century BC of there being in Arabia animals
such as gazelles, lions, wolves, hyenas, hares,
onagers, horses, dogs, ostriches, bustards, and
perhaps wild dromedaries.'^ Many of these ani-
mals, as well as other aspects of pastoral life, can
be seen in other constellation images from pre-
Islamic Arabia. Gazelles were imagined, and the
footprints of their leaps as they ran before the
large lion could be seen in the area of Ursa
Major. The stars we now view as the body of
Cetus were envisioned as ostriches, and the stars
of Corona Australis were said to depict an os-
trich's nest. Eight more ostriches were imagined
in the area now occupied by Sagittarius, four
going to the river of the Milky Way and four
leaving the river. Camels that had recently foaled
were seen in the head of Draco, while some saw
a camel formed of some of the stars in Cassopeia
while the brightest star of that group was called
the camel's hump.'^ In the area of Ursa Minor
there were visualized two calves turning a grist
mill, and between the two calves and the camels
in the head of Draco were wolves. A herd of
goats was seen in the area of Auriga, and the
stars of Cepheus were viewed as a shepherd with
his dog and sheep. The stars Vega and Antares
were said to be dogs barking on account of the
cold weather. A row of horsemen with an outri-
der behind were seen in the stars comprising the
breast and wings of Cygnus with the outrider
made of the tail star. The stars of Corvus formed
a tent, and horses and foals were the stars near
the stable comprised of Crater. The large star in
the tail of Cetus and the star in the mouth of the
Southern Fish were called the two frogs, while a
leather bucket and rope occupied the areas of
Aquarius, Pegasus, and Pisces. Bunches of grapes
were seen in the stars of Centaurus and Lupus,
and a meadow was imagined enclosed by two
rows of stars in Hercules and Serpentarius.
  The Milky Way (known in Arabic as tariq al-
tabbdnah, majarrah [the course], umm al-samd""
[the mother of the heavens], and other terms)"
was important in the early imagery, even though

it is not indicated on any known Islamicate globe.
  The most important modern critical source
for the early Arabic star names are the studies
by Kunitzsch, which have been used extensively
in the following discussion, as well as other sec-
ondary sources'^ and some primary ones, in par-
ticular the tenth-century writings of al-SQfT and
al-BTrunT and thirteenth-century tract by al-Qaz-
wTnT.'^ In the following two sections some of the
Greek accounts will be given for the Ptolemaic
constellations which overlaid the earlier Bedouin
ones. For these Greek sources the writings of
Aratus and early poets such as Homer and He-
siod have been consulted and cited, as well as the
considerable number of secondary sources con-
cerned with Greek star names.^"
  In the two subsequent subsections of this part
of the study the lunar mansions and the 48
constellations will be discussed in detail. The
lunar mansions are treated separately from the
constellations because they are not part of the
Ptolemaic constellation outlines, but are in-
scribed at regular intervals along the ecliptic.
Furthermore, additional comments are required
for background to the complex subject of lunar
mansions. In the section on the "Constellations"
in this chapter, each constellation will be dis-
cussed in terms of the star names appearing on
the Smithsonian globe. The order of the constel-
lations will be the order given by Ptolemy and in
all derivative Islamic star catalogs. When a star
is described as being a particular number in a
constellation, this again refers to numbering in
the al-SQfT and Ulugh Beg catalogs. The minor
differences in their numbers, when pertinent,
will be noted. Minor spelling errors on the part
of the maker and inconsistencies in genitive con-
structions will not be noted (see Chapter 3).
References to other globes when comparing
terms or images will be by the catalog numbers
given in Chapter 6. Arrows are placed on the
photographs to designate plugs or other special
features, in order to distinguish the plugs from
stars. Compass directions have been noted on
each photograph of the individual constellations
so that proper orientation can be maintained.
Descriptions of the figures will be given in terms
  
NUMBER 46

119



of the direction of the compass rather than right
or left to avoid confusion.
  The unequivocal modern identification of a
star is not always possible and modern writers
occasionally disagree on this. Generally Kun-
itzsch's identification has been followed,^' but
sometimes that of Peters and Knobel or Baily
was used. Quite frequently the number of the
star in the al-SQfT star catalog is given just before
the modern identification. The nomenclature for
the latter is for the most part that of Bayer,
employing Greek letters, and when necessary
lower case Latin letters, followed by the genitive
of the Latin name for the constellation. Occa-
sionally a star not covered by the Bayer system is
designated by a Flamsteed number (abbreviated
Flam,),
LUNAR MANSIONS
  The origin of the system of lunar mansions is
obscure, with at least four theories having been
put forward regarding the matter. It has been
suggested that it began with the related Chinese
system called hsiu and spread from China to India
and the Near East,^^ The Indian system called
naksatra of 27 or 28 junction-stars has also been
mentioned as the original source,^^ Others have
contended that it was Babylonian in origin and
extended thence to India, China, and Arabia,^^
or that Hellenistic astronomy played a role in the
diffusion either as a point of origin or through
Hellenistic astronomical and trigonometric tech-
niques current in India.^^ It is fairly evident that
the Arabic version is an accretion of the Indian
upon an earlier Bedouin grouping of fixed stars,
applying the traditional Bedouin star names to
the Indian lunar mansions dividing the zodiac.
  The Bedouins on the Arabian peninsula in pre-
Islamic times were chiefly nomadic camel owners
and herders of goats who needed to estimate the
passage of time and to predict meteorological
events so as to locate winter and spring grazing
lands whose locations varied greatly depending
upon the rainfall,^® These early Arabs had a
primitive system called anwd"' based on a series
of 28 prominent star groups. The solar year was
then divided into 28 intervals, each marked by

observing the star group that set in the West at
dawn (i.e., cosmical setting called raqib) and the
star group that rose in the East with the sun (i.e.,
heliacal rising, called naw").^^ Naturally these two
asterisms were opposite each other in the heavens
and in the anwd"" sequence of 28 star groups. The
stars themselves were held responsible for
weather conditions; thus the early Bedouins
could attempt to predict the weather and mark
the passage of time.
  Sometime before the advent of Islam the
Bedouins assimilated from India a system of 27
or 28 "stations" or "mansions" of the moon,
called in Arabic mandzil (singular manzil).^^ The
mansions corresponded to places in the sky
through which the moon passed in its course
from new moon to new moon in 27 or 28 nights.
The course of the moon is inclined to the ecliptic
at an angle only slightly more than 5°, but its
brilliance is such that nearby stars cannot be
observed. For this reason the mansions were
named for stars in the vicinity of but not directly
along the ecliptic. Each mansion came to repre-
sent approximately one day's travel of the moon
and therefore corresponds to roughly 13° along
the ecliptic beginning at the vernal equinox. Ac-
cording to al-BTrunT,^^ it seems that by the time
the system was assimilated in pre-Islamic Arabia
the system of mandzil had come to be more
frequently associated with intervals of the ecliptic
beginning with the vernal equinox than with the
original 27 or 28 junction-stars or asterisms that
gave them their names.
  The resulting anwd^-mandzil system began
with a star group in Aries (probably to be iden-
tified with fiy Arietis), which corresponded to the
first segment of the House of Aries near the
vernal equinox about 300 BC. These two systems
are not, however, entirely compatible, for one is
calculated on the basis of the risings and settings
of fixed star groups and the other reckoned on
regular intervals of the ecliptic taken from the
vernal equinox. With the precession of the equi-
noxes no fixed star will long maintain the same
distance from the vernal equinox. Consequently
one star group cannot be successfully aligned
with one segment of the ecliptic measured from
  
120

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



the vernal equinox for an extended period of
time.
  This attempted compounding of anwd"' and
the lunar mansions {mandzil) gave rise to a type
of Arabic literature known as anwd"" literature,
in which lexicographers, such as Ibn Qutaybah,
born in Iraq in AD 828,^" attempted to record
the Bedouin association of meteorological phe-
nomena with the anwd" star groups associated
with the 28 lunar mansions. Such works would
usually contain an explanation of the anwd'^ star
groups and lunar mansions, the dates of the
beginning of visibility in the east in the morning
of each star group, and an explanation of the
system of rains, winds, cold, and other weather
conditions associated with these risings, illus-
trated with appropriate proverbs, poetry, and
folklore.^' A second type of literature concerned
with the anwd""-mandzil system was arranged in
the form of a calendar and enumerated natural,
celestial, and meteorological events of concern
to peasants and herdsmen.^^
  About the same time that the lexicographers
were recording this Bedouin material in the
ninth century AD, the first listing of these 28
mansions by an Arabic astronomer was made by
al-Fargham.^^ Astrologers also became seriously
interested in the division of the zodiac into lunar
mansions and the assignment of good or ill char-
acteristics to each. In the tenth century al-SQfT
used much of the anwd"' literature recorded by
the lexicographers for his discussion of the fixed
stars, and identified these star groups in terms of
the stars cataloged by Ptolemy.
  In the eleventh century al-BTrQnT presented in
his Chronology of Nations, a lengthy discussion of
lunar mansions. In this work he criticizes the
traditional Arab use of the rising and setting of
star groups to determine the mansions, which
cannot remain accurate because of the preces-
sion. When criticizing the practice of an astron-
omer of Rayy in Persia, al-BTrunT says:
That the truth is the very reverse of his theory, that the
nature and peculiarities which are attributed to the first
Station [Mansion], and that which the Hindus relate of the
connection of this Station with others, are peculiar to the
first part of [the House of] Aries, and never leave this place.

although the star or stars which form the Lunar Station may
leave it. In a similar way, all that is peculiar to Aries does
not move away from the place of [the House of] Aries,
although the constellation of Aries does move away.^^
  Al-BTrunT then outlines a modification of the
anwd^-mandzil system, which he considered bet-
ter since it is not affected by the precession of
the equinoxes. It became the system used in most
astrological prognostication. Each House or 30°
interval of the zodiac is divided into two and one-
third lunar mansions, A rising of a lunar mansion
means that the sun is in that particular segment
of the zodiac, while the second mansion preced-
ing it rises in the east between the beginning of
dawn and the rise of the sun. That is to say, since
the sun's brightness makes it impossible to see
the actual rising of a lunar mansion occupied by
the sun, the observer notes, between the begin-
ning of dawn and the appearance of the sun, the
rising of the second mansion preceding the lunar
mansion interval in the established sequence. For
example, the third lunar mansion is said to be
rising if the first mansion is the last one whose
rising is visible before the sun rises.
  The order of succession of the mansions is the
same as that of the zodiac, that is from one to
the one east of it. For example, sharatayn to
butayn, Pleiades to Aldebaran. The term naw"',
which applies to the rising of a lunar mansion is
also given to the influence of its cosmical setting,
which occurs at a six-month interval from its
rising. The intervals of the ecliptic designated as
lunar mansions are of such a length that the sun
remains in each for approximately 13 days,
though there is some slight variation since the
length of mansions is not totally uniform. Each
lunar mansion was given the name attributed to
one of the 28 anwd"" star groups, even though the
segments of the zodiac were no longer in close
proximity to that anwd"" group.
  On the Smithsonian globe the names of the
lunar mansions are written at intervals along the
ecliptic of about 13°, indicated by a small dash
at right angles to the ecliptic. This clearly indi-
cates that the globe was made in a tradition which
interpreted the anwd^-mandzil system as seg-
ments of the zodiac, and not in the earlier man-
  
NUMBER 46

121



ner in which the rising of the star groups them-
selves was the the determining factor. In most of
the Islamicate astronomical-astrological litera-
ture the 28 lunar mansions are numbered with
the first corresponding to the vernal equinox,
and the second one occurring about 13° later
further into the zodiacal House of Aries, On the
Smithsonian globe the first lunar mansion in the
usual order {sharatayn) is placed at 13° House of
Aries, while the twenty-eight {rishd") is at the
vernal equinox. This identical arrangement is
also to be found on globe No. 11, made by
Qa^im Muhammad of the Lahore workshop,
probably the same maker who produced the
Smithsonian globe. On other globes, such as No.
74, No. 60 (Figure 7), and No. 59 (the earliest
of the Class B globes dated 535 H/AD 1140-
1141), the usual order of having sharatayn at the
vernal equinox is maintained. There was also a
tradition in Islamicate and in later Latin treatises
of associating with the lunar mansions abstract
patterns of dots or stars in small geometrical
designs frequently having little to do with the
actual conformation of the original asterisms.^^
One extant Islamicate globe (No. 60, see Figure
7) dated 718 H/AD 1318-1319 and signed by
■^Abd al-Rahman ibn Burhan al-MawsilT (authen-
ticity questionable), represents the lunar man-
sions by similar patterns of inlaid silver dots along
the ecliptic, apparently in this same tradition.
  Many of the Arabic terms applied to the lunar
mansions were so ancient by the time the lexi-
cographers recorded them in the ninth century
that their significance had been lost even at that
time. For this reason only a tentative translation
can frequently be given. The names appear to
be older than many of the other traditional Ar-
abic pre-Islamic star names, and the imagery
behind them more obscure, often apparently
coming from a very ancient poetry.^^ In the
following descriptions each lunar mansion will
be given in the commonly accepted order, with
the position along the ecliptic where it is en-
graved on the Smithsonian globe. At the end of
the discussion of each mansion the position ac-
cording to al-BTrunT^' will be given for compari-
son with that on the Smithsonian globe. See the

photographs of the corresponding zodiacal
houses in the subsequent section for illustrations
of the distribution of the lunar mansion names
along the ecliptic.
Lunar Mansion 1.    Sharatayn
(Engraved at 13° House of Aries)
  The word sharat can mean "sign" or "signal"
or "the beginning of a thing," and sharatayn
means "the two sharat." Al-QazwTnT (Ideler, 134)
says the two stars of this lunar mansion are named
"the two signs" because they mark the beginning
of the new year. Among the Bedouins these two
stars were also called qarnd al-hamal (the horns
of the ram), referring to an older indigenous
concept of a young ram that was not the same as
the Ptolemaic form of Aries, Ibn Qutaybah says
this lunar mansion was also called al-nath, al-
ndtih, or al-natih, all meaning the act of butting
(Kunitzsch [1961], nos, 197-200), Al-SQfT {Su-
war, 142), however, was understandably con-
fused by the two traditions and at one point
equated the lunar mansion sharatayn with the
two bright stars on the horns of the Ptolemaic
ram Aries, nos. 1 and 2 [fiy Arietis] (see Kunitzsch
[1961], no. 286). Al-SQfT also said another tra-
dition indicated that the "two signals" were made
up of one of the two on the horns of Aries (no,
1) and the first unformed star of Aries, at the
top of the head [a Arietis]. Al-BTrQnT {Chron.
[trns.], 343) says that sometimes three stars are
included in the mansion [afiy Arietis] in which
case it was usually called al-ashrdt, which is the
plural instead of the dual of sharat. Al-BTrQnT
gives as the position of this mansion along the
ecliptic 0° House of Aries, i.e., the vernal equi-
nox.
Lunar Mansion 2.    Butayn
(Engraved at 25'/2° of the ecliptic)
  The name means "the small belly," being as al-
BTrQnT points out a diminutive form of batn
(belly), so as to contrast with the twenty-eighth
lunar mansion, sometimes called "the belly of the
  
122

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



fish." Al-SQfT {Suwar, 142) adds that it is some-
times called simply batn, while some anwd"' au-
thors (Kunitzsch [1961] no. 63) called it batn al-
hamal (the belly of the ram).
  There is not complete agreement on the iden-
tification of the stars comprising the second lunar
mansion with those of the Ptolemaic catalog. Al-
SQfT and later al-BTrQnT identify al-butayn with
three stars that form a triangle in the Ptolemaic
Aries: one near the outgrowth of the tail (no. 7),
the first of the three in the tail (no. 8), and one
on the back of the thigh (no. 11), which are edp
Arietis, respectively, Al-SQfT also says that some
anwd"" authors (Kunitzsch [1961], no, 68) align
this mansion with four stars to the west of al-
mankib, the star on the shinbone of the Ptolemaic
Perseus [^ Persei] or according to some, the two
on the lower foot [of Persei]. These four stars
would, according to al-SQfT, be nos, 2-5 of the
unformed stars of Aries, the four over the rump
of Aries near Perseus (Flam. 41, 39, 35, 36
Arietis). Al-BTrQnT gives the position of the second
lunar mansion as 12°51'26" of the House of
Aries.
Lunar Mansion 3.    Thurayyd
(Engraved at 38'/2° of the ecliptic)
  The third lunar mansion is named for the
famous open star cluster called the Pleiades lo-
cated in the Ptolemaic constellation of Taurus,
of which six or sometimes seven stars are visible
with the naked eye. The Arabic name al-thurayyd
is a very old Arabic star name whose origin and
etymology are obscure (see Hommel, 595; Kun-
itzsch [ 1961 ], no. 306). Al-SQfT {Suwar, 151) says
that they called it al-thurayyd (possibly translata-
ble as "the precious gem") because the Arabs
were blessed by these stars and their risings. The
word thurayyd can also mean "lustre," and comes
from a root meaning "to be abundant" or "to
increase." Al-BTrunT {Chron. [trns.], 343) says
"some people maintain they were called so be-
cause the rain, which is brought by their naw"",
produced tharwa, i.e. abundance."
  In both the Iliad (XVIII, 486) and the Odyssey
(V,    272),    Homer    mentions    the    Pleiades

{irXeLabes). His near contemporary, Hesiod
{Works and Days, 615, 383), states that the cos-
mical setting of the Pleiades in the Fall was a
time for sowing and rough seas, while their helia-
cal rising in the Spring was a time for harvest
and the beginning of the sailing season. Hesiod
adds that they are the daughters of Atlas, for the
Pleiades were commonly held to be the daugh-
ters of Atlas by Pleione, hence the name Pleiades;
another etymology put forward in classical liter-
ature was that the word Pleiades came from TrXeiv
(to sail). Aratus (lines 253-268) gave the names
by which the poets know the seven small stars:
Alcyone, Merope, Celaeno, Electra, Sterope,
Taygete, and Maia, but adds that only six can be
observed (lines 254-255). Hipparchus {Comm.
arat., I, 6.14) criticizes Aratus for this statement,
saying that on a clear moonless night a person
can observe seven stars. In the work attributed
to Eratosthenes {Catast, 14 and 23) the legend
is put forward of seven stars, one of which is
obscure {capavrjs) because six of the daughters of
Atlas and Pleione mated with immortals, while
the seventh, Merope, was wife to the mortal
Sisyphus and through repentance hides from
sight. The Latin poet Hyginus in his Astronomica
(II, 21) gives an additional story that the obscure
star might be the daughter Electra who does not
appear because of grief over the fall of Troy.
This myth of the "lost Pleiad" was repeated by
Ovid {Fasti, IV. 170-178) and others. (For a
history of the legend of the Pleiades and the "lost
Pleiad" see Martin [1956] 72-79 and 89-94.)
  Even though the poetical and literary sources
available in Greece, as well as the one extant
work of the astronomer Hipparchus, clearly show
a knowledge of six or seven Pleiades, Ptolemy
nonetheless lists only four stars as comprising the
Pleiades in his catalog (nos. 30-33 of the con-
stellation Taurus). Al-SQfT and Ulugh Beg both
followed Ptolemy in listing only four Pleiades in
their catalogs. Al-SQfT, however, explained
clearly that there were two or three additional
stars nearby (see the discussion of Figure 64 for
al-SQfT's statement and a discussion of the
Pleiades on the Smithsonian globe). Al-BTrQnT,
on the other hand, said specifically that al-thu-
  
NUMBER 46

123



rayyd (the Pleiades) consisted of six stars close
together, resembling a bunch of grapes, incor-
rectly called seven by the poets {Astrol., sec. 164),
Al-BTrQnT {Chron. [trns,], 343-344) added that
"Ptolemy mentions only four stars of the Pleiades
since he had not observed more of them, because
to eye-sight they seem to lie quite close together,"
  In some anwd" literature the Pleiades were
called alyat al-hamal (the tail of the ram). This
title appears to refer to an early Bedouin concept
of a ram that differed in outline from the Ptole-
maic Aries (see Kunitzsch [1961] no. 9; al-BTrunT,
Chron. [trns.], 343), and which was displaced by
the introduction of the Greek constellations into
the Islamic world. Al-SQfT seems, however, to be
aware only of the Ptolemaic Aries and criticizes
the anwd" authors for using such a term in ref-
erence to the Pleiades.
  In the Bedouin tradition, however, al-thurayyd
(the Pleiades) was most commonly associated with
the head of a woman (also named al-thurayyd)
who had one arm passing through the area of
Perseus (Figure 57) with a henna-stained hand in
Cassiopeia (Figure 56). Her other hand was in
the area to the southwest where the head of Cetus
is now visualized (Figure 75).
  In keeping with this image of a woman, many
other stars were given special names by the early
Bedouins. Thus we have the following star
names:
  In the shoulder-blade of al-thurayyd {"dtiq al-
thurayyd) are two stars on the lower foot of the
Ptolemaic Perseus, which according to al-SQfT
are nos. 25 and 26 of Ptolemy's list [of Persei].
Ibn Qutaybah says it is one "not bright" star
between al-thurayyd and the shoulders, which for
him are the two in the lower foot of Perseus; to
this interpretation al-SufT objects (see Kunitzsch
[1961], no. 41).
  According to al-SQfT "the shoulder [of the
extended right arm] of al-thurayyd" {mankib al-
thurayyd ) is the star in the shinbone of the lower
leg of Perseus, immediately above the two com-
prising the shoulder-blade. Al-SQfT identifies the
star with no. 24 of Ptolemy's catalog [| Persei].
Ibn Qutaybah says, however, "the shoulder" con-
sisted of two stars identified as of Persei, that is,

the two on the foot, which al-SQfT had called "the
shoulder-blade" (Kunitzsch [1961], no. 157).
  "The upper arm of al-thurayyd" {"adud al-thu-
rayyd) is for both al-SQfT and Ibn Qutaybah the
row of three stars, two of which are in the upper
leg and thigh closest to the ogre's head, and the
third one the most northern of the three small
stars on the stomach of Perseus. Al-SufT identifies
these as nos, 10, 22, and 23 of Perseus [bve Per-
sei]. This group is not labeled on the Smithsonian
globe.
  The elbow of the outstretched arm of al-thu-
rayyd {mirfaq al-thurayyd) is the large star in the
stomach of Perseus (no. 7 of the constellation [a
Persei^. The middle of the three small stars on
the stomach of Perseus is called traditionally ibrat
al-mirfaq (the point of the elbow of al-thurayyd),
which is aligned by al-SQfT with no. 9 of Perseus
[\p Persei]; it is not labeled on the Smithsonian
globe.
  The last and lowest of the three small stars
close together on the stomach of Perseus is called
ma^bid al-thurayyd (the fatty part [of the upper
arm] of al-thurayyd). Al-SufT knew only one star
by that name, and identified it as no. 8 of Perseus
[a Persei]. Ibn Qutaybah, on the other hand, saw
two stars in that area standing near one another,
between which was an estimated visual distance
of one dhird", which for al-SQfT was about 2° 20'
(See Kunitzsch [1961] no. 152). This label does
not appear on the Smithsonian globe.
  The forearm, sd"id al-thurayyd, was made up,
according to al-SQfT, of the star in the upper arm
of the arm raised above Perseus's head and the
large star immediately below on the chest, which
al-SQfT identified with nos. 2 and 3 of the con-
stellation [777 Persei]. Ibn Qutybah, however, saw
three stars as the forearm, which he called al-
dhird" (forearm or cubit), instead of al-sd"id (Kun-
itzsch ([1961], nos. 81 and 258). These three
stars were between the elbow and the wrist,
perhaps including the small star in the beard of
Perseus on the Smithsonian globe [no. 5 of Per-
seus; T Persei]. This label also does not appear on
the Smithsonian globe.
  The wrist of the right outstretched arm is
called mi"sam al-thurayyd, and is represented by
  
124

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



the two open star clusters marked as one star in
the hand of Perseus over his head [xh Persei].
See the discussion of Perseus and Figure 57 for
further details.
  The hand of the right arm was visualized as
spreading out from the wrist (hand of Perseus)
toward Cassiopeia, with the fingers being repre-
sented by the five bright stars of Cassiopeia form-
ing the well-known W-shaped asterism. This
hand was called al-kaff al-khadib, that is, a hand
dyed with henna {hinnd"), a red dye made from
Lawsonia inermis. The expression "the dyed
hand" referred originally to five stars (nos. 12,
2, 4, 5, 6 of the Ptolemaic catalog [fiayde Cassio-
peiae]; see Figure 56). However, as al-SQfT noted
{Suwar, 77) it sometimes happened that astrono-
mers would take a name originally used for a
group of stars and apply it to only one of them;
in this case it was applied to no. 12, the star on
her back elbow [fi Cassiopeiae], as a star for use
on astrolabes. For this reason the label on the
Smithsonian globe appears to apply to only that
one star rather than the entire group of five.
  The other hand of their early conception of a
female figure centering around the Pleiades is
found in the head of the Ptolemaic constellation
Cetus, the sea monster or whale (see Figure 75).
It is called al-kaff al-jadhmd" (the cut-off hand)
probably because no star groups connect the
hand with al-thurayyd, the Pleiades, representing
the head. Al-SQfT identifies the cut-off hand with
the first six stars of the constellation Cetus
[Xaybvp. Ceti].
  To the third lunar mansion, al-thurayyd, al-
BTrunT assigned the position of 25° 12'50" in the
House of Aries.
Lunar Mansion 4.    Dabardn
(Engraved at 51 V2° of the ecliptic
or 21 '/2° House of Taurus)
   Al-Dabardn is from a root meaning "to follow,"
and refers to the fact that it follows the Pleiades,
for which reason it was sometimes called tdli al-
najm "the follower of The Star," since the
Pleiades were occasionally called "the star {al-
najm). Al-SQfT and the anwa" authors (Kunitzsch
[1961], no. 69) referred to al-dabardn as a large

bright red star that corresponded to the one in
the eye of the Ptolemaic constellation of Taurus,
no. 14, which is the famous star a Tauri, called
today Aldebaran (see Figure 64). Aldebaran is
the most prominent of the open cluster compos-
ing the asterism of the Hyades,
  The Bedouins also called the star comprising
the fourth lunar mansion al-faniq (the camel-
stallion), while the stars around it (the other
Hyades) were called al-qilds (the young camels;
see al-BTrQnT, Chron. [trns,], 344). Al-BTrunT gives
the position of the fourth lunar mansion as
8° 34'18" of the House of Taurus, which is
38° 34'18" from the vernal equinox.
Lunar Mansion 5.    Haq"ah
(Engraved at 63° of the ecliptic,
or 3° House of Gemini)
  The word haq"ah means a tuft of hair, a brand-
ing mark usually on the neck or breast of a horse,
or any other distinguishing mark of a horse. Al-
BTrQnT said the stars of this lunar mansion were
called haq"ah because they were compared with
a circles of hairs on the leg of a horse near the
foot; a horse having such a distinguishing mark
was called mahqu" which is from the same root as
haq"ah. This lunar mansion was said to corre-
spond to the first star of the Ptolemaic constel-
lation of Orion. In his catalog Ptolemy clearly
considered this to be a single nebulous
{veipeXoeibrjs) star in the forehead of Orion, while
the anwd" authors (Kunitzsch [1961], no. 115a),
al-SQfT {Suwar, 268), and al-BTrQnT {Chron.
[trns,], 344) saw three small stars next to one
another like a small triangle [Xip^ip^ Orionis].
These stars were sometimes also called al-athdfi
(a three-legged support for a cooking pot) be-
cause they resembled such a tripod. Al-SQfT adds
that the three stars were occasionally called
haq"at al-jawzd", the tuft of hair or distinguishing
mark of jawzd" (the Bedouin name for a giant,
possibly feminine), visualized in the area we con-
sider occupied by the Ptolemaic constellations of
Orion and Gemini. (See the discussion of Orion
[Figure 76], and the section on "Classical Greek
and Pre-Islamic Sources.")
  
NUMBER 46

125



  Al-BTrunT gives the position of the fifth lunar
mansion as 21 ° 25'44" of the House of Taurus,
or 51°25'44" from the vernal equinox.
Lunar Mansion 6.    Han"ah
(Engraved at 77° of the ecliptic,
or 17° House of Gemini)
  The root of this word can mean either to fold
or bend, or to brand a camel on the neck. It is
from the former meaning that al-BTrunT {Chron.,
344) derived his explanation of the word "as if
each of them were winding and twining round
the other." Han"ah is more often translated as a
brand on a camel's neck, maintaining a parallel
with the fifth mansion, and was traditionally ap-
plied to two stars (nos. 17 and 18 of the Ptolemaic
constellation of Gemini [y^ Geminorum]; see Fig-
ure 65). These two stars mark the feet of the
more southern of the two twins. Both han"ah and
haq"ah (the sixth and fifth mansions) are in the
area of the sky called al-jawzd" by the Bedouins,
which covered the area of Ptolemy's Orion and
part of Gemini, and seems to have been a giant
form, possibly feminine. The star y Geminorum,
whose modern name is Alhena from the name of
the sixth mansion, was also commonly called al-
zirr, which has many meanings including button,
socket of the thigh, a pivot, or a bud of a plant;
the intent is obscure. The second star [^ Gemi-
norum] bore the additional title of al-maysdn (a
bright star). The traditions regarding these two
names are not uniform (see Kunitzsch [1961],
no. 321).
  Some people in the anwd" tradition called the
three stars standing before y^ Geminorum (the
three in the feet of the northern twin [nos. 14,
15, 16; rjnv Geminorum]) al-tahdyi, whose precise
spelling is not uniform nor its meaning clear.
They then termed all five stars together al-
han"ah (Kunitzsch [1961], nos. 114, 296), Some,
however, interpreted al-tahdyi (or al-tahiyydt or
al-tahiyyah as it sometimes was written) as the
name of the three comprising the fifth lunar
mansion, al-haq"ah (al-SufT Suwar, 166-167),
  Al-SufT also states that the five stars in the feet
of the twins were called qaws al-jawzd" (the bow

of al-jawzd") which he throws at the foreleg of
the lion. The anterior of the two forelegs of the
huge lion of the early Bedouin traditions was
formed by the two bright stars in the heads of
the Ptolemaic Gemini; this foreleg constituted
the seventh lunar mansion.
   It is possible that at one time han"at al-jawzd"
and haq"at al-jawzd" were combined into one
lunar mansion (see Hommel, 601-602), Al-BT-
runT gives han"ah the position of 4 ° 17' 11" House
of Gemini, or 64° 17'11" from the vernal equi-
nox.
Lunar Mansion 7.    Dhird"
(Engraved at 90° of the ecliptic,
or 0° House of Cancer, at the
point of the summer solstice)
  "The foreleg" (of the lion) was a name applied
to the two stars in the heads of the Ptolemaic
Gemini [afi Geminorum]. In the anwd" tradition
these two stars were seen as forming one of the
forelegs of an enormous lion, which covered an
area of the heavens much larger than the Ptole-
maic Leo. The lunar mansions numbered 7-11
and 14 all reflect the image of this large lion.
  The traditions were confused as to whether
the "foreleg" formed by the two stars in Gemini
was the "drawn up" leg {al-maqbudah) and the
other foreleg, which was formed by the two stars
of Canis Minor [afi Canis Minoris] was the "ex-
tended" leg {al-mabsutah), or whether it was the
other way around. There was, moreover, disa-
greement as to which of these two groups of
stars, those in Gemini or those in Canis Minor,
constituted the lunar mansion. Authors fre-
quently confused these names (Kunitzsch [1961],
no. 83); al-BTrQnT {Chron. [trns,], 345) adds that
some interpreted the dhird" or seventh lunar
mansion as being the two Sirii, that is Sirius [a
Canis Majoris] and Procyon [a Canis Minoris].
  Al-SufT {Suwar, 165) astutely and logically ar-
gues that the groups of stars that rise first should
be the "extended" leg. Since the northern dhird"
[afi Geminorum] rises before the southern dhird"
[afi Canis Minoris], the former rightly should be
termed "the outstretched" and the latter "the
  
126

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



drawn up." In addition al-SQfT argued that since
the northern dhird" [afi Geminorum] was closer
to the ecliptic it must be the lunar mansion. Al-
BTrQnT gives the position of the seventh lunar
mansion as 17°18'35" of the House of Gemini,
which is 77° 18'35" from the vernal equinox.
Lunar Mansion 8.    Nathrah
(Engraved at 103° of the ecliptic,
or 13° House of Cancer)
   Al-nathrah means "the cartilage of the nose,"
belonging to the image of the large lion of Bed-
ouin tradition. Al-BTrQnT {Chron. [trns.], 345;
Astrol., sec. 164) applies this name, apparently
using it in the broader sense of nose, to three
stars in a row on the back of the Ptolemaic
constellation of Cancer, the outside two stars
being yd Cancri and the middle one the open
cluster M44, called Praesepe or the Beehive,
which was considered in both the Greek and
Arabic traditions to be nebulous.
  Al-SQfT, on the other hand, uses the word
nathrah only for "the smudge {latakhah) which
resembles a bit of a nebula"—that is, for Praesepe,
which was considered the first star of the con-
stellation of Cancer. The two stars on either side
[yb Cancri] he said {Suwar, 173) were called al-
mankhardn (the two nostrils); the "smudge" in
between was also named al-mukhatah (the mu-
cus). The three stars together, according to al-
SQfT, were traditionally called fam al-asad (the
mouth of the lion). Al-BTrunT and al-SQfT record
that this open cluster M44 was also traditionally
called the uvula {al-lahdh) of the large lion. The
position given by al-BTrQnT for the eighth man-
sion is 0°0'0" of the House of Cancer, or the
point of the summer solstice.
Lunar Mansion 9.    Tarf
(Engraved at 11574° of the ecliptic,
or 25y4° House of Cancer)
  The word tarf means the vision or glance (also
of the large lion). It was applied by al-SQfT (5M-
war, 173, 181) to two stars: the second formed

star of the Ptolemaic Leo, which is the lower of
the two on the mouth [X Leonis], and the second
unformed star of Cancer [K Cancri]. This tradi-
tion was also followed by al-BTrQnT (cf Kunitzsch
[ 1961 ], no. 304a; Hommel, 602). Navigators des-
ignated the two stars "ayn al-asad, "the eye of the
lion" (Kunitzsch [1961], no, 46a). Al-BTrunT as-
signed the position of 12°51'26" House of Can-
cer, or 102°15'26" of the ecliptic, to the ninth
lunar mansion.
Lunar Mansion 10.   Jabhah
(Engraved at 128'/2° of ecliptic,
or 8'/2° House of Leo)
   Al-jabhah (forehead), sometimes called jabhat
al-asad (the forehead of the lion), was a name
applied to four stars collectively and referred to
the forehead of the large lion of the Bedouin
tradition (Kunitzsch [1961], nos, 103a, 103b),
These four stars are the two on the shoulder of
the Ptolemaic Leo (nos, 5 and 6), the one to the
southwest (no, 9), and the large star on the upper
part of the front leg of Leo (no. 8 called Regulus;
see Figure 67). These are ^yvot Leonis, respec-
tively. The position of the tenth mansion, ac-
cording to al-BTrQnT, should be 25°12'52" of the
House of Cancer or 115°12'52" from the vernal
equinox.
  Lunar mansions seven through ten were also
commonly called kildb al-shitd" (the dogs of the
winter or rain) (Kunitzsch [1961], no. 147).
Lunar Mansion 11.    Zubrah
(Engraved at 141'/2° of ecliptic,
or 21'/2° House of Leo)
  Al-SQfT says {Suwar, 181) of zubrat al-asad that
it is "the mane of the lion, that is, his withers and
shoulders" and identifies the mane of the large
lion as nos. 20 and 22 of the Ptolemaic Leo [bd
Leonis], which are the large star on the rump of
Leo and the star at the top of the thigh (see
Figure 67). See al-BTrQnT {Chron. [trns.], 346) for
further explanation of the word zubrah.
  
NUMBER 46

127



  These two stars were also called al-khardtdn,
whose meaning is unclear; al-SufT says the sin-
gular is khard, so that al-khardtdn would mean
the two khard. Al-BTrQnT apparently takes the
word from the root khrt, meaning to pierce with
a hole, for he says they are called the khardtdn
"as if each of them were penetrating into the
interior of the lion" (cf. Hommel, 603 note 2;
Kunitzsch [1961], no. 128).
  Al-BTrunT assigns the tenth mansion the posi-
tion 8°31'18" of the House of Leo, or
128°31'18" from the vernal equinox.
Lunar Mansion 12.    Sarfah
(Engraved at 154'/4° of ecliptic,
or 4'/4° House of Virgo)
  The term sarfah (change [of weather]) was
applied to a bright star behind the eleventh lunar
mansion; al-SQfT identified the star as the one in
the tail of the Ptolemaic Leo, no, 27 [fi Leonis].
The star was called sarfah, according to al-SQfT
{Suwar, 181) and al-BTrQnT {Chron. [trns.], 347)
because of the change in the weather from heat
with its rising at dawn before the sun and the
change from cold weather with its setting at
dawn. As part of the image of the larger lion of
the Bedouin tradition, this star was called "the
seed of the lion'' {qanb al-asad) according to al-
SQfT, who added by way of explanation "that is
to say, his scrotum" {wi"d" qadibihi). The tail of
the large Bedouin lion was formed by the stars
of the Ptolemaic Coma Berenices, called in Ara-
bic al-hulbah, meaning the coarse hairs at the
end of the tail, Al-BTrQnT assigned the position of
21°25'44" House of Leo, or 141°25'44" from
the vernal equinox, to the twelfth mansion.
Lunar Mansion 13. "Axvwd"
(Engraved at I67'/4° along ecliptic,
or 17'/4° House of Virgo)
  The meaning of the name "awwd" is uncertain,
but it appears to be from a root meaning to howl
or yelp, or to twist or bend. According to al-SufT
the name was applied to five stars in the Ptole-

maic constellation of Virgo, which formed
roughly a 90° angle (see Figure 68). These five
are the one at the top of the southern wing, the
one on the southern shoulder, the one at the side
of the waist nearest the ecliptic, the large one in
the middle of the waist, and the large one on the
northern wing near the elbow, nos. 5, 6, 7, 10,
and 13, respectively [fi-qybe Virginis]. Al-SQfT {Su-
war, 193) adds, however, that some people omit-
ted no, 10, the one in the middle of the waist [b
Virginis] and thus recognized only four stars in
the thirteenth lunar mansion.
  Al-BTrQnT {Astrol., sec. 164) follows the latter
tradition of four stars and accounts for the name
"awwd" by saying that the Arabs spoke of dogs
barking behind the large lion, thus interpreting
"awwd" as howling. Al-SQfT too had mentioned
that some saw four or five dogs here following
the lion, and also added that the lunar mansion
was sometimes called "awwd" al-bard "the "awwd"
of the cold." The tradition of five stars, however,
al-BTrQnT presents in another writing {Chron.
[trns.], 346, 351) where he takes "awwd" to mean
"to turn" because they are five stars in a line, the
end of which is turned to form the 90° angle.
See Hommel (604 note 1) and Kunitzsch
[1961], no. 44 for further information on the
meaning of "awwd". Al-BTrQnT gives the thirteenth
mansion the position of 4°17'10" in the House
of Virgo, or 154° 17'10" from the vernal equi-
nox.
Lunar Mansion 14. Simdk
          (Engraved at 180° of ecliptic,
or 0° House of Libra, at the autumnal equinox)
  Two Stars each bear the name simdk: the large
star on the southern hand of Virgo [a Virginis;
Spica] and the large star at the hemline of Bootes
[a Bootis; Arcturus], which is due north of Spica
(see Figures 53 and 68). The name simdk is of
ancient, perhaps Babylonian, origin. Its meaning
has been obscured with time; hence it is impos-
sible to give it an intelligible translation (see
Hommel 595-596; Gundel &: Boker, cols. 509-
522; al-BTrQnT, Chron. [trns.], 347). These two
stars, each called simdk, formed the two hind legs
  
128

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



of the large lion visualized in the anwd" tradition.
Only the one in Virgo, however, constituted a
lunar mansion. The full name of this star was al-
simdk al-a"zal (the unarmed simdk), which distin-
guished Spica from Arcturus, which bore the
title al-simdk al-rdmih (the armed simdk). Arctu-
rus was considered "the armed simak" because
there are stars near it forming a lance or weapon,
while no such weapon was seen to accompany
Spica. As a lunar mansion rather than a star,
however, Spica was mostly referred to simply as
al-simdk. The fourteenth lunar mansion is given
the position of 17°8'35" of the House of Virgo,
or 167°8'35" from the vernal equinox.
Lunar Mansion 15. Ghafr
(Engraved at 193° along ecliptic,
or 13° House of Libra)
  The meaning of the name al-ghafr is not at all
clear; it was applied to three stars in the Ptole-
maic constellation Virgo: the star in her foot
nearest the ecliptic and the two at the hemline,
nos. 22, 23, and 25, [LKX Virginis]. Al-BTrQnT
{Astrol., sec. 164) states that only the two stars
on the hemline {dhayl) constitute the lunar man-
sion, while elsewhere {Chron. [trns.], 347) he has
the more common tradition of three inconspic-
uous stars.
  Many etymologies are presented in the Arabic
astronomical literature for the word ghafr, the
most common being that the name was applied
because the stars were inconspicuous. Al-SQfT
{Suwar, 194) in his discussion of Virgo says of
this mansion "they also say that it is called al-
ghafr, for, because of the imperfection of the
light of its stars, they say it was veiled (from the
root ghfr)—that is, it was obscured." Shortly
thereafter he continues "they also say it is called
al-ghafr because it is above the two claws of the
scorpion, and for this reason it is called al-mighfar
(the helmet) which is over the head of a man."
Another derivation is that the root means to have
a covering of hair, that is, it is seen as the coarse
hair which is at the tip of the lion's tail. This
explanation, however, shows confusion in the use
of the star names, for the hair at the end of the

lion's tail was represented by the group of stars
called al-hulbah (coarse hair) in the anwd" tradi-
tion, which corresponded to the Ptolemaic aster-
ism of Coma Berenices translated into Arabic as
al-dafirah, meaning a lock of human hair. There
appears to have been some confusion in the
names applied to Coma and those of the four-
teenth and fifteenth lunar mansions,
  Al-BTrQnT {Chron. [trns.], 347) briefly gives all
three derivations, adding that al-ghafr was con-
sidered the best of the lunar mansions because it
stands behind the lion and before the scorpion.
"The evil of the lion lies in his teeth and claws,
the evil of the scorpion lies in its venom and the
stingof its tail." He says it was commonly asserted
that the horoscopes of all the prophets lay in this
mansion—a point that he disputes. He gives the
position of the fifteenth mansion as 0°0'0" of
the House of Libra, which is the autumnal equi-
nox.
Lunar Mansion 16. Zubdnd
(Engraved at 205y4° along ecliptic,
or 25%° House of Libra)
   Zubdnd (the two claws [of the scorpion]) ap-
plied to the two large stars, each of which is in
one of the pans of the balance of the constellation
Libra [a''^ fi Librae]. This lunar mansion reflects
the more ancient Babylonian concept of a scor-
pion larger than the now familiar Scorpio, of
which the constellation now known as Libra
formed the claws (Hartner [3], 501; Gundel &
Boker, 534-535). Ptolemy in his catalog did not
call this constellation a balance but rather the
claws {xvXotl) of the scorpion, and so these two
were nos. 1 and 2 in his listing for the claws. Not
long after Ptolemy, however, the concept of a
balance in this area of the heavens was well
established and firmly superimposed upon the
earlier claws of a large scorpion. In the Arabic
tradition the older Bedouin names reflect this
ancient image of a large scorpion. See Hommel
(597) and Gundel and Boker for a discusion of
the origin and meaning of the word zubdnd. Al-
SQfT {Suwar, 202) adds that the two stars were
called also yadd al-"aqrab "the two hands of the
  
NUMBER 46

129



scorpion" and qarnd al-"aqrab "the two horns of
the scorpion," The position given by al-BTrunT
for the mansion is 12°51'26" of the House of
Libra, or 192°5r26" from the vernal equinox.
Lunar Mansion 17. Iklil
(Engraved at 218y4 ° of ecliptic
or 8V4° House of Scorpio)
  The tradition is not consistent with regard to
this lunar mansion, which is called al-iklil (the
crown), Al-SQfT {Suwar, 202-204) in his discus-
sion of the constellation Libra (see Figures 69
and 70) says that opinions differ among the Arabs
with regard to the seventeenth lunar mansion,
and presented five different interpretations of
iklil.
  (1) Three stars across the bar supporting the
scales of Libra; the one at the southern end of
the bar was not observed by Ptolemy and is not
indicated on the Smithsonian globe. These three
stars are no, 8 of the formed stars of Libra at the
north end of the bar [6 Librae]; no, 6 of the
unformed stars of Libra above the bar [K Librae,
or according to Peters &c Knobel it is O'' Arg,
14782, while the fifth unformed is AC Librae], and
the one not observed by Ptolemy at the southern
tip. On globes Nos, 4 and 5 the word al-iklil is
engraved at the head of the scales, indicating that
this interpretation was in the mind of the makers.
  (2) Three stars in a row in the constellation of
Scorpio, nos. 1-3 of the formed stars. These are
no. 1 in the middle of the three contiguous stars
on the forehead of Scorpio, no. 3 in the southern
claw, and the star in between, which is no, 2 [fibir
Scorpii, respectively], Al-SufT rejected this inter-
pretation and asserted that iklil must be one of
the other four interpretations that are above the
head of Scorpio, Al-BTrQnT {Chron. [trns,], 348)
objected to al-SufT's position and asserted that
the common view, not withstanding al-SQfT, was
that iklil was composed of these three stars in a
straight line,
  (3) Three stars, one of which is no, 8 formed
of Libra [6 Librae] at the northern end of the bar
supporting the scales; the other two are no, 6
unformed of Libra [K Librae}] above the bar, and

no, 8 unformed of Libra which is the northern
of the two on either side of the southern claw of
Scorpio [Flam, 39],
  (4) Five stars which include the three of the
previous interpretation along with no, 9 of the
unformed stars of Libra, which is the southern
of the two stars on either side of the southern
claw of Scorpio [Flam, 40] and a star of sixth
magnitude not mentioned by Ptolemy near the
no. 6 unformed of Libra near the center of the
bar supporting the scales,
  (5) Five stars, which consist of the star at the
northern end of the bar of Libra [no, 8 formed;
6 Librae] the star above the bar and to the south
[no, 6 unformed; K Librae ?], and three stars not
seen by Ptolemy, all of which are above the three
in the forehead of Scorpio and which were given
as the second interpretation of iklil.
  Al-BTrQnT gives the position of the seventeenth
lunar mansion along the ecliptic as 25°42'52"
House of Libra, or 205°42'52" from the vernal
equinox.
Lunar Mansion 18. Qalb
(Engraved at 231 V2° of ecliptic,
or 21 '/2 ° House of Scorpio
   Qalb (heart) or qalb al-"aqrab (heart of the
scorpion) was the given to the eighth star in
Scorpio, second from the end of the body (see
Figure 70), which is the sixteenth brightest star
in the heavens {a Scorpionis). The notion of the
"heart of the scorpion" arises from the anwd"
tradition and not from the Greek tradition of
Scorpio, in which this star was called Antares,
meaning "similar to Mars" since its color, red,
resembled that of the planet Mars (see Gundel &
Boker, colums 509-522). The position given by
al-BTrunT is 8°34'18" of the House of Scorpio,
or 218°34'18" from the vernal equinox.
Lunar Mansion 19. Shawlah
(Engraved at 244° of the ecliptic,
or 4° House of Sagittarius)
  The title al-shawlah (the raised tail) or shawlat
al-"aqrab (the raised tail of the scorpion) was
  
130

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



applied to the two stars at the tip of the tail of
Scorpio, nos. 20 and 21 [Xv Scorpionis]. Al-SQfT
{Suwar, 209) also gives the name ibrat al-"aqrab
(the sting of the scorpion) as a common designa-
tion of these two stars comprising the nineteenth
lunar mansion. Al-BTrQnT gives it the position of
21°25'14" of the House of Scorpio or
231°25'14" from the vernal equinox.
Lunar Mansion 20. Na"d"im
(Engraved at 257° of the ecliptic,
or 17° House of Sagittarius)
  The title al-na"d"im (the ostriches) was applied
to a group of eight stars identified by al-SQfT
{Suwar, 219-220) as being in the Ptolemaic con-
stellation of Sagittarius. The Milky Way passes
through this constellation across the figure's bow
and arm. In the Bedouin tradition the Milky Way
{majarrah) was viewed as a river with one group
of four ostriches going toward the river and
another group of four leaving the river on the
other side. The arriving ostriches {al-na"d"im al-
wdrid or al-na"dm al-wdrid) were nos. 1,2,3, and
25 of Sagittarius: the one of the arrowhead, the
one in the hand holding the arrow, the one at
the southern tip of the bow and the one in the
hoof of the raised front leg [yber] Sagittarii]. The
departing ostriches {al-na"d"im al-sddir or al-
na"dm al-sddir) were nos. 6, 7, 21, and 22 of
Sagittarius, consisting of the one in the hand
drawing the bow, the one in front of the hand,
and the two below [aipr^ Sagittarii]. The eight
taken together were called the ostriches. Al-BT-
rQnT gives the twentieth lunar mansion the posi-
tion of 4° 17'10" in the House of Sagittarius, or
244° 17'10" from the vernal equinox.
Lunar Mansion 21. Baldah
(Engraved at 270°, or the point of
the winter solstice)
  The area beneath the fluttering band extend-
ing behind the head of Sagittarius was called al-
baldah (the place) and was said to contain no stars
(see Figure 71). Al-SQfT {Suwar, 220) called it an
open space under al-qilddah [i.e., ^oirdpv Sagit-

tarii] in which there is no star. Al-BTrunT {Chron.
[trns.], 348; Astrol., 164) does not mention such
a view of al-SQfT's, but compares this desert area
{mafdzah) to the space between two eyebrows.
Kunitzsch ([1961], no. 51) identifies the area as
that around Flam. 53, x^'^ h'-^ Sagittarii, while
Hartner ([3] 502) says it is near TT Sagittarii. Al-
BTrQnT gives this lunar mansion the position of
17°8'35" House of Sagittarius or 257°8'35"
from the vernal equinox.
Lunar Mansion 22. Dhdbih
(Engraved at 28274° of the ecliptic,
or 1274° House of Capricorn)
  The full name of the twenty-second lunar man-
sion is sa"d al-dhdbih, the meaning of which is so
obscure that it is virtually impossible to translate.
The twenty-second lunar mansion is one of ten
groups of sa"d stars, each of which, with two
exceptions consists of two stars. Four of these
sa"d star groups are lunar mansions; between the
two stars comprising each group there is a dis-
tance of one dhird", which for al-SQfT is about
2° 20'. Of the other six 5a"<i-stars, four are in the
Ptolemaic constellation of Pegasus, one in Capri-
corn, and one in Aquarius. The sa"d is of such
ancient origin that its significance seems to have
been lost by the time the sa"d stars were recorded
by the lexicographers in the ninth century. Hom-
ell (p. 606, notes 3 and 4) traces sa"d back to a
Babylonian word meaning "demon," Occasion-
ally the word has been translated as "lucky star"
but rather inappropriately; "omen" might be
somewhat better. The Islamic writers them-
selves, however, always explained only the adjec-
tive or noun modifying the word sa"d and then
left sa"d unexplained. Such will be the procedure
in this discussion.
  The lunar mansion sa"d al-dhdbih (the sa"d of
the sacrificer or butcher refers to stars in the
constellation of Capricorn. One star is at the base
of the horn closer to Aquarius, no. 1 [a''^ Capri-
corni], while the other is the larger star to the
south, no. 3 [fi Capricorni] (see Figure 72). The
small star standing close to no. 1, the second star
of the constellation [v Capricorni] is called, ac-
  
NUMBER 46

131



cording to al-SQfT {Suwar, 227), shd", the sheep
or goat that he slaughters (see Kunitzsch
[1961], no. 257.5). Al-BTrQnT assigned the
twenty-second mansion the position of 0°0'0"
House of Capricorn, or the winter solstice.
Lunar Mansion 23. Bula"
(Engraved at 29574° along ecliptic,
or 2574° House of Capricorn)
  The full name of the lunar mansion, sa"d bula"
(sa"d [of the] devourer or swallower) refers to
two of the three stars in the hand of Aquarius
extended over the back of Capricorn: no. 6 [v
Aquarii or Flam. 7], which is the star on the lower
palm, and no. 8 [e Aquarii] furthest of the other
two. The star in between, no. 7 [ix Aquarii] is the
star that has been swallowed and is on its way
from the throat to the stomach. Al-SQfT {Suwar,
238) names all three, but al-BTrunT follows Ibn
Qutaybah in having only two stars actually com-
pose the lunar mansion, with the middle one too
obscure to count (Kunitzsch [1961], no. 257.4).
Al-BTrQnT assigned this mansion the position of
12°51'26" in the House of Capricorn or
282°51'26" from the vernal equinox.

Lunar Mansion 25. Akhbtyah
(Engraved at 321 '/4° of the ecliptic,
or 21'/4° House of Aquarius)
   Sa"d al-akhbiyah {sa"d of the tents) is composed
of four stars in the east hand and wrist of Aquar-
ius, nos. 9-12 [yvrfT; Aquarii]. They form a tri-
angle with one star in the middle, while some
early writers (Kunitzsch [1961], no. 257.1; al-
BTrQnT Astrol., sec, 169) noted that the pattern
was that of a duck's foot {rijl battah). It was said
that the central star was sa"d and the three sur-
rounding ones were his tents, which perhaps
gives support to Hommel's contention that the
word sa"d was originally the personified concept
of a demon. Akhbiyah is the plural of khibd" mean-
ing a tent made of wool or camel's hair and
supported by three poles. The root of the word
means "to conceal"; on the basis of this derivation
al-BTrQnT {Chron. [trns.], 439) and al-SQfT {Suwar,
239) said the group was given this name because
when this mansion rises all that has been hidden
in the earth comes out into the rain. These four
stars form the modern asterism of the "Y" of
Aquarius, also called the Water Jar. Al-BTrQnT
gives the twenty-fifth lunar mansion the position
of 8° 34'18" House of Aquarius, or 308° 34'18"
from the vernal equinox.
  

  
Lunar Mansion 24. Su"ud
(Engraved at 308° of the ecliptic,
or 8° House of Aquarius)
  The full name sa"d al-su"ud could be roughly
translated as "omen of good fortune'' and was
applied to two stars on the west shoulder of
Aquarius toward Capricorn, nos, 4 and 5 [fi^
Aquarii], along with a third star in the end of the
tail of Capricorn, no. 28 [c' Capricorni] (see Kun-
itzsch [1961], no 257.10 who corrects the Hy-
derabad text of al-SQfT [Suwar, 238]). According
to al-BTrQnT {Chron. [trns.], 349) the mansion was
given this name "because it rises when the cold
decreases, when the winter is past and the season
of the continuous rains sets in." The position
along the ecliptic given by al-BTrQnT is 25° 42'52"
House of Capricorn, or 295°42'52" from the
vernal equinox.

Lunar Mansion 26. Muqaddam
(Engraved at 334'/2° of the ecliptic,
or 4'/2° House of Pisces)
   Muqaddam (anterior) is part of the term al-
fargh al-muqaddam (the anterior spout), which
refers to a leather bucket envisioned by the
Bedouins in the area of the constellation of Pe-
gasus (see Figure 61). The bucket was formed by
the four bright stars on the body of Pegasus (one
is shared with the head of Andromeda), nos. 1-
4, which form the modern asterism of the Square
of Pegasus. The two foremost of the stars (no. 4
immediately above the wing [a Pegasi] and no. 3
near the join of the wing and lower leg [fi Peg-
asi]) were called "the anterior spout [of the buc-
ket]." Occasionally these two were called al-fargh
al-awwal (the first spout) (al-SQfT, Suwar,  122;
   
132

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Kunitzsch [1961], no. 92a,b). Al-BTrQnT assigned
this mansion the position of 21°25'44" House
of Aquarius, or 321°25'44" from the vernal
equinox.

the fish) (Kunitzsch [1961], no. 217; cf 252), He
assigned this mansion the position of 17°8'35"
House of Pisces, or 347°8'35" from the vernal
equinox.



Lunar Mansion 27. Mu"akhkhar
(Engraved at 347'/4° of the ecliptic,
or l7'/4° House of Pisces)
   Mu"akhkhar is part of the expression al-fargh
al-mu"akhkhar meaning "the posterior spout of
the bucket." It refers to the posterior two stars
of the Square of Pegasus: no. 2, which is on the
corner of the back where it is cut off [7 Pegasi],
and the one shared with the head of Andromeda
[b Pegasi or a Andromedae] These two were
viewed by the Bedouins as the posterior spout of
a leather bucket, and were also called al-fargh al-
thdni (the second spout; see discussion of Lunar
Mansion 26). The position of the twenty-seventh
mansion given by al-BTrQnT is 4° 17'10" House
of Pisces, or 334° 17'10" from the vernal equi-
nox.
Lunar Mansion 28. Rishd"
(Engraved at 360°, or the vernal equinox)
  Al-rishd" (the rope) was one of several names
for the twenty-eighth lunar mansion. It was ap-
plied to the star on the south side of the waist of
Andromeda next to the fish of Pisces [no. 12; fi
Andromedae, Mirach] and was designated a rope
in order to supply a rope for the leather bucket
seen in the area of Pegasus (see Lunar Mansions
26 and 27).
  In a different Bedouin tradition this mansion
was also called batn al-hut (the belly of the fish)
because in the early Bedouin conception of this
region a large fish was seen to lie across what we
now view as Andromeda, making this star the
belly of the fish. This fish was much larger than
the lower one of the two Pisces, which in the
Greek tradition was viewed as being next to
Andromeda. The name batn al-hut is written on
the Smithsonian globe next to this star in Andro-
meda. Al-BTrQnT {Chron. [trns.], 350) adds that
this star was also called qalb al-hut (the heart of

THE FORTY-EIGHT CONSTELLATIONS
Constellation 1. The Lesser Bear
[Ursa Minor]
FIGURE 49
  The Lesser Bear, like the Greater Bear, was
called both a bear and a wagon in Greek times,
the latter probably being the older image. One
can easily observe our asterism of the Little Dip-
per in the seven formed stars comprising the
constellation. The Lesser Bear does not appear
in Homer. Aratus (lines 30-44) recounts the
story of Zeus having as a child been hidden and
nurtured by two bears, the other being the
Greater Bear, for which he rewarded them by
placing them in the heavens. He also says that
the constellation is called Kvvbaovpa, the Cyno-
sura (dog's tail) as well, and that the Phoenicians
sailed by the Lesser Bear, while the Greeks
steered by the Greater Bear. In the Islamic world
the Lesser Bear was also used as a guide in travel.
  The notion of a bear for this constellation as
also for the Greater Bear, unfortunately requires
that a nearly tailless creature have a long, incon-
gruous tail. The depiction of the Lesser Bear on
the Smithsonian globe is with an exceedingly
long tail, even longer than that of the Greater
Bear, giving it a striking resemblance to a bintu-
rong, indigenous to India, rather than a bear.
The Greek notion of the Lesser Bear was super-
imposed in the Islamic world over an earlier Arab
conception of the region. The title of the con-
stellation in Figure 49, written in front of the
bear, reflects the Greek asterism and reads su.rat
dubb al-asghar (constellation of the Lesser Bear).
  In the Bedouin tradition a bier or corpse-
bearing plank with three accompanying mourn-
ing daugthers was also seen here, parallel with
the image seen in the Greater Bear. This tradi-
tional image was called bandt na"sh al-sughrd (the
  
NUMBER 46

133


FIGURE 49.—The Lesser Bear [Ursa Minor]. Arrows indicate plugs. (Photo: Smithsonian no.
72-8586)

134

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



smaller form of the daughters of the bier), the
bier being the four stars in the square [fiy^v Ursa
Minoris] and the daughters the three in the tail
[eba Ursa Minoris]. This name does not appear
on the Smithsonian globe, but does occur on
globes No. 4 and No. 5 (see Kunitzsch [1961],
no. 96).
  Two calves were also seen by the Arabs as
belonging to the bier, which al-SQfT identified as
the two bright stars of the square, those on the
shoulder and front leg [nos. 7 and 6; yfi UMi],
respectively. In Figure 49 the star on the shoul-
der [7 UMi] is labeled over the top of the back
akhfd al-farqadayn (the more obscure of the two
calves). The inscription across the middle of the
bear applies to the one on the front leg [fi UMi,
Kochab] and reads anwar al-farqadayn (the
brighter of the two calves). The particular des-
ignations of the two stars appear to be unique to
the globes made by the Lahore family, though
the terms appear in al-SQfi's text and catalog as
well as other writings, such as those of Ulugh
Beg.
  The star at the end of the tail, the Pole Star
[a UMi, Polaris], is labeled al-jadi (the goat) in
Figure 49. This name is also of ancient Arab
origin. The large hole drilled next to it is the
hole of the celestial (equatorial) North Pole.
  The constellation has in addition to seven
formed stars, one unformed one beneath the
stomach of the Lesser Bear. This star is labeled
fa"s al-rahd "the axis of the millstone" [Flam. 5,
UMi] in Figure 49. Al-SQfT {Suwar, 28) said it
resembled the axis of a millstone that had in its
center the North Pole; some people called the
star simply "the axis of pole" (al-BTrQnT Astrol.,
sec. 160; Kunitzsch [1961], no, 97c), The word
fa"s more precisely means the protuberance on
the edge of a millstone where the axis or rod was
attached to turn it. The Pole was seen as being
the center of the millstone turning in a socket,
while at the fa"s a beam or rod was attached to
the millstone that had at the other end of the
rod the two ca\\es, farqaddn [fiy UMi], who turn
the grist mill.
  It was also said (Ideler, 15-16; Kunitzsch
[1961], no, 970) that around the North Pole

were obscure stars that together with the
"smaller form of the daughters of the bier"
formed the shape of a fish in whose middle is the
North Pole. Ulugh Beg states that the small stars
around the Pole form ihlilaj (the fruit of the
myrobalan), the latter being the name of a tree
native to Asia bearing edible red or yellow fruit.
These images are not known to be represented
on any celestial globes.
Constellation 2.  The Greater Bear
[Ursa Major]
FIGURE 50
  This constellation is the best known of all
constellations, for a brilliant part of it forms what
is known in our time as the Big Dipper, This
latter asterism, however, forms only a part of the
area covered by the Greater Bear, Because of the
prominence of this area of the heavens, several
different images have been superimposed upon
the region.
  Homer {Od., v,273) speaks of the Bear {apKTos)
"which men also call the Wagon {aixa^a) which
ever circles where it is," Aratus (lines 26-44) uses
both names, giving it the additional name Helice
{eXlKTj), which implies convolutions and may refer
to its being a circumpolar asterism, Hyginus
(11,1) presented the legend that it was a compan-
ion of Artemis named Callisto who became preg-
nant by Zeus and was changed into a bear by
Artemis.
  The oldest and most persistent image in the
Arabic world is that of a bier or corpse-bearing
plank, formed by the four bright stars in the
bowl of our Dipper, followed by three mourning
daughters (the three stars of the tail or Dipper
handle). Thus in Figure 50 there is written on
the rump of the back al-na"sh (the bier), referring
to the four stars forming the square [afiyb Ursae
Majoris]. Under the tail and parallel to it is the
word al-bandt (the daughters), referring to the
three bright stars of the tail [r}^e UMa].
  Since the Greek image of the Bear is super-
imposed over the Bedouin bier, the constellation
is titled surat dubb al-akbar (the form of the
Larger Bear) written along the head. The long
  
NUMBER 46

135



FIGURE 50.—The Greater Bear [Ursa Major]. (Photo: Smithsonian no. 72-3098)

136

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



tail on the bear on the Smithsonian globe, as well
as its mane, gives it an unbearlike appearance.
There are, however, other globes (e,g,. No, 1)
where the bear is shown with a very short, stumpy
tail. The individual names of the four stars form-
ing the square in Figure 50 reflect the Ptolemaic
image of the bear. Thus the one in the middle
of the back is labeled zahr al-dubb al-akbar (the
back of the Greater Bear [no. 16; a UMa,
Dubhe]); at the base of the tail the star is named
maghriz al-dubb (base of the bear [no. 18; 5 UMa,
Megrez]); the one immediately below is called
fakhdh al-dubb (thigh of the bear [no, 19; 7 UMa,
Phecda]); and the star in the stomach is called
mirdt al-dubb (loin of the bear [no, 17; fi UMa,
Merak]).
  The names of the stars in the tail arise from
the Bedouin tradition. The first star of the tail,
no. 25, is called al-jawn, the black horse (c UMa,
Alioth; see Kunitzsch [1961], no. 106). The large
star in the middle of the tail is called al-"andq
(the goat [no. 26; ^UMa, Mizar]) and the last star
of the tail [no, 27; r] UMa, Alkaid] is al-qd"id (the
leader), possibly being viewed as the leader of
the funeral train. The small star next to the
middle of the tail [g UMa, Alcor] was not listed
by Ptolemy. Consequently al-SQfT {Suwar, 32)
named it al-suhd (the overlooked one), a name it
bears in Figure 50. Al-SQfT added that this is a
star by which men test their vision, but did not
assign it a place in his catalog.
  Another Bedouin image is reflected in the
designation of the twin stars in each of the three
prominently depicted feet of the Bear, These
stars are labeled, beginning with the rear paw,
al-qafzah al-uld (the first leap) al-qafzah al-thdni-
yah (the second leap) and al-qafzah al-thdlithah
(the third leap) [nos. 23, 24, 20, 21, 12, 13;
v^XfiLK UMa, respectively], and refer to the image
of a deer or gazelle running before a lion. Each
pair of stars was so named because they resem-
bled the track of a deer's cloven hoof. In Figure
50 the lower front paw bears the label tatimmat
qafzah al-thdlithah (complement of the third
leap), which apparently refers also to LK UMa.
This name occurs on only two other globes, Nos.
11 and 13, and is unknown in astonomical liter-

ature. The maker wrote the name "the third
leap" in front of the front legs, perhaps intending
for it to refer to the two stars above those in the
lower front paw. The maker of the globe may
not have understood the image of the deer, for
the term for "leap" is actually incised as al-faqrah
"vetebra" (a difference of two diacritical points
over the word). An identical spelling occurs on
globes No. 11 and No. 13.
  The constellation is composed of 27 formed
stars, to which this maker adds the additional
"overlooked star." There are eight unformed
stars, five grouped beneath the bear, with one
isolated in front of the front paw and two beneath
the tail. The image of the large lion chasing the
deer is reflected in the name of one of the two
unformed stars beneath the tail, which is labeled
kabid al-asad (liver of the lion [no. 1 unformed;
a Canum Venaticorum; see Kunitzsch [1961] 135).
Constellation 3.    The Dragon
[Draco]
FIGURE 51
  Draco was not one of the star groups men-
tioned by Homer, Aratus (lines 45-62) says the
dragon winds between the two bears like a
branch of a river. Its head is under the foot of
the Kneeler (Hercules) and it coils about the pole
of the ecliptic, to which the great circles dividing
the ecliptic into 12 houses can be seen converg-
ing. Several dragons of Greek legend were asso-
ciated with the figure: (1) the dragon that
guarded the Garden of the Hesperides, thus as-
sociating it with Hercules, whose form is to be
seen above the dragon's head; (2) the dragon
killed by Cadmus; and (3) the Python killed by
Apollo.
  With the introduction of Ptolemaic astronomy
into the Islamic world, the concept of the dragon
was quite literally transferred into this region,
overlaying the earlier Bedouin tradition. The
constellation is called surat al-tinnin (the constel-
lation of the dragon) written in Figure 51 above
its neck (see also Figure 49 for details). The large
star behind the eye is labeled ra"s al-tinnin (head
of the dragon [no, 5; 7 Draconis]).
  
NUMBER 46

137






FIGURE 51.—The Dragon [Draco], This view shows the entire area around the northern ecliptic
pole. (Photo: Smithsonian no. 72-3082)

  The remaining stars named in Draco reflect
the early Bedouin tradition of camels, wolves,
and hyenas. Thus, the four stars forming a square
on the head (nos. 5, 4, 3, 2) are labeled "awd"idh
(the camel-mothers [y^fiv Dra; Kunitzsch (1961),
no. 42]). The star on the tongue [no. 1; p. Dra]

is labeled al-rdfid (the wandering [camel]). This
is probably a mistake, a matter of the distribution
of the diacritical points, for al-rdqis (the ambling
[camel]) as it reads in the text of al-SufT {Suwar,
41; see Kun., 240). However al-QazwTnT (Ideler,
32) and globe No. 4 also read rdfid. Written

138

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



along the second turn of the dragon, along the
back of the spine under the chin, is the word al-
dhib (the wolf). This probably refers to the large
star south of the word [no. 25; 17 Dra]. Al-SufT
and others (Kunitzsch [1961], no. 79) also spoke
of two wolves, the other being the third star to
the north of the latter [no. 24; ^ Dra]. Along the
spine of the back at the tip of the tongue there
reads al-dhikh (the manlike hyena), which refers
to the second star to the south of the wolf [no.
27; t Dra; see Kunitzsch [1961], no, 80],
  According to Ptolemy the dragon consisted of
31 internal stars. The Smithsonian globe also
shows 31 stars in this constellation. Although
some are actually external to the form of the
dragon, this was probably an error on the part
of the artisan. On the neck of Draco before the
first coil are two plugs, which appear in Figures
49 and 51 to be stars, but are not. They are
indicated by arrows, Al-SQfT recorded another
star, without a number, which had not been
observed by Ptolemy, This star was in the midst
of the four "camel-mothers" near the eye and
was called al-ruba" (the young camel). Only one
globe, No. 3, is known to have this particular
label.
Constellation 4.    Cepheus
                    FIGURE 52
In the Greek world Cepheus was the name of
four or five mythological figures, of which per-
haps the best known was the king of Ethiopia and
the husband of Cassiopeia, whom he faces as a
constellation. Aratus (lines 179-188) says he is
located in the area near Cynosura, the Lesser
Bear, which is below his feet. Hipparchus {Comm.
arat, 1, 2,11), however, criticizes Aratus for hav-
ing the size of the constellation too large. In
classical descriptions of the constellations, Ce-
pheus was the only bearded person, indicating
that he was a foreign king. On the Smithsonian
globe, however, there are three other bearded
forms: Perseus, Aquarius, and the Centaur, as
well as the ogre's head carried by Perseus. Both
al-BTrQnT {Astrol., sec. 160) and al-SQfT {Suwar,
49) refer to Cepheus wearing a conical hat {qal-
ansuwah).
  
In Figure 52 the title is written near his exotic
headress, sdrat qiqd"us (the constellation of Ce-
pheus), the latter being a transliteration of the
Greek Kr]ipevs (Cepheus).^^ All star names reflect
an older indigenous Arabic tradition of a shep-
herd, his dog, and his flock, rather than the
Ptolemaic tradition of a human figure. In Figure
52 the star on his forward foot, the second star
of the constellation [7 Cephei] is called al-rd"i{the
shepherd). The star on the northern tie of his
sash (no. 3) is fi Cephei, whose modern name
Alfirk is derived from a name applied by the
Bedouins not just to this star but also the one
above on his upper arm, the fourth of the con-
stellation [a Cephei]. Kawkabd al-firq, as al-SufT
{Suwar, 46) called these two, means "the two
stars of the flock" referring to the flock tended
by the shepherd [afi Cephei]. The word al-firq
(the flock) is engraved on the Smithsonian globe
under the arm of Cepheus near Draco.
  In keeping with the image of a shepherd and
his flock al-SQfT {Suwar, 47) recorded a star be-
tween the two legs, which had not been observed
by Ptolemy and to which he consequently did
not assign a number. This star he called kalb al-
rd"i (the shepherd's dog [Flam. 28, 29 Cephei]).
It does not appear on the Smithsonian globe,
since only one of the unnumbered stars of al-
SQfT is indicated (the "overlooked star" in Ursa
Major). On some globes (e.g., Nos. 4 and 5) Kalb
al-rd"i is written between the legs but with no
star inlaid there, so that it was interpreted as
applying to the star on the back foot [K Cephei].
  In addition to "the shepherd's dog" al-SufT
recorded four external stars near the elbow of
the back arm and four stars on the forward leg,
which had also not been observed by Ptolemy,
These nine stars, however, he did not number.
Thus his catalog has 11 formed stars and 2
unformed stars, on either side of the head, as did
Ptolemy's, Figure 52 also has 11 formed and 2
unformed stars. Two medium sized plugs can be
seen on and near the southern leg, as well as a
row of small plugs between the feet (indicated by
arrows). What appears to be a star on the front
(southern) tie of his sash is actually a plug and
not a star. The star on Cepheus's back (northern)
  
NUMBER 46

139


FIGURE 52.—Cepheus. (Photo: Smithsonian no. 72-3099)

140

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



foot was put in but has now fallen out. The other
circles of the same size are in reality plugs.
Constellation 5.    The Howler,
that is, the Ox-Driver [Bootes]
FIGURE 53
   The name "ox-driver" given by the maker of
the Smithsonian globe as the second name of the
constellation {al-baqqdr) is a translation of the
Greek name fiocorrjs (ploughman), which appears
in Greek literature as early as Homer {Od., V,
272) where it apparently applies not to the entire
constellation, but to its brightest stars. In the
Odyssey the name of ploughman given this aster-
ism is in keeping with the name of wagon, given
to the Greater Bear, Thus Bootes is the plough-
man or driver of the wagon represented by part
of the Greater Bear, Aratus (lines 91-95,
[trns.], Mair, 215) calls the constellation
apKTOipvXa^ (guardian of the bear) "whom people
also call Bootes since he seems to lay hand on the
wain-like Bear.' Beneath his belt is the bright
star apKTovpos (guard of the bear). From the latter
name we derive the star's common name today
Arcturus, the fourth brightest star of the heav-
ens. The Smithsonian globe is unique among
studied globes in that it indicates a knowledge of
the correct Greek name of the constellation.
  The first title, surat al-"awwd" (the form of the
Howler), is the usual Arab designation, and is
derived from a misunderstanding of the word
fioojT-qs (the ploughman) as reading fiorjTris (the
howler or shouter). It may be that the figure was
viewed as a herdsman driving the Great Bear
before him. Indeed in nearly all instances he is
depicted as following Ursa Major, except on
globe No. 4, where he is facing away with his
head turned back toward the bear. The second
part of the title "that is to say, the ox-driver {al-
baqqdr)" reflects a closer rendering of the Greek
fioCiT-qs (see Kunitzsch [1974], 174-176, for other
names such as hdris al-shamdl, "the guardian of
the north").
  The constellation consists of 22 internal stars,
one of which it shares with Hercules (the one at
the tip of the sword). Immediately above the hilt

of the sword there are actually two small inlaid
stars next to one another; on the globe they
appear almost as one). There is one unformed
star below the hemline of the tunic. In Figure 53
can be seen a damaged area of the globe. A small
area is missing and a crack occurs above it; im-
mediately to the left can be detected one of the
two large plugs in the globe. There is also faintly
visible on Bootes's western chest and thigh to-
ward the south a smaller, nearly rectangular
plug. Only two stars of this constellation are
named on the Smithsonian globe. Both star
names seem to reflect the Bedouin image of a
lancer, or at least an armed form, rather than a
"howler' or an ox-driver derived from Greek
tradition. On the forward leg the largest of the
three stars is named mufrad al-rdmih (the solitary
[star] of the lancer). Mufrad (the solitary) may
have been seen originally as outside the lines of
the armed figure. This star (no, 20; r; Bootis,
Mufrid) appears labeled only on globes made by
the Lahore family, and the globe made by "AIT
KashmTrT (No. 10). The star is given this title in
the illustrations to Ulugh Beg's copy of al-SQfT
(see al-SQfi Suwar, figure 5) but al-SQfT himself
calls it rumh (the lance), as do most other writers
(Kunitzsch [1961], no, 255).
  The second labeled star in Bootes is the very
large one beneath the hem of the tunic, the first
unformed star of the constellation. This is a
Bootis or Arcturus. The precise meaning of the
Arabic name for this star, simdk al-rdmih, is dif-
ficult to determine. The word simdk occurs for a
star in Virgo, which gave its name to the four-
teenth lunar mansion. The word simdk is perhaps
of Babylonian origin, but its meaning is so ob-
scure that it cannot be translated. The word al-
rdmih usually acts as an adjective meaning
"armed" or "fighting," and so we have a lance-
bearing or fighting simdk. On the Smithsonian
globe the word al-rdmih does not actually func-
tion as an adjective, but this is probably another
grammatical error of the maker (see Kunitzsch
[1961], no. 270; on globes Nos. 1, 3, and 34 the
word simdk does not appear, only al-rdmih). The
star Arcturus was also by some (Kunitzsch
[1961], no. 77) given the name al-dhakar "the
  
NUMBER 46

141


FIGURE 53.—The Howler, that is, the Ox-Driver [Bootes] and the Crown [Corona Borealis].
(Photo: Smithsonian no. 72-3090)

142

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



penis." This may not be a true Arab name, how-
ever, but only a designation arising after the
introduction of the Ptolemaic constellation. For
the star stands in Bootes baina fakhidhayhi (be-
tween his two thighs), as al-SQfT says {Suwar, 52)
from which the term al-dhakar could easily have
arisen. It seems that in the Bedouin tradition the
two simdks (in Virgo and Bootes) were viewed as
the two hind legs of the large lion.
Constellation 6.    The Crown
[Corona Borealis]
FIGURE 54
  To the west of Draco, between the two figures
of Hercules and Bootes near the damaged section
of the Smithsonian globe, are two concentric
circles representing the constellation of the
Northern Crown (see Figure 54 for detailed
view). Aratus (lines 70-71, [trns.], Mair, 213)
called it areipavos, a wreath or garland "which
glorious Dionysus set to be a memorial of the
dead Ariadne." On the two earliest globes (Nos.
1 and 34) the Greek title is literally translated as
al-iklil al-shamdli (the northern wreath), and so
it is found in al-SQfT where it is drawn as two
concentric circles. Al-BTrQnT {Astrol., sec. 106)
says it is generally known as "the dish of the
orphans and poor people" {qas"at al-yatdmd wa
al-masdkin).
  The etymology of the more common Arabic
name al-fakkah, engraved on the Smithsonian
globe above the outer ring, is puzzling. The basic
root of the word means "to break" or "to release."
There is a break or gap between the two north-
ernmost stars [-Kl Coronae Borealis] of the ring of
eight stars making up the constellation. The
word al-fakkah may refer to this break (Kunitzsch
[1961], no. 85). At this gap the knot of the
wreath is tied, when it is pictured as a garland.
The gap could also represent a break in a dish
or plate, which might account for al-BTrum's
term "the dish of orphans and poor people."
  On globes No. 26 and No. 1 it is drawn as an
open horseshoe but on all others it is a closed
plate or ring, except the early eighteenth-century
globe. No. 31, where it is actually a crown. The

largest and brightest star of the constellation [a
CrB, Alphecca], the fourth star counting from
the southern side of the gap, is labeled, within
the inner ring, nayir fakkah. This is poor gram-
mar for al-nayir min al-fakkah (the brilliant star
of al-fakkah). Al-SQfT {Suwar, 57) also calls it al-
munir min al-fakkah.
Constellation 7.    The Kneeling Man
[Hercules]
FIGURE 54
  The position of Hercules is up-side-down,
kneeling on his rearmost (western) knee, with his
foot stepping on the head of Draco; he is head
to head with the serpent bearer, Serpentarius. In
the Arabic tradition he is usually called simply
al-jdthi (the kneeling man) from the form of the
Greek constellation, although al-SQfT adds that
he was also called al-rdqis (the dancer; Ideler,
381). On the Smithsonian globe the figure is
drawn so that the raised hand appears to be a
right rather than a left hand, and both his feet
are shown with the large toe on the same side of
the foot. In Figure 54 a large plug in the globe
can be seen to extend across the back (western)
leg of Hercules. A fairly large rectilinear plug is
over his eastern shoulder and a small rectangular
plug can be seen on his eastern thigh.
  When Aratus wrote his poem in the third
century BC the attribution of Hercules to the
constellation had not been made. Aratus (lines
64-66; [trns.] Mair, 213) says "that sign no man
knows how to read clearly, nor on what task he
is bent, but men simply call him On His Knees."
The Greek reads eyybvaais, i.e., 6 ev ybvaai
Kadrjuevas. Aratus (lines 270; 64; 73) also several
times calls him "the unknown phantom" or sim-
ply "phantom" {etbiiiXov). Gradually, however, the
mythological figure of Hercules, or Heracles
{ripaKXijs) as he was called in the Greek world,
came to be attached to the form of a kneeling,
toiling man. Hyginus and later mythographers
identified him with Theseus, Ixion, Prometheus,
Orpheus, and others. It is interesting, however,
that Ptolemy calls the constellation only "he who
is on his knees" and never Hercules. The kneel-
  
NUMBER 46

143


FIGURE 54.—The Kneeling Man [Hercules] and the Crown [Corona Borealis]. Arrows indicate
plugs. (Photo: Smithsonian no. 72-3094)

144

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



ing figure has his forward foot on the head of
Draco, so that he could be interpreted as Her-
cules fighting the dragon that guarded the Gar-
den of the Hesperides, the twelfth labor assigned
to Hercules. The attribution of Hercules to the
constellation works in nicely with the interpre-
tation of some of the other nearby constellations.
For instance Hydra, the sea serpent, could be
interpreted as the Lernean Hydra of his second
labor. A centaur also plays a role in the story of
Hercules, and Sagitta the arrow could be thrown
by Hercules toward Aquila (the eagle), Lyra, and
Cygnus, distorted representations of the Stym-
phalian birds of his fifth labor.
  The figure was frequently depicted with a
lion's skin over his forward arm and a shepherd's
crook in the back raised hand. In the Islamic
world the lion's skin is not shown, and in place
of the shepherd's crook is the oriental scimitar.
Wellesz ([1959], 5), suggested that this scimitar
was an early misunderstanding of the shepherd's
crook on the part of the illustrator of al-SQfi's
treatise. On the earliest globe (No. 1) there is no
weapon in the hand of the Kneeler, but on all
other globes there is a small scimitar-like weapon
in the raised back hand.
   The Arabic title for the constellation, siirat al-
jdthi (the constellation of the kneeling man), lit-
erally translates the older Greek title. It is written
along the figure's western side between Corona
Borealis and Hercules. The constellation is com-
posed of 29 formed stars and one unformed star
above the raised upper shoulder. Al-SQfT num-
bers them 1 unformed, 28 formed, and 1 unnum-
bered formed star on the back foot, which it
shares with Bootes.
  The only star labeled on this globe is the one
in the head, which is called ra"s al-jdthi {the head
of the kneeling man [a Herculis]). While this
designation for the first star of the constellation
is given by al-SQfT, he also gives the Bedouin
term kalb al-rd"i (the shepherd's dog). (See the
discussion of Serpentarius for the shepherd and
another one of his dogs.) The term "the shep-
herd's dog" is found on globe No. 4 for this star.
On the earliest globes (Nos. 1 and 34) as well as
No. 5, a Herculis is not labeled. On globe No. 3

the star is called simply ra"s (head). But on the
globes of the Lahore family as well as the Smith-
sonian globe, and the later globe No. 31, it is
labeled "the head of the kneeling man," reflect-
ing the Greek rather than the Bedouin asterism.
The Smithsonian globe does not label a promi-
nent row of stars across his two arms [nos. 4,3,
2, 5, 6, 7, 8, 9 and 10 of Hercules; KyfibXfiov^
Herculis], two stars of the constellation Lyra at
the end of his forward hand [nos. 7 and 9; fiy
Lyrae], and two from the Serpent near his back
raised elbow [nos. 4 and 3; fiy Serpentis]. These
stars are called by al-Sufi al-nasaq al-sha"dmi {the
northern row), a name which they bear on all
globes except the two earliest (Nos. 1 and 34)
and the Mughal ones. This row of stars was
thought to be the northern boundary of a
meadow, the southern boundary being formed
by a row of stars in Serpentarius (Kunitzsch
[1961], nos. 243, 192a).
Constellation 8.    The Lyre
[Lyra]
FIGURE 55
  The Arabic title of this constellation, surat al-
shulydq, may be derived from the Greek word
X^Xvs, meaning tortoise. Since the lyre was made
by stretching seven strings on a tortoise shell
which then acted as a sounding board, x^Xw came
to mean lyre. Aratus (lines 268-274) speaks of
the tortoise, which Hermes pierced for strings
and named the lyre {Xvpa), and was placed by
Aratus as a constellation between the Bird (Cyg-
nus) and the forward knee of Kneeling Man. Al-
SQfi {Suwar, 67) says it is the constellation al-
lurd, a transliteration of the Greek lyrfl and adds
that it is also called al-shulydq, al-sanj (a Persian
word for cymbals or brass castanets), al-mi"zaf{a
triangular harp with seven strings equivalent to
the Greek Kidapa), al-iwdzz (a goose), and al-
sulahfdt (a tortoise). The latter word appears on
the two oldest globes, Nos, 1 and 34. Indeed on
No. 1, the form engraved is that of a tortoise.
The exact origin of the word shulydq as used on
the Smithsonian globe seems uncertain (see al-
SQfi [trns,], 75 note 2); it is possible it arose as a
  
NCMBER 46

145



FIGURE 55.—The Lyre [Lyra] and the Fowl [Cygnus]. (Smithsonian no. 72-3092)

146

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



misunderstanding and incorrect writing of sulah-
fdt. In any case it appears that the meaning of
the word as a musical instrument was soon lost,
for on globe No. 3 it is merely a decorative
device, while Nos. 4 and 5 show the constellation
more as a cup or jug. Some illustrations of al-
SQfi manuscripts depict it as an urn (Wellesz
[1959], figure 22b; Upton, 143), while others
show it as a tortoise (al-SQfT Suwar, figure 8). On
the Smithsonian globe and No. 26 (also by the
Lahore family) it is drawn to resemble a crown.
The constellation is composed of 10 formed
stars. Al-SQfT {Suwar, 68) notes an extra star
which he had seen on globes. In light of this it is
interesting that the eleventh century globe. No,
34, has 11 stars on it.
  Only one star is labeled on the Smithsonian
globe, Vega [a Lyrae], the fifth brightest star of
the heavens. It is the lower (westernmost) star,
at the vertex of the triangle of stars on the
northern side, indicated by a slightly larger silver
dot. It is labeled nasr wdqi" (a falling eagle), the
Vega being a corrupt transliteration of wdqi"
(falling). Al-SQfT {Suwar, 68) and al-BTrQnT {As-
trol., sec. 163) apply the name only to a Lyrae,
while the three stars of the triangle together (nos.
1-3) are called al-athdfi (the legs of a tripod).
Others (Kunitzsch [1961], no, 195a) interpreted
"the falling eagle" as referring to the three stars
ae''^f''^Lyr, suggesting the form of a bird with
half-closed wings and hence falling and in peril
(see Figure 23, globe No. 46, which depicts it as
a falling bird).
Constellation 9.    The Fowl
[Cygnus]
                        FIGURE 55
In Greek literature this constellation was con-
ceived of as a flying bird, usually a swan. Aratus
(lines 275-281) used the word opvis, a general
word for bird, both a wild bird of prey and a
domestic fowl, saying it is a bird riepbeis (wreathed
in mist; i.e., in the Milky Way), whose western
wing tip stretches out toward Cepheus, whose
hand we see near Draco, while its other wing is

near the prancing horse (Pegasus). Later writers
(e.g., Erat. Catast., 25) refer to it as the KVKVOS
(the swan), and traditions associate it with the
story of Leda and the swan. In the Arabic world
the word al-dajdjah is used, which can mean any
type of fowl, but in particular domestic fowls,
hens and cocks. Al-SQfT also uses the even more
general word for bird, td"ir, but this does not
occur on globes. Al-BTrQnT {Astrol., sec. 160) adds
that al-dajdjah is like a duck {battah) with its neck
and wings stretched out as if it were flying. On
all globes studied prior to those of Mughal India,
the bird is clearly a type of barnyard fowl with a
wattle or comb and with wings partially closed so
that it appears not to be flying. On the Smithson-
ian globe, however, the bird, whose title is writ-
ten below its head, has its wings fully out-
stretched as in flight, the neck and head are
straight, and there is no sign of a wattle or comb,
giving it the general appearance of a goose or
swan.
  The constellation consists of 17 formed stars
and 2 unformed ones. The very large star at the
base of the tail is labeled dhanab al-dajdjah wa
huwa al-ridf (the tail of the fowl, that is to say,
the follower). This star is no. 5, a Cygni, named
today Deneb, obviously derived from dhanab,
meaning tail. The word al-ridf (the follower)
arose from the Bedouin concept of four stars
across the upper edge of the wings [nos. 6, 4, 10,
12; bye^ Cyg] as Horsemen (labeled al-fawdris on
globes Nos. 4 and 5; see Kunitzsch [1961], no.
98). The star a Cyg was then "the follower" or
"the one riding behind" the horsemen. The label
on the Smithsonian globe gives both the Ptole-
maic and Bedouin star names, while on some of
the other later globes, such as Nos. 26 and 31,
only the Ptolemaic "tail of the fowl" is given.
  In Figure 55 the star in the bill of the bird is
called minqdr al-dajdjah (the bird's beak [no, 1;
fi Cyg\} and the star in the bend of the northern
leg is labeled rukbat al-dajdjah (the bird's knee
[no, 17; co''^ Cyg]. These basically Ptolemaic des-
ignations do not seem to occur on Class A globes
prior to the Lahore family, but were considered
by al-SQfT to be "astrolabe" stars, and hence are
  
NUMBER 46

147



found on some Class B globes. The star in the
bill and the one in the tail plus the Horsemen
form what today is known as the asterism of the
Northern Cross [afiybe^ Cyg].
Constellation 10.
The Lady with a Chair
[Cassiopeia]
FIGURE 56
  Aratus (lines 188-196) identifies this group
with Kaaaleireia, the wife of Cepheus, who stands
behind her toward the north. He also says the
stars (the well-known W-shape) resemble "the key
of a twofold door barred within, wherewith men
striking shoot back the bolts" ([trns.] Mair, 223).
A Greek tradition that developed by the time of
Ptolemy drew her as a queen sitting on a throne
with her arm outstretched toward her daughter
Andromeda, whose foot can be seen above her
open hand. The legend goes that Cassiopeia
boasted she was fairer than the sea nymphs, the
Nereides, for which brashness Poseidon sent a
flood and a sea monster (Cetus, another constel-
lation to be discussed in Figure 75) to ravage the
coast of her realm in Ethiopia (Hyginus, II, 10).
The only way in which Cepheus and Cassiopeia
could rid their realm of the sea monster was to
sacrifice their daughter Andromeda to the mon-
ster. Their subjects thus compelled them to chain
her to a rock to be devoured. She was later
rescued by Perseus, who can be seen as a con-
stellation south of Cassiopeia's legs (see Figure
57), Cassiopeia, however, was banished by Posei-
don to the sky, but so near the Pole Star that she
hangs downward half the year as a lesson in
humility.
  In the Arabic world her name was not transli-
terated into Arabic, but was lost completely. She
is simply called the lady with the chair {dhdt al-
kursi), as can be seen in the title written above
her head in Figure 56, The constellation consists
of 13 formed stars. Near her forward leg is
written rukbat dhdt al-kursi (the knee of the
seated lady). This does not occur on any other
globe studied, but it is to be found in Ulugh

Beg's copy of al-SQfT's treatise {Suwar, figure 10;
see also Wellesz [1959], figure 68), where it
clearly refers to the large star in her lap [b Cas-
siopeiae], the fifth of the constellation. It is no
doubt intended to apply to that star on the Smith-
sonian globe as well, and not to one of the two
in the leg that were considered of fourth mag-
nitude, and thus seldom given individual names.
The large star on her raised elbow [fi Cas] is
labeled kaff al-khadib (the dyed hand), that is, a
hand dyed with henna {hinnd"), a red dye made
from Lawsonia inermis. The star name is more
frequently and correctly written as al-kaff al-
khadib (cf Wellesz [1959], figures 68, 62, 47;
Kunitzsch [1961], no. 136, 136b). On our globe,
as on all known globes, it appears to refer only
to this one large star on her elbow, no. 12 of the
constellation. However, as al-SQfT notes {Suwar,
77), names referring to a configuration of stars
were sometimes used by astronomers to refer to
only one of the stars of the group, as in this case
when this became an important star and was put
on astrolabes. Accordingly, in the Bedouin tra-
dition al-kaff al-khadib (the dyed hand) refers to
five white stars in the Milky Way {majarrah),
which are identified by al-SQfi as referring to the
W-shaped asterism, i,e,, nos, 12, 2, 4, 5, and 6
of the constellation [fiaybe Cas]—one in her el-
bow, one in her cheek, at her waist, on her lap
and the upper of the two on her northern leg.
These five stars were viewed by the Bedouins as
being the open hand of the upper extended arm
(passing through Perseus) of a figure named al-
thurayyd, whose head was the Pleiades, (See the
discussion of Lunar Mansion 3, the Pleiades,) Al-
SQfT {Suwar, 77) also designated the star on her
elbow as kaff al-khadib wa huwa "aid sandm al-
ndqah (the dyed hand, and it is at the hump of
the she-camel). This reflects another Bedouin
tradition of envisaging a large she-camel {ndqah)
in the stars composing the region of Cassiopeia
and Andromeda (see Kunitzsch [1961], no, 190).
While terms related to this camel are not found
on celestial globes, the early notion of a camel in
this region of the heavens is illustrated in two
copies of al-Sufi's treatise (Wellesz [1964], 91).

148

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 56.—The Lady with a Chair [Cassiopeia]. (Photo: Smithsonian no. 72-3100)

NUMBER 46

149



Constellation 11.    Barsd"iis,
that is, the Bearer of the
Ogre's Head [Perseus]
FIGURE 57
  In the title sHrat barsd"ils wa huma hdmil ra"s
al-ghUl, which is written in two lines along the
northern side of the figure, the name Perseus is
transliterated as Barsd"ds, after which it is added
by way of explanation that he is the bearer of the
ogre's head (see Kunitzsch [1974], 181 for var-
ious ways in which the Greek name Perseus was
transliterated into Arabic). The figure holds by
the hair the bearded and mustachioed head of a
male demon, while in the hand raised over his
head he wields a sword, which appears at first
glance to resemble more a club until the hilt is
noticed. Perseus is one of four bearded figures
on the Smithsonian globe, not including the
bearded orge's head. The artist has again drawn
the figure with his hands reversed.
  According to Greek tradition Perseus {-wepaevs)
was the son by Zeus of Danae who, because her
father had been told by an oracle that his grand-
son would kill him, had been set adrift on the sea
with her son. They were taken in by King Poly-
dectes, who years later contrived to send Perseus
to fetch the head of the Gorgon Medusa. The
Gorgons were three hideous sisters, of whom
only Medusa was mortal; snakes grew on their
heads instead of hair, and anyone who beheld
them was turned to stone. With the aid of
Athena, Perseus beheaded Medusa and carried
off her head, holding it by the snakes. When
the blood dripping from the head fell to earth
it gave rise to Pegasus, another constellation.
Before returning to King Polydectes Perseus
rescued Andromeda from the sea monster Ce-
tus and married her. In Islamicate iconography
the snakes on Medusa's head became ordinary
hair and the female Medusa became a male
ghoul. It has been suggested by Wellesz
([1959], 9) that the blood flowing from Me-
dusa's neck in Greek representation may have
suggested a man's beard to the Arabic artists.
To the west of Perseus are to be seen the feet

of his wife Andromeda, and directly above him
his mother-in-law Cassiopeia. When Aratus (lines
248-253) described the constellation he men-
tioned him by name but did not describe him as
carrying Medusa's head. However, later mythog-
raphers described him as having a sword in his
right hand. Medusa's head in his left, and wings
on his feet, although this is not reflected in
Ptolemy's description of the figure in his star
catalog. The winged boots decorated with a scale
pattern he wears in Figure 57 are actually a part
of typical Mughal dress.
  The constellation consists of 26 formed stars
and 3 unformed stars. To the left of the ghoul's
head is written ra"s al-ghdl (head of the ghoul),
which is intended to refer to the largest star in
the head, [fi Persi, Algol], a star commonly la-
beled on globes.
  Within the blade of the sword there is written
mu"tasam al-thurayyd (the shelter of al-thurayyd),
which is probably a mistake for mi"sam al-thu-
rayyd (the wrist of al-thurayyd). The former read-
ing is not known to occur elsewhere, except on
globes No, 11 and No, 13, which were probably
made by the same maker who produced the
unsigned Smithsonian globe. The latter star
name has a long tradition in the Islamic world,
being used by anwd" authors (Kunitzsch [1961],
no, 169) and on all globes except the two earliest,
Nos, 1 and 34, where the label is omitted. This
name refers to the star, no. 1 (actually a pair of
open star clusters) in the hand of Perseus [hx
Persei], which was considered a nebulous group
{■n veipeXoeibrjs av^poipi]) by Ptolemy; al-Sufi (5M-
war, 85) calls the star sahdbi (cloudy) (cf. Kun-
itzsch [ 1961 ], no. 169). On globe No. 26 by Diya^
al-DTn of the Lahore workshop the label reads
mi"sam al-thurayyd sahdbi (the nebulous wrist of
al-thurayyd).
  As to the significance of al-thurayyd, we must
look again to an early Bedouin tradition over
whose terminology there is superimposed the
Ptolemaic image of Perseus. This concept of a
large female figure named al-thurayyd deter-
mines the labeling of all the other stars in this
constellation. The Pleiades were seen as the head
  
150

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 57.—Barsa^Hs, that is, the Bearer of the Ogre's Head [Perseus]. (Photo: Smithsonian
no. 72-3097)

NUMBER 46

151



of a woman whose upper arm stretched out
through Perseus with her open hand with finger
tips stained with henna forming the constellation
Cassiopeia. With this configuration in mind, the
star (or pair of open star clusters) is called the
wrist of al-thurayyd. The large star in the stomach
of Perseus is labeled in Figure 57 mirfaq al-
thurayyd (the elbow of al-thurayyd [no. 2; a Per,
Mirfak]). The star on the shin-bone of the back
leg is labeled mankib al-thurayyd (the shoulder of
al-thurayyd [no. 24; ^ Per]). The two stars in the
lower foot [nos. 25, 26; o^Per] are usually titled
"dtiq al-thurayyd (the shoulder-blade of al-thu-
rayyd). The Smithsonian globe reads "dtif al-
thurayyd (which occurs on only two other globes,
Nos. 11 and 13)—a certain mistake, a td" having
been written instead of td", and one dot omitted.
See the discussion of Lunar Mansion 3 for fur-
ther details on stars related to al-thurayyd.
Constellation 12.    He who Holds
the Reins [Auriga]
FIGURE 58
  In the Greek world there were two images
given to this area between the Pleiades and the
Greater Bear, both described by Aratus (lines
156-166). There was seen in the area the chari-
oteer 0 rjvloxos (one who holds the reins) and also
the Goat, di^, which is said by Aratus to be on
his left (western) shoulder, and young goats or
kids, epiipoL, which were placed on the chari-
oteer's waist below. The mythological signifi-
cance of the charioteer and goats is obscure.
Hyginus (II, 13) said the figure was called
Erichthonis who first invented the four-horse
chariot and first established sacrifices to Athena.
Cleostratos of Tenedos is said to have first
pointed out the kids among the stars. Ptolemy
called the constellation rjvloxos (Heniochus; he
who holds the reins) and the third star the goat
and the eighth and ninth the kids. The goat and
the kids were considered important indicators of
weather. The bucolic poet Theocritus (VII, 52-
54) said, for example, "fair voyage to Mitylene
shall Ageanax have when the Kids stand in the

evening sky and the south wind speeds the wet
waves, and when Orion stays his feet upon the
Ocean" ([trns,] Gow, 59). In Latin literature the
constellation was called Auriga (Hyginus, II, 3;
Virgil Aeneid, IX, 668).
  In the Islamic world the Greek conception of
the charioteer and the goats was quite accurately
transported, unchanged, into the Arabic culture.
Thus the constellation is called siirat mumsik al-
a"innah (the constellation of him who holds the
reins), in Figure 58 written above the head. The
large star on the eastern shoulder is labeled man-
kib dhi al-"indn (the shoulder having the rein [no.
4; fi Aurigae]), from which we get the modern
name Menkalinan. This is not a common label
on extant globes, but does occur elsewhere (e.g.,
Nos. 11 and 13) and in Ulugh Beg's copy of al-
SQfi's treatise {Suwar, fig. 12).
  The large star on the western shoulder is la-
beled "ayyuq [no. 3; a Aur] and is Capella, the
sixth brightest star in the heavens. The meaning
of the Arabic word is not clear (see Kunitzsch
[1961], no. 47). The usual meaningof the Arabic
root is "to delay" or "to impede." It has been
suggested that it is derived from a Babylonian
word iqu (goat), and that perhaps this word is
the source for both the Greek word a'l^ and
the Arabic "ayydq (Hommel 595). The anwd"
author Ibn Qutaybah (Kunitzsch [1961], no.
47) says that "ayyUq is behind al-thurayyd (the
Pleiades) on the right side of the Milky Way,
and "it is a whiter, brighter, and more brilliant
star and is located much closer to the Pole than
the Pleiades." Al-Siifi also called Capella raqib
al-thurayyd (guardian of the Pleiades), since in
many locations it rose at the same time.
  Al-SufT {Suwar, 91) also referred to the two
stars close together on the western arm bent
in toward his waist as al-jidydn "the two Kids"
[nos. 8,9; f?; Aur], apparently following the
Greek tradition that may have been derived
from a Babylonian source. The star near the
elbow of the same arm is also named "the
goat," using a different word "anz, which al-
Siifi stated could also apply to Capella, but
more appropriately was used for the star below
  
152

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 58.—He Who Holds the Reins [Auriga]. Arrow indicates plug. (Photo: Smithsonian no. 72-3096)

NUMBER 46

153



[no. 7; 6 Aur]. The asterism of these three stars
[cT^f Awr] is even today known as "the Kids."
  Thirteen stars are shown on the Smithsonian
globe as being in the constellation Auriga. The
star in the lower foot, which is shared with the
horn of Taurus below, is given the number 11
of the constellation Auriga. What appears to be
a round star above the hand on his waist is in
fact a plug and not an inlaid star; it is indicated
by an arrow in Figure 58. Ptolemy cataloged 14
stars in this constellation; al-Sufi, however, omit-
ted the fourteenth star of Ptolemy's listing, with
the observation that he was not able to observe
it. In this, al-SQfi was followed by Ulugh Beg five
centuries later.
  Al-Sufi {Suwar 92) also mentions that in the
Bedouin tradition there were two stars between
"dtiq al-thurayyd (the shoulder-blade of al-thu-
rayyd), which are the two in the lower foot of
Perseus to the west, and al-"ayyu,q [a Aur, Capella]
on the western shoulder of Auriga. He identifies
these as being the star in the upper foot of
Perseus [no. 21; e Persei] and the one in the
upper foot of Auriga [no. 10; i Aur]. He says
they are called al-mrjf and al-birjis, though which
is which is not clear. The significance of these
two star names is very uncertain, although some
suggest that they indicate camel imagery (Kun-
itzsch [1961], no, 66), Neither star name is
known to occur on celestial globes.
Constellations 13 and 14.    The Serpent
Charmer and the Serpent
[Serpentarius and Serpens]
FIGURE 59
  Aratus (lines 74-90) described the serpent-
holder oipiovxos (Ophiuchus) as having his head
toward the north, near that of "the phantom"
(Hercules), with his feet trampling a monster
(Scorpio). He notes that the stars on his shoulders
gleam brightly while those on this hands are
dimmer. Ophiuchus clutches in his two hands the
oipis, a serpent or snake which is coiled a little
about his eastern hand but is greatly coiled above

his western hand with its jaw turned toward
Corona Borealis, Later mythographers (e,g.,
Erat. Catast, 4) identified Ophiuchus with Ascle-
pius and so with the art of healing.
  In Figure 59 to the east of the head is written
surat al-hawwd (the constellation of the serpent
charmer), which consists of 24 formed stars and
5 unformed ones. The large star in the head, no.
1, is labeled directly over the head ra^s al-hawwd
(the head of the serpent charmer), from which
we get the modern name Ras Alhague [a Ophi-
uchi]. This is not a common term on Class A
globes prior to the Lahore workshop, although
al-SQfi uses it and considered it an astrolabe star.
The usual Arabic designation for this star on
early globes is al-rd"i {the shepherd), a term that
is also engraved on the Smithsonian globe to the
east of the head, near the graduated celestial
equator which can be seen passing across his
chest. The star nearest this word on the Smith-
sonian globe is no. 5 of the unformed stars [Flam.
72, Oph], a star considered of fourth magnitude.
Since stars of this magnitude were seldom given
special names, the name must surely be intended
to apply to the large star in his head [a Oph], in
keeping with all the astronomical literature (Kun-
itzsch [1961], no. 276).
  Al-Sufi {Suwar, 102) adds that there are two
dogs near the shepherd. One of the dogs is the
higher and larger of the two stars on the eastern
shoulder [no. 2; fi Oph] called Kalb al-rd"i (the
shepherd's dog). The other is the star in the head
of Hercules [a Herculis]. This probably reflects
a Bedouin tradition, although no source but al-
SQfi mentions it (see Kunitzsch [1961], no. 143).
The dogs of the shepherd are not labeled on the
Smithsonian globe.
  The serpent in Figure 59 was titled beneath
the head near the missing piece of the globe,
[iwra^] hayyat al-hawwd (the Constellation of the
Serpent Charmer's Serpent). On this globe there
appear to be only 15 stars in the Serpent itself.
However, three (nos. 1, 4, 5) were placed in the
area now missing, bringing the original total to
18 formed stars, corresponding with the number »
in the catalogs.
  
154

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 59.—The Serpent Charmer and the Serpent [Serpentarius and Serpens]. Arrows
indicate plugs (one is missing). (Photo: Smithsonian no. 72-8587)

NUMBER 46

155



  In Figure 59 there is one star labeled on the
serpent, one that is difficult to see for it is located
in the badly tarnished area in the western coil.
The star to which the label refers is no. 9 of the
serpent [a Serpentis], clearly labeled "ayn al-
hayyah (the eye of the serpent). This label is quite
certainly an error, for the eye of the serpent is
nowhere near, and this name occurs in no other
globe or text. The maker must mistakenly have
thought of the word "ayn (eye) when seeing the
word "unq (neck), which differs from the word
for eye by only the placement of diacritical points
and a slightly different formation of the last
letter. In any case, this star is called "unq al-
hayyah (the neck of the serpent) by al-SQfi {Suwar,
107) and on all globes studied except the two
earliest (nos. 1 and 34), where it is not labeled.
The modern name of the star is Unuk, clearly
derived from "unq.
  In reference to the fourth star of the serpent
[fi Ser; a star which should be where the piece is
missing], al-SQfi says "the place where the neck
begins, i.e., the beginning of the Northern Row"
{al-nasaq al-sha"dmi). While not appearing on the
Smithsonian globe, this name refers to the
Bedouin notion of a Northern Row of stars which
includes two in the serpent [fiy Ser], the stars
across the arms of Hercules and two in Lyra,
which form the edge of a meadow. See the dis-
cussion of Figure 54 (Hercules) for further de-
tails.
  Of the seventh star of the serpent (the high,
very visible star in the western coil; b Ser), al-Sufi
remarks that it is "the beginning of the Southern
Row," which is another row of stars, south of the
one through Hercules. It is composed of four
stars in the tarnished area of Serpens [nos.
7,8,9,10; bXae Ser], two stars in the western hand
of Serpentarius [nos. 7,8; be Oph], the star in the
serpent southeast of the western hand of the
serpent-charmer [no. 12; u Oph], and the star
near the hem of the serpent charmer's pleated
skirt and the two uppermost stars of his eastern
leg [nos. 12, 19, 13; 7;f^ Oph]. The area between
the Northern and Southern Rows was called al-
rawdah (the meadow), where the shepherd and

his two dogs roamed (al-SQfi Suwar, 102). In
Figure 59 on each foot of Serpentarius is a name
of a lunar mansion which applies to an interval
of the ecliptic (see Lunar Mansions 19 and 20).
Constellation 15.    The Arrow [Sagitta]
FIGURE 60
  Aratus (lines 311-312) says that above Sagit-
tarius, the archer, is an arrow {olarbs) alone with-
out a bow. In Figure 60 Sagittarius cannot be
seen, but would be to the south (right) of the
small bird. The arrow has been associated with
the legend of Hercules killing the eagle gnawing
the liver of Prometheus, In the Islamic world the
name was usually translated sahm (arrow) as in
Figure 60. The globes of the eleventh century
(Nos. 1 and 34) use the word "anajah, which
means javelin. The constellation consists of five
stars in a nearly straight line. No stars are greater
than fourth magnitude, and none bear special
names.
Constellation 16.    The Eagle [Aquila]
FIGURE 60
  Aratus (lines 312-315) says that near Cygnus,
whose head is seen at the north side of Figure
60, "another bird tosses in storm, of smaller size
but cruel in its rising from the sea when the night
is waning, and men call it the Storm-bird {ayjTos)"
([trns.] Mair, 231). The later word used was aerbs
(an eagle). The eagle (Hyginus 11,16) was
interpreted as either that killed by Hercules or
the eagle Zeus used to carry Ganymede to
Olympus. In Figure 60 the title reads surat "uqdb
(the constellation of the eagle), a simple
translation of the Greek a'erbs. It consists of nine
formed stars and six unformed ones, which are
south of the celestial equator passing through
Aquila, with one in the graduations. Al-BTrunT
{Astrol., sec. 160) says the eagle is lighting upon
the arrow, and indeed on Smithsonian globe,
Aquila does appear to be landing upon Sagitta.
On the earliest Arabic globes the eagle actually
sits on the arrow.
  
156

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY










FIGURE 60.—The Arrow [Sagitta], the Eagle [Aquila], the Dolphin [Delphinus], and the Part
of a Horse [Equuleus]. (Photo: Smithsonian no. 72-3087)

NUMBER 46

157



  Written on the breast of the eagle is the word
td"ir (the flying bird) and applies to the large star
on the upper part of the northern wing. [no. 3;
a Aquilae]. This is the eleventh brightest star,
and the Arabic term is the source for the modern
name Altair. A Bedouin tradition (al-SQfi, Suwar,
111) applied the name al-nasr al-td"ir (the flying
eagle) to three bright stars in a row: a Aql; the
large star in the head [no. 2; fi Aql]; and the large
star on the inside edge of the northern wing [no.
5; 7 Aql]. Al-Sufi also applies the name "the flying
eagle" to the constellation as a whole, while al-
BTrunT {Astrol., sec. 163) restricts it to the
brightest star, a Aql. It may be that there was a
Babylonian common antecedent for both the
Greek and Bedouin concepts of an eagle for this
region of the sky (Hommel, 595; Kunitzsch
[1961], no. 194a). The ninth star, in the tail, is
labeled in Figure 60 dhanab al-"uqdb (the tail of
the eagle). This is not known to occur on any
globe prior to those made in Lahore.
Constellation 17.    The Dolphin
[Delphinus]
FIGURE 60
  Aratus (lines 316-318) speaks of "the Dolphin
{beXipls) with few bright stars and body wreathed
in mist, but four brilliants adorn him, set side by
side in pairs" ([trns.], Mair, p. 231). One legend
has it that the dolphin was placed in the heavens
by Poseidon as a reward for persuading Amphi-
trite to marry Poseidon (Erat. Catast ,31). Others
say it is the dolphin who bore Arion the citharist
(a poet of the late seventh century BC) to shore
safely when servants wishing to kill him threw
him overboard.
  The Arabic name, written near the mouth in
Figure 60, is siirat al-ddlfin, a transliteration of
the later Greek form of the word for dolphin
beXiplv (see Kunitzsch [1974], 186), Al-BTrunT
{Astrol., 160) says that al-ddlfin is a marine animal
that resembles an inflated wine-skin {al-ziqq al-
manfukh). The constellation contains 10 formed
stars, of which only one is named on the Smith-
sonian globe—the largest of the three in the tail,
labeled dhanab al-ddlfin (tail of the dolphin [no,
1; 6 Delphini]).
  
The four bright stars that form a rhomboid
mentioned by Aratus are the four large ones
toward the front of the dolphin [nos, 4,5,6,7;
fiaby Del]. These four were in the Bedouin tra-
dition called al-qa"ud, which is difficult to trans-
late, but seems to mean "the young camel" (see
Kunitzsch [1961], no. 234). There are variant
readings for the word, such as al-"uqud (neck-
lace). Another Bedouin tradition gave the name
al-salib (the cross) to these same four stars. In
keeping with this image al-SQfi {Suwar, 116) also
called the star in the tail [e Del] "amdd al-salib
(the column of the cross). Al-BTrunT [Astrol. sec.
163) speaks of salib al-td"ir (the cross of al-td"ir
[a Aquilae'^. This asterism is known today as Job's
Coffin. Another group of four stars was given
the parallel name salib al-wdqi" (the cross of al-
wdqi" [a Lyrae]). Al-Sufi {Suwar, 63) identifies
this as no. 19 of Hercules [t Herculis] and nos.
3,5,4 of Draco [fiy^ Draconis]. Later writers iden-
tify it with the "camel-mothers" in Draco [vfi^y
Draconis; see Kunitzsch [1961], no. 278].
      Constellation 18.
Part of a Horse [Equuleus]
FIGURE 60
  The constellation having the form of a head of
a horse is not mentioned by Aratus. It is called
by Geminus and Ptolemy 'iinrov TrpoTOfiT], mean-
ing the front part of a horse. It can be seen imme-
diately above the Dolphin. In Figure 60 it is
labeled near the muzzle qit"at al-faras (the part
of a horse). Al-BTrQnT {Astrol., sec. 160) calls it
"the first horse" in reference to Pegasus, whose
head can be seen next to this constellation.
Equuleus is composed of four obscure stars, none
of which are given individual names.
Constellation 19.    The Larger Horse
[Pegasus]
FIGURE 61
  Aratus (lines 205-215) wrote of this constel-
lation after describing Andromeda, whose head
can be seen in the upper corner of Figure 61
([trns.] Mair, 223, 225):
  
158

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY







FIGURE 61.—The Larger Horse [Pegasus]. (Photo: Smithsonian no. 72-3088)

NUMBER 46

159



Beneath [Andromeda's] head is spread the huge Horse,
touching her with his lower belly. One common star gleams
on the Horse's navel and the crown of her head. Three
other separate stars, large and bright, at equal distance set
on flank and shoulders, trace a square upon the Horse. His
head is not so brightly marked, nor his neck, though it be
long. But the farthest star on his blazing nostril could fitly
rival the former four, that invest him with such splendor.
Nor is he four-footed. Parted at the navel, with only half a
body, wheels in heaven the sacred horse.
  Later mythographers (e.g., Hyginus II, 18;
Erat. Catast, 18) associated the horse with
Pegasus, the offspring of Poseidon and the Gor-
gon Medusa and the horse used by Bellerophon,
as well as with Melanippe, who was the daughter
of Chiron the Centaur, who when born was
placed in the skies by the gods so that Chiron
(Centaurus) could not see her and with only half
a body so that her sex could not be determined,
Aratus had not mentioned wings, and Eratos-
thenes {Catast, 18) notes that Pegasus's flight is
unbelievable since it has no wings. But by Ptole-
my's time wings were clearly a part of the con-
stellation.
  In Arabic this Ptolemaic front half of a winged
horse is most frequently called "the larger horse''
{faras a"zam), as is written in Figure 61 along
the neck. The larger horse is also occasionally
called "the second horse" {al-faras al-thdni) as on
globes Nos. 1 and 34, (Equuleus was the first
horse). It was also known as the "winged horse"
{faras mujannah). It consists of 20 formed stars.
A medium-sized plug is seen in Figure 61 near
the tip of the wing. The bright star on its nostril,
no. 17, is labeled fam al-faras (the mouth of the
horse). This star is also called anf al-faras (the
nose of the horse) on globe No. 3. The star is e
Pegasi, whose modern name Enif is derived from
the Arabic word for nose (cf. Kunitzsch [1974],
259).
  The square formed by the four bright stars
was called by the Bedouins al-dalw (the bucket).
See discussion of Lunar Mansions 26 and 27 for
this traditional Arab asterism. These are known
today as the Square of Pegasus. The star at the
upper edge of the wing [no. 4; a Peg] is labeled
in Figure 61 matn al-faras (side of the horse).

The star near the join of wing and lower leg [no.
3; fi Peg] is labeled mankib al-faras (shoulder of
the horse). The star in the southeast corner of
the body [no. 2; y Peg, Algenib] is named jindh
al-faras (wing of the horse), while the one shared
with the head of Andromeda [no. 1; b Peg or a
Andromedae, Sirrah] bears two labels: surrat al-
faras (navel of the horse, the source of the mod-
ern name), and ra"s al-mar-'ah al-musalsalah
(head of the chained woman, referring to Andro-
meda). This latter star is numbered as a star in
Pegasus. These individual names of the four stars
composing the square all reflect the Ptolemaic
image rather than the earlier Bedouin imagery.
  The traditional image of the leather bucket is
evident on the Smithsonian globe in the title
given the pair of stars close together at the back
of the wing [nos. 5,6; ri; Peg]. The inscription
appears to read al-kub (the cup; written in an
identical manner on globes Nos. 11 and 13 by
the maker of this unsigned globe), but this is
surely an error for al-karab, which differs only
slightly in the formation of the middle letter. Al-
karab means the place where the rope is attached
to a bucket, and was used by al-SQfi {Suwar, 122)
in reference to these two stars.
  There are in addition three groups of two
stars, each bearing traditional Arab labels on the
Smithsonian globe. These are three of the 10
groups of sa"d stars, sa"d being roughly trans-
lated as omen. The attachment of sa"d to certain
stars is very ancient and its meaning obscure;
consequently it is best to leave it untranslated
(see discussion of Lunar Mansions 22-25 for
further details on the sa"d stars). The two largest
of the four stars grouped together on the neck
are labeled sa"d al-humdm {sa"d of the hero; but
could also mean sa"d of the sleet or hail [nos.
11,12; f^ Peg]). On the stomach between the two
legs, two stars are labeled sa"d bdri" (an excellent
sa"d [nos. 9,10; Xp. Peg]). The two stars together
on the lower leg are named sa"d matar (a rainy
sa"d [nos. 7,8; rjo Peg^. Al-SQfT calls the two in
the head [Bv Peg] sa"d al-bihdm {sa"d of the young
animals [see Kunitzsch [1961], no. 257, 2a, 2b,
2c for variants]). But this last term is not pro-
  
160

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



vided on the Smithsonian globe.
  The star near the hoof of the lower leg [no.
18; TT Peg] bears an undecipherable inscription:
mrhlat al-faras. This star does not bear an indi-
vidual name in any of the literature or on other
studied globes, with the exception of Nos. 11
and 13, which have the same words and are no
doubt made by the same maker who produced
the Smithsonian globe.
Constellation 20.    The Chained
Woman [Andromeda]
FIGURE 62
  Aratus (lines 197-204) described Andromeda
{avbpbpeba) as having her arms outstretched and
bearing bonds or fetters {beapa), for according
to legend she was bound to a rock as a sacrifice
to the sea-monster Cetus, by her parents Cassio-
peia and Cepheus, until she was rescued by Per-
seus, whose figure can be seen beneath Andro-
meda's.
  In Figure 62 she is depicted with her arms
outstretched, but with no chains. Occasionally on
globes she is depicted wearing chains: on No. 3
they are between her hands, and on No. 26 they
are between her feet. To the south along her
side is the northern of the two fishes comprising
the constellation Pisces. Two additional repre-
sentations of Andromeda are found in al-SQfi
manuscripts, in which a fish lies across her feet
or across her body. These conceptions are drawn
from the Bedouin rather than the Ptolemaic
traditions and do not appear on any known
globe.
  The constellation contains 24 formed stars
including the one which in Figure 62 is outside
the figure, above the northern hand. The star
on her head [a Andromedae] is counted in the
star catalogs as a star of Pegasus, so that the total
of formed stars for this constellation in the cata-
logs is 23, with no unformed stars. A medium
sized plug can be seen under the north arm.
  The title written near the northern side of her
head reads surat al-mar"ah al-musalsalah (the
constellation of the chained woman), which is the

the usual designation. Al-SufT {Suwar, 125) adds,
however, that she is also called al-mar"'ah allati
lam tarra ba"lan (the woman who never had a
husband) in an attempt to derive the name An-
dromeda from avbpa pv eibe (she who does not
know a man) (cf Kunitzsch [1974], 187-188).
The star in her head [a And or b Pegasi; Sirrah]
is labeled ra"s al-mar"ah al-musalsalah (the head
of the Chained Woman) and also surrat al-faras
(navel of the horse), for it is shared with Pegasus.
  The star on her waist near the fish is given
two names as well: jan^ al-musalsalah (the side of
the Chained One) and batn al-hut (belly of the
fish [no. 12; jS And, Mirach), the first reflecting
the Ptolemaic form and the latter the Bedouin
image of a large fish. The latter name was given
also to the twenty-eighth lunar mansion (see sec-
tion on Lunar Mansions), Al-Sufi {Suwar, 126)
also called the star, in addition to the two pre-
vious names, al-rishd" (the rope) which is also a
name for the twenty-eighth lunar mansion. In
this Bedouin image the rope was furnished for
the leather bucket which was seen in the asterism
of the Square of Pegasus.
  On the Smithsonian globe rijl al-musalsalah
(the foot of the Chained One) is the title given
the star in her eastern foot [no. 15; 7 And].
While this Ptolemaic name is not common on
Class A globes, it does occur on Mughal globes
e.g., Nos. 11 and 13), Al-Sufi {Suwar, 126) says
it is given this name on astrolabes, and it occurs
on Ulugh Beg's personal copy of al-SQfi's treatise
(see Wellesz [1959], figure 70), The traditional
Bedouin term for the star was "andq al-ard (the
wild lynx), which gave rise, rather circuitously,
to the modern name of the star, Almach (see
Kunitzsch [1961], no, 34a, 34b), The latter term
occurs on some early globes (e,g., Nos. 3, 4, and
5).
  Al-Sufi {Suwar, 132-134) appended to his dis-
cussion of Andromeda a description and illustra-
tion of three early Bedouin constellations—two
different fishes overlapping Andromeda and a
complete horse {al-faras al-kdmil). In these dis-
cussions al-SQfT clearly described the Andromeda
Nebula (M31), which would be located about
  
NUMBER 46

161


FIGURE 62.—The Chained Woman [Andromeda] and the Triangle [Triangulum]. (Photo:
Smithsonian no. 72-3085)

162

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



where the plug under Andromeda's arm is on
the Smithsonian globe, but he did not place it in
the catalog. It is not known to be indicated on
any extant globe, though it is shown in some
illustrations of the fish in Andromeda in al-SQfT
manuscripts (see Wellesz [1965], plate 14).
Constellation 21.    The Triangle
[Triangulum]
FIGURE 62
  The constellation, following Ptolemaic speci-
fications, consists of four stars. Aratus called it
beXrijiTov, and mythographers (Hyginus, II, 19)
suggested that Hermes placed it above the head
of Aries so that its brightness would mark the
dimness of Aries and so that its triangular shape
would form the first letter of the name for Zeus
(Ais, an older form of Zevs). While the constella-
tion bears no formal title in Figure 62, the star
at the top of the triangle is called ra"s al-muthal-
lath (the apex of the triangle [a Trianguli]). Al-
SQfi adds that this star, together with the most
northerly of the three at the base, were tradition-
ally called al-anisdn (the two friends) [afi Tri].
While this name does not appear on the Smith-
sonian globe, on globe No. 5 the name al-anisdn
is inlaid in silver while ra~'s al-muthallath (the
apex of the triangle) is merely engraved, in an
attempt to indicate that the former is the older,
more traditional term within the Arabic world,
and the latter the later, secondary designation.
Constellation 22.    The Ram
[Aries]
FIGURE 63
  Aries the Ram is the first of the zodiacal signs,
having been at one time the sign in which the
vernal equinox occurred. Due to precession,
however, the equator and ecliptic on the Smith-
sonian globe meet in the constellation of Pisces
instead of Aries. The first segment of the zodiac
(1 ° to 30°) after the junction marking the vernal
equinox is still labeled al-hamal (the young ram).

written on the north side of the ecliptic in long
stretched out letters, for this segment of the
ecliptic is, for astrological purposes, the House
of Aries. The segment of the zodiac in which the
constellation Aries, or most of it, occurs on the
Smithsonian globe is labeled al-thaur (the bull,
or Taurus). Thus the line separating the House
of Aries from the House of Taurus passes
through the horns and near front leg of the
constellation Aries. On the southern side of the
ecliptic three names of lunar mansions can be
seen: in front of the forward knee of Aries there
is butayn, the second lunar mansion; beneath the
belly thurayyd, the third lunar mansion; and near
the tail of Aries is dabardn, the fourth mansion
(see the section in Chapter 5 on Lunar Mansions).
  Aratus (lines 225-232) placed the Ram, Kpibs,
near Andromeda and the Triangle. This is the
legendary Ram that bore the golden fleece, the
object of the Argonaut's quest; it was the Ram
that carried Phrixus and Helen through the Hel-
lespont (Hyginus, ii.20). The Ram was sacrificed
to Zeus, who placed it among the heavens,
though others say (Erat. Catast., 19) that the ram
gave its golden fleece to Phrixus as a memorial
and then passed up to the heavens of its own
accord. Hyginus (II, 20) adds the story that when
Dionysus attacked Africa the army found them-
selves without water, A ram showed them the
way to water, for which reason Dionysus placed
the ram in the heavens so that when the sun was
in that constellation all living things would be
refreshed and so that it would be the chief of all
the signs since it had been the best leader of his
army.
  In Figure 63 we see a faithful rendition of the
classical form of the constellation, A fairly large
plug in the globe can be seen across his front leg.
Overhead the ram is labeled surat al-hamal (the
constellation of the young ram). The constella-
tion consists of 13 formed stars and 5 unformed
ones. Of the unformed, four are in a group over
the rump and one is on top the head. Al-Sufi
observed an additional formed star in the back
rear upper leg, but did not number it nor enter
it in his catalog, nor is it indicated on the globe.
  
NUMBER 46

163


FIGURE 63.—The Ram [Aries]. (Photo: Smithsonian no. 72-3089)

164

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



  In Figure 63, three stars are labeled, all reflect-
ing traditional Arab imagery. Of the two stars
very close together in the horns, the one closest
to the outside of the horns is called muqaddam
al-sharatayn (the anterior one of the two signals).
This star is no. 1 of the constellation [fi Arietis].
The second star of the pair in the horns is labeled
mw'akhkhar al-sharatayn (the posterior of the two
signals) and refers to no. 2 of Aries [7 Arietis],
The name al-sharatayn (the two signals) is the
name of the first lunar mansion, which derived
originally from these two stars. On the Smithson-
ian globe, however, the former names apply just
as star-names, while the title al-sharatayn is writ-
ten as a lunar mansion at 13° House of Aries,
rather than in the constellation of Aries. The
star at the base of the tail, the first star of the
three encircled by the tail, and the star on the
back thigh form a triangle, which also gave rise
to a lunar mansion. The name of the mansion is
not written in the constellation of Aries as a star
name, but only along the ecliptic at 25^2° House
of Aries (just before the knee of the constellation
Aries), where it reads al-butayn, the second lunar
mansion.
  The large star at the top of the head, the first
of the unformed stars, is labeled in Figure 63 al-
ndtih (that which butts or gores). This is a Arietis,
whose modern name Hamal derives from the
word for ram used for the entire constellation.
Al-SQfT {Suwar, 142) used al-ndtih for a star
name, while most of the writers in the Arabic
world used it as another name for the first lunar
mansion (cf. Ideler, 132; Kunitzsch [1961], nos.
198, 199, 200). See the section on Lunar Man-
sions for further details on such star names.
Constellation 23.    The Bull
[Taurus]
FIGURE 64
  According to Aratus (lines 167-178) the Bull,
ravpos, is found crouching at the feet of the
charioteer, who can be seen in the upper left-
hand corner of Figure 64, and with whom Tau-
rus shares a star. Aratus adds that the Bull's head

is well marked by the group of stars called the
Hyades. This star group was one of the ones
mentioned by Homer {II., XVIII, 486) as were
the Pleiades {Od., V, 272), which are also in
Taurus. The legend of Europa and the bull was
commonly linked to this constellation (e.g., Hy-
ginus, II, 21). In the star group of the Hyades,
Hesiod (frag. 60) names only five stars: Phaisule,
Koronis, Klaeia, Phaio, and Eudora, and for him
{Works and Days, 615) their cosmical setting was
a time for ploughing. Later mythographers (Hy-
ginus, II, 21) counted seven Hyades and made
them nymphs who nursed Dionysus. They were
named Eudora, Ambrosia, Koronis, Pedile, Po-
lyxo, Phyto, and Thyone. They were also said to
be sisters who cried themselves to death when
their brother Hyas was killed hunting. The word
Hyades {vabes) is according to some derived from
veiv (to rain), although others give an alternative
derivation from vs (swine), since it was sometimes
thought to be a sow with four young (Neuge-
bauer OCD, 282). Cicero {Nat deo., II, 43) says
the word comes from hyein, for they bring rain
"while our nation stupidly names them the Suck-
ling-pigs, as though the same Hyades were de-
rived from the word for 'pig' and not from
'rain.'" For legends of the Hyades, see Martin
([1956], 79-89).
  In Islamic representations only the first half
of a charging bull is depicted, following Ptole-
my's description {'ev rrj aitoToprj). In Figure 64
the line dividing the zodiacal House of Taurus
from the House of Gemini passes through the
body of Taurus, so that the constellation actually
occupies the area between about 50° and 80°
along the ecliptic. Three lunar mansions are
written along the ecliptic: one at the back of
Taurus under the tail of Aries, dabardn; haq"ah
written near the head of Taurus; and han"ah
between the horns. These are Lunar Mansions
4, 5, and 6, respectively. In the middle of the
bull is written surat al-thawr (the constellation of
the Bull).
  The cluster of five small stars, the Pleiades, on
the shoulder of the Bull are called al-thurayyd
roughly   translated  as   "the  brilliant  gem"—a
  
NUMBER 46

165


FIGURE 64.—The Bull [Taurus]. (Photo: Smithsonian no. 72-3095)

166

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



name also used for the third lunar mansion of
which the Pleiades formed the central part (see
discussion of Lunar Mansion 3; included are
Greek accounts of this asterism). Ptolemy, al-
SQfT, and Ulugh Beg all list only four stars mak-
ing up the Pleiades. However, al-SQfT adds that
in the space within the group there were two or
three stars in addition to the four, like a cluster
of grapes {"unqub al-"inab) (cf. Ideler 137, 390).
Al-SQfT {Suwar, 151) explains "the stars of the
Pleiades exceed these 4 which we have men-
tioned, but I confined myself to these 4 because
they are very close and the 4 are of sufficient
greatness in magnitude so that they are cited and
their names not omitted," adding that the Arabs
also considered it to have the dignity of one star
which they called al-najm (the Star par excellence).
They also called it nujdm al-thurayyd (the stars of
al-thurayyd). The Smithsonian globe curiously
has clearly inlaid five stars for the Pleiades.
  The largest star of the five on the head is
labeled "ayn al-thawr (the eye of the bull). It was
numbered fourteenth in Taurus and is the thir-
teenth brightest star in the heavens [a Tauri,
Aldebaran]. Al-SQfT {Suwar, 154) says it is a
bright red large star traditionally called al-da-
bardn from a root meaning "to follow," because
it followed the Pleiades, adding that al-dabardn
is the name of the fourth lunar mansion and is
used on astrolabes. The maker of the Smithson-
ian globe used the name "eye of the bull" for the
star, reserving dabardn for the lunar mansion.
The open cluster called the Hyades in the Greek
world are represented by a group of five stars in
the face of the bull, the brightest being al-da-
bardn [nos. 11, 12, 13, 14, 15; yb^d^^^ae Tau].
Al-SQfT recorded an additional four stars in this
group, two between each pair of large ones, one
being just outside the face, none of which he
entered in his catalog. Another traditional name
for Aldebaran was al-faniq, the camel-stallion,
because, as al-BTrQnT {Chron. [trns.], 344) said,
they called the stars around it (the other Hyades)
the young camels {al-qilds).
  The star in the northern horn, which is shared
with Auriga, is not counted among the stars of

Taurus, thus giving a total of 32 formed stars
for Taurus. The Smithsonian globe has 34 inlaid
stars, one of which is the star in the northern
horn [fi Tau] and one an extra Pleiade. There
are 11 unformed stars, one near the western
hoof of the bull (in the line of single degree
graduations of the equator), two above the head
of Orion whose figure is beneath the bull's horns,
and eight between the horns, extending into the
line of the ecliptic.
Constellation 24.    Al-Jawza",
or The Twins [Gemini]
FIGURE 65
  The twins, which Aratus (line 147) called
blbvpoi, were identified by later mythographers
with various pairs including Apollo and Heracles,
and Castor and Pollux. The latter were the most
affectionate of all brothers and were placed in
the heavens as a record of unselfish friendship
(Hyginus, II, 22). The myth of Castor and Pollux
gave rise to the modern names for the two most
prominent stars in the constellation. The twins
were described in Greek literature as one having
his arms around the other, as they appear in
Figure 65, The constellation of the twins seems
to go back to Babylonian texts (see Gundel 8c
Boker, cols. 522-528).
  The figure of Gemini is drawn in the zodiacal
House of Cancer, the name of which {al-saratdn)
is written along the north side of the ecliptic.
The line dividing the House of Gemini from that
of Cancer passes through the feet of the western
twin. Three lunar mansions are inscribed along
the ecliptic: dhird", the seventh mansion derived
from stars in the constellation of Gemini, is writ-
ten north of the ecliptic at the division of the
House of Cancer and Gemini (the summer sol-
stice); nathrah, the eighth mansion, is written on
the stomach of the second twin, marking the
point at 103° of the ecliptic; and tarf the ninth
mansion, is engraved south of the ecliptic be-
tween the easternmost hand of the twins and
Cancer the crab, marking the point at 115y4°.
  
NUMBER 46

167


FIGURE 65.—Al-Jawzd^ or the Twins [Gemini]. (Photo: Smithsonian no. 72-8588)

168

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



  The constellation is titled surat al-jawzd" (the
constellation al-jawzd"). The meaning of al-
jawzd" is somewhat obscure. It comes from a root
meaning "to marry' and is the traditional Bed-
ouin name for the region, as al-BTrQnT {Astrol.,
sec. 159) notes. The name al-jawzd" possibly re-
ferred to a large feminine figure envisioned in
this region of the heavens in pre-Islamic Arabia
(see section on Classical Greek and Pre-Islamic
Sources and Kunitzsch [1961], 23-25). Al-SQfT
{Suwar, 160) employed the name al-taw"amdn
(the two twins) for the constellation. This name
is the Arabic translation of the Ptolemaic blbvpoi
(twins) and occurs on the early globes (Nos. 1
and 34) as well as some later ones (No. 31). For
the most part, however, later astronomers re-
turned to the early Bedouin name for the region
{al-jawzd") for the name of the constellation as
well as the corresponding zodiacal house.
  Gemini is composed of 18 formed stars and
seven unformed ones. Of the unformed ones,
one is west of the western hand of the second
(eastern) twin, one is west of the forward foot of
the first (western) twin, one between the legs of
the two twins, and four near the eastern arm of
the back twin. The small dot between the legs of
the eastern twin is in fact only a smudge on the
surface of the globe.
  Over the head of the westernmost twin in
Figure 65 reads the label ra"s al-taw"am al-muqad-
dam (the head of the foremost twin), referring to
the first star, the one in the forehead [a Gemi-
norum, Castor]. The star in the face of the eastern
twin is labeled ra"s al-taw"am al-mu"akhkhar (the
head of the rear twin), the second star of the
constellation [fi Gem, Pollux]. Only these two
stars are labeled on the Smithsonian globe; both
labels reflect the Ptolemaic constellation image.
  The more common traditional Bedouin des-
ignation for the two stars taken together was al-
dhird" (the foreleg), which refers to the pre-
Islamic concept of a lion covering a large part of
the heavens. When the two stars are given this
name they are generally considered as the sev-
enth lunar mansion, for which reason this name
appears on the ecliptic as the seventh lunar man-

sion at the division of the House of Gemini and
the House of Cancer (the summer solstice) (see
discussion of Lunar Mansion 7 for various inter-
pretations of dhird"). On all globes examined,
except those by the Lahore family and the two
earliest (Nos. 1 and 34), which have very few star
names, the stars in the heads of the twins are
labeled with the Bedouin term al-dhird" rather
than with Ptolemaic labels. Moreover, on all the
globes except the two earliest and the Mughal
ones, the stars in the two feet of the eastern twin
[nos. 17, 18; 7^ Gem] were labeled al-han"ah.
This term is also the name of the sixth lunar
mansion and gave rise to the "modern" name for
the star 7 Gem, Alhena. On the Smithsonian
globe these two stars are not labeled in the con-
stellation of Gemini, but the name appears as a
lunar mansion along the ecliptic (see the discus-
sion of Lunar Mansion 6).
Constellation 25.    The Crab
[Cancer]
FIGURE 66
  In the Hellenistic world this constellation, the
most inconspicuous of all zodiacal signs, was usu-
ally called KapKlvos (a crab), latinized as Cancer.
Aratus (lines 892-908) calls it not only KapKlvos,
but the manger or crib {iparvrj) between two asses
{OVOL). This image of the crib and asses was com-
mon in later poetry. Theocritus (XXII, 21) says
"the clouds disperse this way and that, the bears
are seen again, and between the asses the dim
crib, betokening that all is fair for voyaging"
([trns.] Gow, 159; cf Virgil Georgics, 1,356, ff).
Mythographers (Hyginus, 11,23) said that Dion-
ysus placed the two asses in the heavens for they
had helped him flee across a swamp to a temple
in order to restore his sanity. Another story (Erat.
Catast, 11) says that the two asses brayed so
loudly when Zeus was fighting the Giants, that
the Giants took flight and were defeated. The
crab was explained (Hyginus, 11,23) as being one
which had snapped at the foot of Heracles (Her-
cules) from the swamp when he fought the Ler-
  
NUMBER 46

169


FIGURE 66.—The Crab [Cancer]. (Photo: Smithsonian no. 72-8589)

170

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



naean Hydra. Ptolemy named the constellation
Cancer {KUPKIVOS) and noted that its first star was
nebulous and called the crib {iparvrj latinized as
Praesepe; see Gundel & Boker, col. 522 for Me-
sopotamian parallels).
  In Arabic the constellation is called al-saratdn
(the crab). The inscription can be seen in Figure
66 behind the southern back leg. It is a fanciful
representation, the back feet even having toes.
  The line dividing the House of Cancer from
the House of Leo passes through the constella-
tion of the crab. Two lunar mansions are written
on the south side of the ecliptic: one at the tail
of Cancer, tarf the ninth mansion marking
115y4° of the ecliptic; and jabhah the tenth lunar
mansion, written next to the head of the crab at
128° along the ecliptic.
  No stars in the constellation are named on the
Smithsonian globe. The three stars in a row
across the breast of the crab, however, frequently
do bear names on other globes and in the astro-
nomical literature, all derived from Greek con-
cepts of the constellation. The middle star of
these three (no. 1) is the open cluster [M44,
Praesepe]. Al-SQfT termed it ma"laf (a manger,
or stable) calling it sahdbi {cloudy). The two stars
on either side of the Manger, nos. 4 and 5 of the
constellation [yb Cancri], were called by Aratus
and Ptolemy "asses." Al-SQfT also termed the two
stars al-himdrayn (the two asses).
  Alongside this terminology of Greek origin,
however, al-SQfT also presented a traditional Be-
douin system of terms. In this system the open
cluster Praesepe was called al-nathrah (the carti-
lage of the nose), and as such was the name of
the eighth lunar mansion (see subsection discus-
sion of Lunar Mansion 8 for related terms). This
name arose also from the Bedouin conception of
the lion, larger than the Ptolemaic Leo, in that
part of the heavens. On the Smithsonian globe
the term nathrah appears not in the constellation
of Cancer, but in the zodiacal House of Cancer
along the ecliptic as a lunar mansion. In the
Bedouin tradition the open cluster Praesepe was
also called al-lahdh (the uvula [of the lion]).
In keeping with the image of the large lion.

the two unformed stars in front of and slightly
above the head of the crab, nos. 3 and 4 of the
unformed stars, were traditionally called al-tarf
(the glance of the eyes [of the lion]; [v^ Cancri]).
Al-tarf constitutes the ninth lunar mansion, and
as such is written, not alongside the two stars,
but in the House of Cancer along the ecliptic
immediately behind the crab. Al-SQfT {Suwar,
173) adds that the Arabs called the first un-
formed star (above the southern claw), along with
a small one behind al-tarf al-ashfdr (the places
where the eyelashes grow).
  The constellation consists of nine formed stars
and four unformed ones, the latter being the two
above the head and the two near the southern
claw. It is assumed that the position of the star
touching the northern claw is an error of the
artisan and that it was intended to be a formed
star.
Constellation 26.    The Lion [Leo]
FIGURE 67
  Aratus (lines 148-155) associated the Lion
{X'eojv) with the sun's hottest path. Greek mythog-
raphers said that Leo was the Nemean lion killed
by Heracles during his first labor. In the Arabic
world the constellation was called surat al-asad
(the constellation of the lion), written in Figure
67 above the head. A larger lion was seen in this
region of the sky in the Bedouin tradition, only
part of which overlaps with the Ptolemaic Leo.
The constellation of Leo occupies the area be-
tween 130° and 170° of the ecliptic, its head in
the House of Leo and its rear half in the House
of Virgo. Three lunar mansions are found along
the ecliptic: at 141 V2° between the front paws is
zubrah. Lunar Mansion 11; at 154V4° under the
belly is sarfah. Lunar Mansion 12; and at 167^4°
on the right back paw is "awd". Lunar Mansion
13.
  The star on the mouth farthest from the eclip-
tic is labeled mankhar al-asad (the nose of the
lion [no. 1; AC Leonis]). The star between the eyes
is labeled ra"s al-asad al-shamdli (the northern
  
NUMBER 46

171


FIGURE 67.—The Lion [Leo]. (Photo: Smithsonian no. 73-2658)

172

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



head of the lion [no. 3; p Leo]). The very large
star of the two contiguous ones on the upper
part of the lower forearm is termed qalb al-asad
(the heart of the lion [no. 8; a Leo]). This first
magnitude star was also called al-maliki{the royal
one) from Ptolemy's term fiaaiXloKos (a little
king), hence its modern name Regulus. The large
star on the rump is labeled zahr al-asad (the back
of the lion [no. 20; b Leo]). All of these star names
reflect the Greek leonine outline transposed over
the larger Arab one. The latter star [b Leo] bears
an additional title in Figure 67: min al-zubrah
(one of al-zubrah). The name is written again at
the star on top of the thigh, southeast of the
other [no. 22; 6 Leo]. These names indicate the
two stars that represent the mane of the larger
lion, al-zubrah meaning a mane. These stars gave
their name also to Lunar Mansion 11.
  Near two stars on the shoulders (nos. 5 and 6)
of Leo in Figure 67 there is written al-jabhah
(the forehead), which in the Bedouin tradition
referred not only to these two stars but also to
the one to the southwest (no. 9) and the very
large one further south (Regulus). These stars [f
y T] a Leo] were collectively called "the forehead
[of the large lion]" and formed Lunar Mansion
10. The modern name of 7 Leonis, Algeiba, is
derived from the Arabic word for forehead. The
word jabhah, representing the lunar mansion, is
also written at 128'/2° of the ecliptic.
  The star in the tail, no. 27, is labeled sarfah
(change of weather [fi Leo]). Since sarfah consti-
tuted Lunar Mansion 12, this same word is writ-
ten again as a mansion at 154'/4° of the ecliptic
(see section on Lunar Mansions for further dis-
cussion).
  The constellation of Leo consists of 27 formed
stars and 8 unformed ones. The latter are the
two over the back of Leo, the three under the
belly and in front of the hind legs, and the three
above the tail between Leo and Virgo and Ursa
Major. These last three, over the tail (nos. 6-8
of the unformed stars), constitute the asterism
called by Ptolemy irXoKapos, known in Latin as
Coma Berenices (Berenice's Hair). The asterism
Coma Berenices was identified and named by the
court astronomer to Ptolemy Euergetes in Alex-
andria, named Conon of Samos (see  Bulwer-

Thomas). He was a close friend of Archimedes
and named the asterism in honor of Ptolemy's
consort Berenice, who had vowed to dedicate a
lock of her hair in a temple if her husband
returned victorious from the Third Syrian War,
which began in 246 BC. Ptolemy III did return,
and the court astronomer preferred to place the
lock of hair in the skies. Mythographers and
poets wrote of the story (e.g., Hyginus, 11,24;
Callimachus 80-85, no. 110; Catullus, 66). The
astronomer Ptolemy, however, refers to it only
as a lock of hair, not mentioning Berenice. Al-
SQfT says that these three stars, which resemble
a right triangle [afiw Comae Berenices], were called
by Ptolemy al-dafirah, "a lock of hair" (cf. Kun-
itzsch [1974], 281-283). Al-SQfT {Suwar, 181-
182) states that they are all of fifth magnitude
and around them are many stars whose counting
is difficult because of their compactness, which
resembles the thickness of the Pleiades. The Be-
douin image was of these stars forming the tail
tuft on the tail of the enormous lion, for as al-
SQfT states {Suwar, 182):
These stars, together with the small close ones outside the
three, are called al-hulbah (the coarse hair), and that is
because outside al-sarfah (the star in Leo's tail) is a bent row
of stars ... so they liken these stars to the tail and the bright
star at the root of the tail [i.e. sarfah] to the scrotum; and
they compare the three which we have mentioned, with the
nearby small ones in their midst, to the hair which is at the
extremity of the tail. . . . The common people call these
stars together "the ripe grain of wheat" {al-sunbulah), and
many of the followers of the anwd^ reckon that for these
stars the House of Virgo is called al-sunbulah because they
resemble that through their compactness and their large
number.
The stars of Coma are unlabeled on the Smith-
sonian globe; on Nos. 4 and 5 they are encircled
by a decorative device which Virgo holds, as if
they were intended to go with Virgo rather than
Leo.
Constellation 27.    The Ear of Wheat
[Virgo]
FIGURE 68
  In Hellenistic times this constellation was de-
scribed as a maiden, wapdevos (Virgo in Latin),
with wings, holding in her hand an ear of wheat,
  
NUMBER 46

173


FIGURE 68.—The Ear of Wheat [Virgo]. (Photo: Smithsonian no. 72-3115)

174

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



arax^^ {spica in Latin). Aratus (lines 96-137)
adds that she was also called blKrj (Justice) and
dwelt on earth until the coming of the Bronze
Age when she ascended to the heavens. In her
role as Justice she was later associated with the
Scales (Libra). The association of an ear of wheat
with a maiden was also a symbol of the harvest,
and indeed the autumnal equinox occurs in this
sign. The northernmost star of the northern
wing (above the bend of the elbow) was called
TrpoTpvytjTTjp (that which rises before the vintage)
by Aratus and Ptolemy (in Latin Vindemiatrix,
the modern name of e Virginis). Mythographers
associated various figures with the maiden, in-
cluding Demeter (because of the wheat) and
Erigone, daughter of Icarus. The Arabic trans-
lation of -jrapdevos was al-"adhrd" (the maiden),
but the Bedouin name for the region soon re-
placed it in the literature. Thus al-sunbulah (the
ear of wheat), probably of ancient Sumerian or-
igin, became the standard name for the constel-
lation and the zodiacal house, as can be seen on
all extant globes, as well as in Figure 68 where
surat al-sunbulah (the constellation of the ear of
wheat) is engraved over the figure's northern
wing.
  Al-BTrQnT {Astrol., sec, 159) describes the Pto-
lemaic figure as "in the form of a maid with two
wings, in a flowing skirt and in her hand an ear
or two of wheat directed to the bottom of her
skirt." On all the globes that were examined the
constellation is depicted with wings, with her
northern arm extending toward Bootes, but
without an ear of wheat in her lower hand. On
globes No. 4 and No. 5 the figure holds a swatch
of hair or a decorative device incorporating the
three stars of Coma in her northern hand. The
three stars of Coma are visible in Figure 68 above
her northern hand. The line separating the
House of Virgo from the House of Libra passes
across the shoulders of Virgo. This is the point
of the autumnal equinox, and consequently the
equator is shown to intersect the ecliptic. On
Virgo's northern shoulder at the autumnal equi-
nox is the word simdk, the name of the fourteenth
lunar mansion. On her hip at 193° of the ecliptic,
the word ghafr is engraved, and near the hemline
at 205y4° is zubdnd, the names of the fifteenth

and sixteenth lunar mansions, respectively. See
the section on Lunar Mansions for further de-
tails.
  The very large star in the southern hand is the
only star named in Figure 68, and it is a Vir, the
fifteenth most brilliant star in the heavens, which
bears the modern name Spica from the Latin
rendering of the Greek word for an ear of wheat.
In the Arabic translation of Ptolemy's Almagest
araxvs was translated by the word al-sunbulah,
but the Bedouin word for the star soon domi-
nated. In Figure 68 the star is labeled simdk a"azal
(an unarmed simdk). The name simdk is of ancient
possibly Babylonian origin and its meaning has
been obscured with time (see Hommel, 596). In
the Bedouin tradition this simdk represents one
of the hind legs of the very large lion. The other
hind leg of the lion was formed by the other star
bearing the name simdk, Arcturus {al-simdk al-
rdmih), which can be seen in a direct line due
north of Spica, at the hemline of Bootes. Of the
star called simdk in Virgo, al-SQfi {Suwar, 193)
says:
They call it al-a^zal (unarmed) because it is opposite al-simdk
al-rdmih which is called ramih (armed) because the lance
{rumh) is on its right; and these are two bright stars [both
called simdk], one of which is near the foot of the Howler
[Bootes], whom they call the ox-driver, and the other one is
in the zodiac and is called "unarmed" because there is no
weapon accompanying it.
The "unarmed simdk" in Virgo gave its name to
the fourteenth lunar mansion.
  The large star in the northern wing was called
in Greek "the vintager" (in Latin, Vindemiatrix),
but bears no traditional Bedouin name and seems
not to have been as important to Bedouin im-
agery.
  Five stars forming roughly a 90° angle (begin-
ning with the star in the south wing, the one on
the southern shoulder, the one at the side of the
waist near the ecliptic, the large one in the middle
of the waist, and the large one on the northern
wing) formed the thirteenth lunar mansion, and
collectively [fivybe Vir] were called al-"awwd". On
the Smithsonian globe the name appears only as
a lunar mansion along the ecliptic. The star in
the southern foot, together with the two at the
hemline [nos. 22, 23, 25; IKX Vir] were tradition-
  
NUMBER 46

175



ally called al-ghafr and comprised the fifteenth
lunar mansion. The stars themselves are not la-
beled on the Smithsonian globe, although they
frequently are labeled on other globes. The con-
stellation of Virgo is composed of 26 formed
stars and six unformed ones. The latter are the
three along the southern arm (in the large grad-
uated portion of the ecliptic) and the three south
of the southern hand on the other side of the
ecliptic along the outstretched writing giving the
name of the House of Libra.
Constellation 28.    The Balance
[Libra]
FIGURES 69, 70
   In Greek times, as in much earlier Sumerian
times, the constellation now known as Libra was
seen as the two claws of the scorpion, with Scor-
pio and Libra combined essentially into one con-
stellation (see Gundel & Boker, cols. 522-529,
534; Hommel, 597). Thus both Aratus (lines 90,
438) and Ptolemy use the word xvXotl (claws) for
the constellation. The concept of the asterism as
a balance or scales {^vybs), which was a later
introduction, calls to mind the equality of day
and night at the autumnal equinox, which occurs
at the break between the House of Virgo and the
House of Libra (see Brown 1:71; Gleadow, 169).
Hyginus (11,26) states "this sign is divided into
two parts on account of the great spread of the
claws. One part of it our writers have called the
Balance. But the whole constellation together
was put in the sky, it is said, for the following
reasons" [trns] Grant). He then recounts that
Orion boasted he could kill anything on earth.
Earth, angered at this, sent the Scorpion, which
is said to have killed him. Zeus then placed Scor-
pio in the heavens as a lesson to men not to be
too self-confident. Artemis, who had suffered an
insult from Orion, requested Zeus that it be
placed so that when Scorpio rises, Orion sets.
  The name of the constellation translated into
Arabic as al-zubdnd, an old word of Sumerian
origin meaning "the claws," and also, using a
more recent word, as al-mizdn, or "the balance"
(see Kunitzsch [1974], 191). Figure 69 shows the
relation of the balance to the other constellations

on a globe made in 998 H/AD 1589-1590 by "AIT
KashmTrT ibn Luqman (No. 10; for details of the
set of scales depicted on the Smithsonian globe
see Figure 70). On the Smithsonian globe the
name of the constellation is surat al-mizdn (the
constellation of the balance). Al-SQfT {Suwar,
199) says that he has seen globes on which this
constellation was depicted as a man holding a
small pair of scales that had no stars in it, and in
fact on globe No. 1 the constellation is drawn in
a similar manner, with all the stars (both formed
and unformed) drawn within the form of the
man, but for one near and two inside the small
scale.
  On the Smithsonian globe the constellation of
Libra is drawn within the House of Scorpio,
taking up about 222° to 235° along the ecliptic.
At 218y4° of the ecliptic, the name of Lunar
Mansion 17, iklil, is written in the southern foot
of the balance (see discussion of Lunar Mansion
17 for the five interpretations of this lunar man-
sion, four of which refer to stars in Libra and
one to stars in Scorpio. Between the two parts of
the balance, at 23U/2°, the eighteenth lunar
mansion {qalb) is written. On the upper foot of
Serpentarius there is written shawlah, the nine-
teenth lunar mansion at 244° of the ecliptic, and
on the other foot na"d"im at 257° of the ecliptic,
the twentieth lunar mansion.
  The names of stars in Libra on the Smithson-
ian globe all reflect the Greek tradition. The
large star in the pan to the north (no. 3), is labeled
kiffah shamdli (the northern plate of the balance
[fi Librae]). The large star on the other pan (no.
1) is termed kiffah janubi (the southern plate of
the balance [a''^ Librae^). These two stars to-
gether make up the sixteenth lunar mansion
called al-zubdnd (the two claws [of the scorpi-
on]). This lunar mansion name reflects the more
ancient, probably Sumerian concept of the larger
Scorpion, of which this constellation formed the
claws. The name zubdnd on the Smithsonian
globe, however, is written only as a lunar man-
sion along the ecliptic and not as a star name.
  Libra is composed of eight formed stars and
nine unformed ones. The latter group consists
of the pair, one on either side of the lines sus-
pending the northern pan, the pair on either side
  
176

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



^      \       ^'

0^

-vvj^ii^



FIGURE 69.—The Balance [Libra]. Illustrated on an overall view of Globe No. 10, dated 998
H/AD 1589-1590, by "AIT KashmTrT ibn LQqman. London, private collection. (Photo: Alain
Brieux)
of the northern claw of Scorpio, one in front of	south of Libra. The star above the bar holding
the mouth of Scorpio, one at the suspensory	the scales and to  the south  is an  extra star
device, the pair, one on either side of the south-	(marked by an arrow in Figure 70) inadvertently
ern claw of Scorpio, and the large star to the	added by the maker.

NUMBER 46

177



Constellation 29.    The Scorpion
[Scorpio]
FIGURE 70
  In ancient Mesopotamia a large scorpion in
this region of the heavens was apparently associ-
ated with the darkness which comes on as the
year approaches the winter solstice, and it is one
of the most frequently depicted zodiacal signs
extant in Babylonian remains (see Gundel &
Boker, cols, 522-528, 696). In Greek literature
it was called aKopirlos. The Ptolemaic Scorpio in
the star catalog covers a smaller area than the
original scorpion asterism of Sumerian origin,
for the constellation Libra was formed from part
of the larger scorpion (see discussion of The
Balance [Libra]). The mythographers associated
with it the legend of the scorpion sent by Artemis
to kill Orion (Hyginus, II, 26 and II, 34; Erat.
Catast, 7 and 32; Ovid Fasti, V,535-544). The
Greek word for scorpion was translated by the
common Arabic word "aqrab, also meaning scor-
pion, as evidenced in the title of the constellation
written in Figure 70 near the base of the tail,
surat al-"aqrab (the constellation of the scorpion).
Al-SufT {Suwar, 207) says it is a famous {mash-
hurah) constellation.
  In Figure 70 the scorpion covers about 235°
to 265° of the ecliptic, its head being in the
House of Scorpio and most of its body and tail
in the House of Sagittarius. On the Smithsonian
globe the constellation is composed of 20 formed
stars and three unformed ones, the latter being
the one near the end of the tail and the two
below the lower foot of Serpentarius. The stars
in the foot of Serpentarius who is stepping on
the top of the Scorpion are counted with Serpen-
tarius and not with Scorpio, Ptolemy, al-SQfl,
and Ulugh Beg all catalog 21 formed stars in
Scorpio. The inexplicably missing star on the
Smithsonian globe should be located immedi-
ately to the north of the third star of the tail,
counting from the base of the tail. This missing
star, no, 14 in the catalogs, is part of a visual
double [f' Scorpii], no, 15 next to it being f^ Sco.
Curiously no. 13 of the constellation (the second
star of the tail) is also a visual double, ^''^ Sco,

but is counted as only one star in the catalogs.
The first unformed star, near the end of the tail,
was said to be nebulous {sahdbi). This is an open
cluster visible to the naked eye [OC 6475; M7].
  Only one star in Figure 70 appears to be
labeled. Near the last star in the body, in small
writing, is written al-"aqrab (the scorpion). One
would naturally apply this to the nearest star, no.
9 of the constellation [T SCO]. I can find no
evidence, however, on any globes nor in al-SufT
or Ulugh Beg for the use of this name with this
star. On the other hand, the middle of the three
stars near the mouth, no. 1, was later called al-
"aqrab, from which the modern name Akrab
arose [fi Sco]. It is unlikely, however, that the
maker would write the title at such a distance
from the star. The maker probably intended it
to go with the second star from the end of the
body, no. 8, which was called qalb al-"aqrab (the
heart of the scorpion [a 5co]) and simply omitted
the qalb. On globe No, 11, quite certainly by the
same maker who produced the Smithsonian
globe, qalb al-"aqrab is clearly written near the
eighth star of the constellation. The "heart of
the scorpion" was the traditional Arab name
given the bright red star called by Ptolemy
avraprjs {Antares), which is used as the name of
the star today (cf Kunitzsch [1974], 291-292),
The name Antares comes from avrl (similar to)
A.TJS (Mars)—that is, similar in color to the planet
Mars (Ideler, 181). In the Bedouin tradition the
two stars on either side of the "heart of
the scorpion" [nos, 7, 9; ar Sco] were called al-
niydt (the artery) for the aorta (see al-SQfT, Suwar,
209; Kunitzsch [1961], no, 205), The "heart of
the scorpion" (Antares) along with Vega, "the
falling eagle'' were together called al-harrdrdn
"the two whimpering [dogs]" who howl because
of the cold that usually sets in with the rising of
these two stars [a Sco and a Lyrae; see Kunitzsch
[1961], no. 116a]. The "heart of the scorpion"
also gave its name to the eighteenth lunar man-
sion and as such is written on the Smithsonian
globe simply as qalb at 231'72° of the ecliptic.
  The two stars on the tip of the tail, nos. 20
and 21, were called al-shawlah (the raised tail [of
the scorpion]) and formed the nineteenth lunar
  
178

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 70.—The Scorpion [Scorpio]. (Photo: Smithsonian no. 72-3114)

NUMBER 46

179



mansion [Xv Sco], The stars themselves in Figure
70 are not labeled, but the word shawlah as a
lunar mansion appears at 244° of the ecliptic.
The modern name of X Sco is Shaula (see the
section on Lunar Mansions).
Constellation 30.    The Archer,
that is, the Bowman
[Sagittarius]
FIGURE 71
  In Greek literature the Archer, TO^OTTJS, was
frequently represented as a centaur, although
this is not mentioned by Aratus (lines 301-307),
who says merely that when the sun scorches the
bow (TO^OI^) and the archer {ro^evrrjs) it is a time
of storm; the sailor should not sail at night but
put into shore in the evening. This is one of the
zodiacal signs that is probably of Babylonian or-
igin (see Gundel 8c Boker, cols. 522-528). The
Babylonian archer was depicted with wings,
which in Greek times became a mantle or cloak.
In Islamicate representations of the Ptolemaic
constellation, there is usually a fluttering end of
a head band or band of a turban, as in Figure 71
(see Wellesz [1959], 10), while al-BTrunT {Astrol.,
sec. 159) says he has long tresses. The mythog-
raphers (Erat. Catast, 28; Hyginus, 11,27) noted
that while some called it a Centaur others did
not because centaurs did not use arrows. The
legend given was that he was Crotus, son of
Euphene, nurse of the Muses. He was very clever
in the arts, for which reason the Muses requested
Zeus to represent him in a star group. Zeus added
the arrow to represent his keenness.
  In the indigenous Arab image for this region,
deriving from the common Sumerian origin, a
bow without a person was seen. The traditional
Arab name was al-qaws (the bow), which gradu-
ally replaced the translation of the Ptolemaic
name for the constellation and the related zodia-
cal house. Thus on the Smithsonian globe the
zodical house is labeled al-qaws (the bow). The
title of the constellation on the Smithsonian
globe, written in front of the head, reflects the
Arabic translation of the Ptolemaic title as well:
surat rdmi al-qawwds. This should be amended to

siirat al-rdmi wa huwa al-qawwds (the constellation
of the archer, i.e., the bowman). In Figure 71
the arm holding the arrow is drawn so that it
covers all but the arrowhead. Most of the con-
stellation, which occupies the area from 265° to
300° of the ecliptic, is in the House of Capricorn,
though the bow is in the House of Sagittarius.
  Three lunar mansions are written along the
ecliptic in Figure 71: baldah, the twenty-first, at
270° or the summer solstice (the name is written
at the bow); dhdbih, the twenty-second, at
282V4°, behind the neck of the archer; and bula",
the twenty-third, at 295y4°, in front of the head
of Capricorn.
  The archer is composed of 31 formed stars.
The small star behind the elbow of the Archer
[no. 19; h''^ Sagittarii] has incorrrectly been
placed outside the outlines of Sagittarius by the
maker of this globe. In addition, the large star
on the elbow [no. 18; x"'^ Sgr] is misplaced, for
it ought to be more to the north; moreover, it
has been indicated by a large silver dot when in
fact it was considered to be only of fifth magni-
tude.
  In Figure 71 the star in the eye is labeled "ayn
al-rdmi {the eye of the archer) and was considered
nebulous [no. 8; J/''^ Sgr]. The star near the hoof
of the lower front leg is labeled "urqub al-rdmi
(the archer's tendon) [no. 23; /3''^ Sgr]. Al-SQfi
{Suwar, 217-218) says this important star was
placed on astrolabes even though it is less than
fourth magnitude, correcting Ptolemy's magni-
tude of two. The star on the knee of the same
leg, no. 24, is labeled rukbat al-rdmi {the archer's
knee [a Sgr]); al-SufT also corrects its magnitude.
There is written near the star in the arrowhead
the word qaws (bow). It is unlikely this term was
intended to apply to this star [no. 1; y Sgr], for
such a term is not found on any globes nor in
texts. Star no. 1 was occasionally referred to as
zujj al-nushshdbah (the arrowhead of a wooden
arrow) or zujj al-sahm (al-SQfT, Suwar, figure 30;
Kunitzsch [ 1974], 294), also meaning arrowhead.
It is possible the maker simply intended it as a
general label for stars in the bow (although in
the literature only the six stars in the eye and
ribbon were called the bow; Al-SufT ^Mwar, 220),
  
180

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 71.—The Archer, that is, the Bowman [Sagittarius]. (Photo: Smithsonian no. 72-3112)

or that he intended it to apply to the star at the
southern tip of the bow [no. 3; e Sgr] whose
common name today is Kaus Aust,
The Milky Way is very rich in this area and

affected the traditional Bedouin terminology,
which is reflected in the names of the lunar
mansions derived from this constellation. Two
groups of four stars each were seen as ostriches.

NUMBER 46

181



four arriving at the river (the Milky Way) and
four leaving. The twentieth lunar mansion was
named for ostriches {na"d"im; see discussion of
Lunar Mansion 20 for details as to which stars in
Sagittarius made up the ostriches).
  Six stars (nos. 9-14) in a curve, consisting of
the three in the face (excepting the nebulous one
in the eye) and the first three of the fluttering
band, were called by al-Sufi al-qald"is (the young
ostriches) as well as al-qilddah (the necklace
[^OTcdpv Sgr\). The area under the fluttering
headband was called al-baldah (the place) and
was said to contain no stars, Baldah gave its name
to the twenty-first lunar mansion.
Constellation 31.    The Goat
[Capricornus]
FIGURE 72
  The tenth zodiacal sign, Capricorn, is also
unquestionably of Babylonian origin (see Gundel
8c Boker, cols, 522-528). It was seen originally
as a sea-goat, most often drawn with the tail of a
fish. In Greek it was called alyoKepoiS, a horned
goat, and according to one Greek legend was the
Cretan goat-nymph named Amaltheia who
reared Zeus in the cave of Dicte on the Aegean
Hill (Neugebauer OCD). Others (Hyginus, II,
28; Erat. Catast, 27) say it was Aegipan {aiylitav)
who had cared for Zeus, He was said to have
frightened the Titans by his braying. The lower
part of his body is shaped as a fish since he hurled
shellfish, instead of stones, against his enemy,
Mythographers also said he was Pan whom Egyp-
tian priests had said threw himself into a river
thereby making the lower part of his body a fish
while the rest of his body remained a goat. In
this way he escaped the monster Typhon, and
Zeus, admiring his cunning, placed him among
the constellations.
  In the Islamic world he was called al-jadi (the
goat), a name which he bears in the title written
at the end of his horns in Figure 72 and also as
the title of the House of Capricorn along the
ecliptic, Al-BTrQnT {Astrol., sec. 159) criticized the
use of this word for the constellation since it

literally means a kid or young goat, suggesting
instead the word tays (adult male goat). Nonethe-
less, the former is the usual title of the constel-
lation. In Figure 72 Capricorn is drawn with no
hind legs, as is customary, but his fish tail is not
distinct. In the manuscripts of al-Sufl and on
most globes he is drawn clearly with a fish or
mermaid-like tail that is a faithful reproduction
of the classical form, except for globe No. 1 (the
earliest), on which it is depicted as having four
legs and resembling a deer.
  The constellation of Capricorn occupies the
area between 295° and 327° of the ecliptic, with
its head in the House of Capricorn and its body
in the House of Aquarius, the line dividing the
two Houses passing through its cheek and for-
ward front leg. Three lunar mansions can be
seen in Figure 72: the twenty-third, bula", at
295y4° and in front of Capricorn's head; su"iid,
the twenty-fourth mansion, at 308° and in the
neck of the sea-goat; and on the rump at 321 '/4°
of the ecliptic, akhbiyah, the twenty-fifth lunar
mansion.
  No stars are labeled on the constellation on
the Smithsonian globe, Capricorn is comprised
of 28 formed stars.
  Two stars in its head, the one at the base of
the inside horn and the larger one below the
three on the horns, nos. 1 and 3, both considered
of less than third magnitude although they are
represented by different sizes of plugs on the
Smithsonian globe, gave their names to the
twenty-second lunar mansion. These two [a''^,9
Capricorni] were called sa"d al-dhdbih, the mean-
ing of which, as with all the sa"d lunar mansions,
has become so obscure with time that it is vir-
tually impossible to translate it. A very rough
translation might be "the omen of sacrifice," The
word dhdbih is engraved on the Smithsonian
globe as a lunar mansion along the ecliptic at
282y4° beside the neck of the constellation Sag-
ittarius in the House of Capricorn (see Lunar
Mansion 22 for a discussion of sa"d stars). Ac-
cording to al-SQfi {Suwar, 227) the Bedouins also
called the lowest of the three stars at the base of
the horns "the sheep {shdt) which is slaughtered"
[no, 2; V Capricorni].
  
182

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 72.—The Goat [Capricornus]. (Photo: Smithsonian no. 72-3116)

NUMBER 46

183



  The two brightest stars in the body of Capri-
corn, the two large ones very close together as if
a double star in the middle of the hind quarters
(nos. 23 and 24) were frequently labeled on
globes, in fact on all except the two earliest globes
(Nos. 1 and 34), on No. 26, and on the present
globe, the latter two from the Lahore workshop.
These two stars [yb Cap] were called sa"d nd-
shirah, which translates roughly as "omen of fer-
tility." The brightest of the two, considered to
be of third magnitude (the one closest to the tail;
b Cap), was called dhanab al-jadi (the tail of the
goat), a name it bore on astrolabes and on some
class B globes such as that made in 1065 H/AD
1655 by Hamid ibn Muhammad MuqTm of the
Lahore workshop (No. 68).
Constellation 32.    The Water-Pourer
[Aquarius]
FIGURE 73
  This zodiacal constellation is called by Aratus
vbpoxbos (the water-pourer), of which he says
(lines 282-298) that when the sun is in him "the
sailor should avoid the open sea, for fierce are
the south winds then and hard the frost." Later
mythographers (Hyginus, II, 29) identified him
with Deucalion (who reigned when the great
flood took place) or Ganymede (cupbearer to the
gods), but the simple "water-pourer" is retained
in the Ptolemaic catalog. In the Ptolemaic con-
stellation the flow of water streams into the
mouth of another constellation, Piscis Austrinus.
Figure 73 presents an overall view, while Figure
87 shows Aquarius in detail. On the Smithsonian
globe the water-pourer is one of four bearded
figures, and the title of the constellation is writ-
ten to the left of the head as siirat sdkib al-md"
(the form of the pourer of water), the same title
used by al-Sufi {Suwar, 231) who goes on to say
wa huwa al-dalw (that is to say, the bucket). The
word al-dalw (the bucket) represents the indige-
nous Bedouin image of the area, which derived
from a Sumerian source shared with the Greek
imagery. The traditional term dalw replaced the
Arabic translation of the Ptolemaic name as a

title for the House of Aquarius. Thus al-dalw is
the title for the House on the Smithsonian globe,
while the constellation bears the Ptolemaic title.
  The figure occupies the space along the eclip-
tic of about 328° to 345° with one arm stretched
backward over Capricorn to about 305°, The
division of the House of Aquarius from the
House of Pisces passes through the eastern arm
and lower leg of the constellation of Aquarius.
Two lunar mansions can be read along the eclip-
tic: at 321 ° akhbiyah, the twenty-fifth at the rear
of Capricorn; and at 334V2° muqaddam, the
twenty-sixth mansion, which is between the skirt
and the stream of water. On the Smithsonian
globe the constellation contains 41 formed stars.
According to all the star catalogs it should con-
tain 42 formed stars. The star that the maker of
the Smithsonian globe inadvertently omitted
should have been located in the stream of water
just below the vessel, and is no, 23 in the listings
[K Aquarii (?) or Flam, 78 (?)]. Three unformed
stars were also considered part of the constella-
tion. These can be seen near the second bend of
the stream of water, between the water and the
tip of the tail of the sea-monster Cetus, One of
the three stars, directly below the two that lie
close together is so tarnished it is difficult to
make it out; these three stars are today consid-
ered a part of the constellation Cetus,
  The only star labeled on the Smithsonian globe
is the large one at the end of the stream of water,
in the mouth of the Southern Fish. This star, a
Piscis Austrini, was considered by all catalogs at
that time to be no. 42 of the constellation of
Aquarius, although today it is numbered as part
of the constellation of the Southern Fish. The
star is labeled/am al-hut (mouth of the fish) from
which arises the modern term Fomalhaut for this
star, the eighteenth brightest in the heavens. Al-
SQfT {Suwar, 239) stated that this star was in the
Bedouin tradition also called al-zalim (the male
ostrich), adding that "the bright star which is in
the end of the river (Eridanus) is also called that;
and I have seen on some globes the word al-zalim
inscribed between these two stars." The image
of the ostriches is in keeping with those seen in
Sagittarius. Another traditional name for Fom-
  
184

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

N




FIGURE 73.—The Water-Pourer [Aquarius] is on the left side of the globe. See also Figure 87.
This view shows the area around the vernal equinox. (Photo: Smithsonian no. 72-3084)

alhaut was al-dafda" al-awwal (the first frog),
because as Al-SQfT {Suwar, 239) said, the bright
star in the tail of Cetus [fi Ceti] was called "the
second frog." These latter two stars were also
singled out by Aratus (lines 395-398), though
he did not call them frogs.
  
The four stars on the hand and wrist of Aquar-
ius, holding the water jug, made up the twenty-
fifth lunar mansion called sa"d al-akhbiyah [nos.
9-12; y-K^ri Aqr]. On the Smithsonian globe the
term akhbiyah is written as a lunar mansion along
the ecliptic. These four stars form the modern
  
NUMBER 46

185



asterism of the Y of Aquarius, also called The
Water Jar. The twenty-fourth lunar mansion was
derived from the two stars on the western shoul-
der of the Water-Pourer, nos. 4 and 5 [fi^ Aqr]
and no. 28 of Capricorn in the end of its tail, and
as such was called sa"d al-su"ud, which translates
roughly as "omen of good fortune." The three
stars in the hand over Capricorn, nos. 6-8, com-
prise the twenty-third lunar mansion [Flam. 7
and pe Aqr] called sa"d bula". The two stars on
the eastern shoulder of Aquarius, nos. 2 and 3,
were frequently called sa"d al-malik, roughly
translated as "the royal omen" [ao Aqr], a title
frequently used for these stars (see globes Nos. 3
and 5; see section on Lunar Mansions for a
discussion of the sa"d-stars).
Constellation 33.    The Fish
[Pisces]
FIGURE 74
  Again the concept of water dominates this area
of the zodiac as it had with the preceding sign,
Aquarius. The fishes, Ixdves, Aratus says (lines
239-247; [trns,] Mair, p, 227) are placed one
higher, more to the north, than the other and
thus it
louder hears the fresh rush of the north wind. From both
there stretch, as it were, chains {deana), whereby their tails
on either side are joined. The meeting chains are knit by a
single beautiful and great star, which is called the Knot of
the Tails. Let the left shoulder of Andromeda be the guide
to the northern Fish, for it is very near.
Mythographers (Hyginus, 11,30; Manilius, IV,
519-581) relate the story that Venus and her
son Cupid came to the river Euphrates in Syria
where the monster Typhon appeared. They both
threw themselves into the river and became
fishes. For this reason the Syrians stopped eating
fish for fear they would seem to oppose the
protection of the gods and so they would not
catch the gods themselves.
  The drawing of the constellation on the Smith-
sonian globe is another faithful rendering of the
classical form. The fishes are drawn in a playful
manner, especially the one nearest the water
poured by Aquarius and the neck of Pegasus,
which even has an eyebrow; they both appear to

have small ties about their necks. The division of
the House of Pisces from that of the House of
Aries (the vernal equinox) passes behind the tail
of the western fish in figure 74. The western fish
actually lies on the equator, which in Figure 74
is coming from the lower left and continues to
the upper right. The graduations of both the
ecliptic and equator begin their numbering at
the vernal equinox, but the numbering of the
celestial equator was repeated after the autumnal
equinox. Consequently the western fish lies along
the equator graduated from 160° to 180°, north
of the section of the ecliptic graduated 340° to
360°. The band and the eastern fish pass
through the first 30° of the ecliptic. Four lunar
mansions can be seen: at 347'A° between the
water poured by Aquarius and the western fish
is mu"akhkhar, the twenty-seventh mansion; at
the vernal equinox by the tip of the tail of the
western fish is rishd", the twenty-eighth. At 13°
of the ecliptic near the first turn of the connect-
ing band there reads sharatayn, the first lunar
mansion, and at 25'/2°, butayn, the second man-
sion.
  The constellation bears the title surat al-hut
(the constellation of the fish), engraved under
the head of the western fish. This is the common
title for both the constellation and zodiacal house
and is the traditional Bedouin term for the region
which replaced the term al-samakatdn (the two
fish), which had been used to translate the Pto-
lemaic title. Clearly the Bedouin image was of
one large fish in this area, no doubt deriving from
a Babylonian image which was also the source
for the two fishes of the Greek world. According
to the star catalogs the constellation consists of
34 formed stars. In Figure 74, 35 stars can be
seen, there being for some unknown reason an
extra star in the band connecting the fish, in the
second twist of the band from the eastern fish
(indicated by an arrow). The star at the end of
the tail of the eastern fish (near Andromeda) is
usually placed in the band; the spacing of several
stars is not very precise. The star at Andromeda's
underarm is to be counted with Pisces, The con-
stellation also has four unformed stars, two of
which are in the ecliptic beneath the western fish
and two just south of the ecliptic near the tip of
  
186

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 74.—The Fish [Pisces]. (Photo: Smithsonian no. 72-3113)

NUMBER 46

187



the tail of Cetus which can be seen in the lower
righthand corner of Figure 74.
  No star in Pisces is labeled on the Smithsonian
globe. There is written across the chest of An-
dromeda batn al-hUt (the belly of the fish). This
term refers to the star on Andromeda's waist
closest to Pisces [fi Andromedae], and arose from
a Bedouin conception of a large fish lying across
the area of Andromeda. This star gives its name
to the twenty-eighth lunar mansion, and as such
was also called in the anwd" tradition al-rishd"
(the rope). The latter term represents the lunar
mansion on the Smithsonian globe, written at the
vernal equinox. The prominent star mentioned
by Aratus as the Knot of the Tails is probably
the third in the band between the fishes, counting
from the eastern fish. This star [no. 19; a Piscium]
was called by al-SufT {Suwar, 252) "aqd al-khaytayn
(the knot of the two ties) from the Ptolemaic
image. Though it is the brightest star of the
constellation, it does not seem to have had as
important a role in Bedouin terminology, for it
bears no traditional Arab name and is not labeled
on any globes. The modern name for the star is
El Rischa, from the name of Lunar Mansion 28.
Arab astronomers and writers of anwd" litera-
ture, however, did not use this term for a Pis-
cium, but only for fi Andromedae, where it does
not mean a knot, but a rope, arising from the
conception of a bucket in nearby Pegasus.
Constellation 34. Qitus
[Cetus]
FIGURE 75
  The southern constellations (those south of the
ecliptic) represent for the most part signs of
darkness, evil spirits, and toil that populate the
vast ocean or deep into which the heavenly bod-
ies sink and through which sailors must struggle.
The first southern constellation was called by
Aratus (lines 354-359) "the mighty sea-mon-
ster," "the hated sea-monster," and "the dark-
blue Cetus," Cetus being the Latin transliteration
of the Greek KfjTos. Aristotle {PA, 669a) used the
word for any animal of the whale kind, "whales.

the dolphin and all spouting cetacea {avaipvaoiVTa
KT]T-q)." Mythographers (e.g., Hyginus, 11,31) as-
sociated the sea-monster with the monster sent
by Poseidon to kill Andromeda. The constella-
tions of Andromeda and Perseus who saved her
are due north of the constellation Cetus, whose
head is just south of the Ram, Aries. Thus his
head is north of the equator.
  In the Islamic world the Ptolemaic constella-
tion was called al-qitus, a transliteration of the
Greek name. In the Greek world the constella-
tion was most frequently seen as a whale, while
in the Islamic world it was often depicted like a
"senmurvs," an oriental winged dragon, and al-
most always shown wearing a collar (Wellesz
[1959], 18). In Figure 75 it has a snarling dog's
head, bird's feet, and a feathered fish tail, pro-
ducing an effect more fantastic than sinister. The
constellation consists of 22 formed stars, includ-
ing the one in the hair, next to the leg of Aries,
and also the four on the chest, which are shared
with the River (Eridanus). The star names in
Figure 75 all reflect the Ptolemaic schema. The
large star in the jaw is labeled fam al-qitus (the
mouth of Cetus [no. 3; 7 Ceti]). The star on the
northern tip of the tail, closest to the equator, is
called dhanab al-qitus shamdli (the northern tail
of Cetus [no. 21; i Get]). The title of the star on
the southern part of the tail is dhanab al-qitus
janubi (the southern tail of Cetus [no. 22; fi
Cet'\). One name often given to the latter star
today is Deneb Kaitos, derived from this Arabic
dtle.
  Another modern name given the latter star
today, Diphda, comes from the Arabic name
given the star in the Bedouin tradition: al-dafda"
al-thdni (the second frog). The first frog is the
large star in the mouth of the Southern Fish [a
Piscis Austrini; Fomalhaut], which is west of Ce-
tus. These two stars were also singled out by
Aratus (lines 395-398), though he did not call
them frogs.
  In keeping with the notion of ostriches in the
area of the Southern Fish and Sagittarius, the
stars nos, 12-16 (the five in the trunk of Cetus
TV^Bri Get] were traditionally called al-na"dm and
al-na"dmdt, both meaning ostriches (Kunitzsch
  
SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

FIGURE 75.—QTtus [Cetus]. (Photo: Smithsonian no. 72-3101)

NUMBER 46

189



[1961], nos. 181, 184), While not marked on the
Smithsonian globe, these stars are labeled in this
manner on globes Nos, 3, 4, and 5,
  In the Bedouin tradition, the first six stars of
the constellation [Xayb^p Get], in the head and
the neck, not including the one in the hair, were
collectively called al-kaff al-jadhmd" (the cut-off
hand [of al-thurayyd]). This group of stars was
seen as part of the human form of which the
Pleiades (called al-thurayyd) formed the center
(see the discussion of Lunar Mansion 3 for details
on this early asterism).
  The four on the chest, shared with the River
Eridanus, were called batn (the belly [nos. 8-11;
paew Get]). On globes No. 1 and No. 34, the two
earliest, they bear this label. The four stars that
are close together in the root of the tail [nos, 17-
20;V'^ 0,198,V'', 0,161 Get] al-Sufi {Suwar, 269)
remarks he has seen labeled on some globes as
al-nizdm (the string of pearls), but that he has not
been able to find that in any of the literature,
(Kunitzsch [1961], no. 206 says the engraver of
the globe seen by al-SQfi simply made an error
and meant to inscribe al-na"dm or "ostriches.")
Al-SQfT also adds that it has been said that the
set of stars forming Cetus was called al-baqar (the
cows), but that is said by Arabs "who have no
knowledge of the stars" (see Kunitzsch [1961],
no. 58a),
  Although the names of Bedouin origin do not
appear on the Smithsonian globe, they do all
appear on one or another of the globes prior to
the seventeenth century, with the exception of
"the string of pearls" and "the cows," which occur
on none of the globes examined.
Constellation 35.    The Giant
[Orion]
FIGURES 76, 78
  The second southern constellation is Orion,
who for the most part lies north of the celestial
equator. The equator passes through the three
stars forming the well-known belt of Orion, Or-
ion was one of the five star groups named by
Homer and was perhaps the first constellation to

which a Greek legend was assigned. Traditionally
he was a great hunter, eventually killed by Ar-
temis, according to Homer {Od., V, 121-124),
or as another tradition has it, by a scorpion
(Scorpio) sent by her. Aratus (lines 323-325)
does not describe the figure of Orion {ooplcov),
but merely says "Let none who pass him spread
out on high on a cloudless night imagine that,
gazing on the heavens, one shall see other stars
more fair" ([trns.] Mair, 233). Ptolemy described
him as he was usually depicted in Greek art:
holding in his right hand a Xayo^fio'Xov, a staff
for flinging at hares (although also used for a
shepherd's staff or crook); a short bent sword or
dagger {paxocipa) hung from his waist; and a
lion's skin {bopa) was draped over his left arm.
He was also depicted as kneeling in a manner
similar to Hercules (see Figure 54).
  Al-SQfT {Suwar, 264) describes the constella-
tion as "resembling somewhat the form of a man
having a head, two upper-arms and two feet; and
he is called al-jabbdr (the giant) because he has
two thrones [a reference to groups of stars in
Lepus and Eridanus] and in his hand there is a
stick and around his middle a sword." He uses
simply the word al-"asd to translate XayojfibXov in
the descriptive text, but does amplify it further
in the catalog, calling it al-"asd dhdt al-kildb (a
stick used for [wild] dogs). On the Smithsonian
globe it is difficult to tell what the weapon is
intended to be. The most radical change that al-
SQfi made was to describe in the text the nine
stars that, according to Ptolemy, were on the
lion's skin as being on a sleeve, kumm. Thereafter
in the Islamic world Orion was drawn with an
elongated sleeve on his left arm, with the excep-
tion of globe No. 1, where he appears to hold
some long object in his left hand, which contains
nine stars. Surprisingly, in his catalog al-SQfT does
not mention the sleeve, but adhers strictly to the
Almagest, stating that the nine stars are "in the
skin {al-jild) worn on the left arm."
  In Figure 76 Orion can be seen between and
below Taurus and Gemini (for a detailed view
see Figure 78), The constellation is titled under
the eastern arm surat al-jabbdr (the constellation
of the giant). The traditional Bedouin term for
  
190

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 76.—The Giant [Orion] (see also Figure 78). This view also shows the area around the
summer solstice. (Photo: Smithsonian no. 72-3079)

the region was al-jawzd", which comes from a
root meaning "to marry"; its precise meaning is
obscure and the word may have arisen through
a complex philological error (Kunitzsch [1961],
24; Ideler, 214-218). In any case the v/ordjawzd"
seems to have referred to a large, possibly femi-

nine, figure (see section on Classical Greek and
Pre-Islamic Sources, for a discussion of this Be-
douin image). The name al-jawzd" appears as the
title of the constellation of Orion on the two
earliest globes (Nos. 1 and 34) and is a compo-
nent of many of the individual star names. On

NUMBER 46

191



the Smithsonian globe the name al-jawzd" is given
to the constellation and House of Gemini, which
was also in the area covered by the ancient al-
jawzd". The very large star on the eastern shoul-
der on the Smithsonian globe is given two names:
mankib al-jawza" (shoulder of al-jawzd") and yad
al-jawzd" (arm of al-jawzd"). The star referred to
is no, 2 [a Orionis], a variable star that is the
twelfth brightest star of the heavens. The mod-
ern name of the star, Betelgeuse, arose from
another name for the star, not on the Smithson-
ian globe, ibt al-jawzd" (armpit of al-jawzd")
  The larger of the two stars next to each other
on the west shoulder is labeled al-mirzam [no. 3;
7 Ori, Belletrix]. The term al-mirzam is a tradi-
tional term applied to three stars in the heavens,
apparently in the sense of companion, since in all
three instances there is a larger star with which
it is paired. This pairing is most evident in Canis
Major and Canis Minor, where it is applied to
companions of Sirius and Procyon (see section
on Classical Greek and Pre-Islamic Sources). The
use of mirzam for a star in Orion may be of later
origin, and was not used consistently (Kunitzsch
[1961], no. 166), Al-SQfi {Suwar, 269) says that
some people called the bright red star [a Ori] by
the name mirzam al-jawza", but that it is incorrect
and that the term mirzam properly belongs to the
third star of the constellation [7 Ori] which pre-
cedes the brighter star, like the two mirzams of
the two Sirii (a reference to Sirius and Procyon).
This third star [7 Ori] is also labeled on the
Smithsonian globe al-ndjidh (? the fortunate one),
another obscure Bedouin term.
  The large star on the western foot of Orion,
which is shared with the River Eridanus into
which the figure is stepping, is labeled rijl al-
jawzd" al-yasri (the left foot of al-jawzd"). This
star [no. 35; fi Ori, Rigel] is the seventh brightest
of the heavens; it was also called by al-SQfT (5M-
war, 269) rd"i al-jawzd" (the shepherd of al-
jawzd").
  The star on the knee on which he is resting,
no. 38, is termed rijl al-jawzd" al-yami (the right
foot of al-jawzd"). The modern name of the star,
Saiph [K Ori], is derived from the name given by
al-SufT {Suwar, 269) to the three stars [nos. 30-

32; c ^''^i Ori] composing the dagger, sayf al-
jabbdr (the sword of the giant), a name clearly
derived after the introduction of the Ptolemaic
constellations. The Bedouin term for these three
stars was al-jawdzi {Y>.un\tzsch [1961], no, 105).
Two of the stars, 0''^ i Ori, comprise the asterism
known today as the Sword of Orion.
  The three stars forming the famous asterism
of the Belt of Orion, be^Ori, were called in Greek
XOipires (the [three] Graces). In the Islamic world
nos. 26-28 were called mintaqat al-jawzd" and
nitdq al-jawzd", both meaning "the belt of al-
jawzd". " The former term is the one from which
the modern name of 5 Ori, Mintaku, is derived.
These three stars were also (al-SQfT Suwar, 269)
called al-nazm and al-nizdm (a string of pearls) as
well faqdr al-jawzd" (the vertebrae of al-jawzd").
These stars are not labeled on the Smithsonian
globe. The traditional Arabic terms for the nine
stars on the elongated sleeve [y^y'o^x'"® Ori],
nos. 17-25, were tdj al-jawzd" (the crown of al-
jawzd") and dhawd"ib al-jawzd" (the hanging locks
of al-jawzd"). These imply a different positioning
of the figure than that of the later Ptolemaic
form. These titles do not appear on the Smith-
sonian globe.
  The star marked on the head of Orion is given
as no. 1 in the catalogs. Al-SQfT {Suwar, 264) says
of it:
The first star which is on the head is nebulous {sahdbi) but it
is three small stars next to one another in the form of a small
triangle. Ptolemy gave to the middle of the triangle the rank
of a star and gave it longitude and latitude in the book
[Almagest].
There is indeed luminosity in the area of these
three stars [XV'V^ Ori], which both Ptolemy and
al-SQfT called nebulous. Al-SQfT in his catalog
counted the three stars as one, following Pto-
lemy, and Ulugh Beg maintained the same des-
ignation. On some globes (e.g., Nos. 4 and 5),
three stars are indicated in the head instead of
one as on the Smithsonian globe. The Bedouins
traditionally called these three small stars al-
haq"ah, the name of the fifth lunar mansion (see
section on Lunar Mansions). Al-SQfi adds that
they also were called al-athdfi (the three-legged

192

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



support of a cooking pot) since they resembled
it.
  The constellation of Orion is comprised of 38
formed stars and no unformed stars. In Figure
78 what appears to be one star on the eastern
side of the chest is actually two contiguous stars.
They are not usually placed so close to one
another. In the hand overhead there are only
four stars and not five as appears at first glance.
The great Nebula of Orion, M42, does not seem
to have been recorded at this time.
Constellation 36.    The River
[Eridanus]
FIGURE 77
  Aratus (lines 360-366) was the first to name
this winding constellation that has the form of a
river, speaking of the poor remains of Eridanus
{■qptbavos) as a river swollen with tears {TTOXVKX-
avTos TTorapbs). The legend referred to is that of
the river Po, into which Phaethon fell when
struck by Zeus's thunderbolt and where his sisters
the Heliades wept for him. The river Eridanus
was partly burnt up, hence the "poor remains."
Eudoxus had called it simply "river," and later
writers associated it with the Rhone or the Rhine
and even the Nile and Oceanus (Hyginus, 11,32).
The river twists from the forward foot of Orion
over to Cetus the Sea-Monster, where it shares
four stars; it turns toward Lepus the Hare and
then back under the feet of Cetus. In Figure 77
the River has the appearance of a snake, biting
at Orion's foot; actually Orion is supposed to be
stepping into the River. A very large round plug
on the globe can be seen across the southern
segment of the river, indicated by arrows (see
Chapter 2 for details regarding plugs).
  In Arabic the constellation is titled surat al-
nahr (the constellation of the river), written in
Figure 77 at the first bend. The only star labeled
in Figure 77 is the one at the end of the river,
beneath the feet of Cetus. This star, no. 34, is
labeled akhir al-nahr (the end of the river). This
is not the star bearing today the name Achernar
[a Eridani], the ninth brightest star of the heav-

ens, for the latter star has a declination (meas-
ured from the equator) of -57°29' while the
star called akhir al-nahr, no. 34 of the constella-
tion, has according to al-Sufi a latitude of
-53°30' measured from the ecliptic. This last
star of the constellation must be 6 Eridani, a
double star resolvable with binoculars, today
called Acamar and having a declination of
—40°3r from the equator. Thus the classical
constellation did not extend as far south as does
the modern outline of the asterism.
  The constellation is composed of 34 formed
stars, not including the 4 shared with Cetus.
There are no unformed stars, Al-Sufi {Suwar,
277-278) reports that in the Bedouin tradition
the thirty-fourth star of the constellation was
called al-zalim (the male ostrich) and that be-
tween this ostrich and the ostrich at the mouth
of the Southern Fish [a Piscis Austrini] are many
stars called al-riydl; which he further explains
with the phrase/iroM al-na"dm (baby ostriches).
  The Arabs traditionally called the first three
stars of Eridanus (the three in the head [Xfiyp
Eri]), along with no, 36 of Orion [r Orionis] (the
one above the very large one on the foot of
Orion) kursi al-jawzd" al-muqaddam (the anterior
throne of al-jawzd") because, according to al-SufT
{Suwar, 277) they are in a square resembling a
throne. The thirty-fifth star of Orion, also on
the foot [fi Orionis], was said to be alongside the
square and resemble the foot {rijl) on the throne.
  The Bedouins (al-SQfi Suwar, 277) also called
nos. 14-20 of the River (the five stars to the
south of the chest of Cetus plus the first two of
those shared with Cetus) udhi al-na"dm (the
hatching place in the sand of ostriches). Al-SQfi
expands his explanation: "that is to say, its nest
and place of its eggs . . , those which are around
these stars are called the eggs and the egg-shells"
[rp^\, W.B.2'788,T'-' Eri plus eir Ceti; cf Kun-
itzsch [1961], no. 308]. Al-SQfi further remarks
{Suwar, 278):
I have seen in Shiraz many stars near the horizon resembling
a skiff {zawraq) in which there is a bright star of third
magnitude which, along with the one on the south branch
of the tail of Cetus, i.e. the second frog, and the star in the
mouth of the fish, is in a triangle. In the spaces are stars of

NUMBER 46

193


FIGURE 77.—The River [Eridanus]. Arrows indicate plugs. (Photo: Smithsonian no. 72-3110)

194

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



fourth, fifth, and sixth magnitudes, all of which are called
al-riydl (baby ostriches) of which nothing is said by Ptolemy.
Constellation 37.    The Hare
[Lepus]
FIGURE 78
  Beneath the feet of Orion is the hare (\a7aj6s
or Xaycos in Greek), continually pursued by the
hunter (Orion) and his dog (Sirius) which is to
the east of Lepus (Aratus, line 338), The
mythographers (Hyginus, 11,33; Erat, Catast., 34)
said that Hermes placed the hare among the
constellations because of its swiftness. Hyginus
gives the account of the extreme multiplication
of the hares on the Island of Leros, after which,
he says, the constellation was placed in the heav-
ens so men could remember "that nothing is so
desirable in life but that later they might expe-
rience more grief than pleasure from it" (trns.
Grant).
  The Ptolemaic name of the constellation was
translated literally into Arabic as surat al-arnab
(the constellation of the hare). In Figure 78 the
rabbit resembles a dog more than a hare. The
expression "arsh al-jawzd" (the seat of al-jawzd")
is written twice near the hare in Figure 78, once
below and once above. These terms refer to four
stars of Lepus, the two on the forepart of the
body and the two on the hind feet, nos. 7-10.
Al-SQfi {Suwar, 238) states that these four stars
[afiyb Leporis] were called kursi al-jawzd" al-
mu"akhkhar and "arsh al-jawzd" (the posterior
throne or seat of al-jawzd") "because they are in
between the two legs of al-jawzd" [a figure in
roughly the area of Orion] in the location of the
throne." The adjective "posterior" is used to
distinguish this throne of al-jawzd" from another
one in the constellation of Eridanus. In Figure
78 the letters m-w are written before the expres-
sion "arsh al-jawzd" inscribed under Lepus. This
is probably an abbreviation of al-mu"akhkhar (the
posterior). Similarly what looks like l-w before
the same label engraved over the back must also
be intended as the abbreviation m-w. The iden-
tical letters occur on globe No. 11, which was
made by the same maker who produced the

anonymous Smithsonian globe. Of these four
stars al-Sufi {Suwar, 283) also adds "and 1 have
read in some books of the anwd" that they are
called al-nihdl." The latter word means men or
beasts (usually camels) going to drink water in
order to quench their thirst, and from this name
the "modern" name of the lower star in the
forepart of the body, Nihal, {fi Leporis), was de-
rived. The modern name of the other star on the
front shoulder, a Leporis, is Arneb or Alarneb,
from the Arabic name for the constellation. The
constellation is composed of 12 formed stars and
no unformed ones.
Constellation 38.    The Greater Dog
[Canis Major]
FIGURE 79
  The brightest star in this constellation, which
is the brightest in all the sky, was mentioned by
Homer (//, V,5-6; XXll,26-30) and called by
the name aelpios (the scorcher). The later Latin
name was Sirius. It was said of the star that it was
"a baleful sign {KaiKov arjpa) for it brings to
suffering mortals much fiery heat," for when it
rises with the sun it marks the season of greatest
heat. Homer mentioned also that it was the star
men call the Dog of Orion. Hesiod {Works and
Days, 584-588) speaks of Sirius when he says:
In the exhausting season of summer; then is when goats are
at their fattest, when the wine tastes best, women are most
lascivious, but the men's strength fails them most, for the
star Sirius shrivels them, knees and heads alike, and the skin
is all dried out in the heat (trns. Lattimore, pp. 87, 89).
By the time of Aratus this star "that keenest of
all blazes with a searing flame" (lines 330-331)
was seen as part of a constellation and marked
the jaw of the Dog, which was the faithful hound
of the Hunter Orion. The star is still called the
dog-star and the days of greatest heat the dog-
days.
  In the Islamic world the Ptolemaic constella-
tion was called al-kalb al-akbar (the Greater Dog).
Al-Sufi {Suwar, 285) says of it, "it is a constella-
tion of a dog behind the constellation of al-jawzd"
(a figure in the area of Orion) and the constel-
lation of the Hare, for which reason it is called
  
NUMBER 46

195


FIGURE 78.—The Hare [Lepus]. (Photo: Smithsonian no. 72-3107)

196

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 79.—The Greater Dog [Canis Major]. (Photo: Smithsonian no. 72-3108)

NUMBER 46

197



kalb al-jabbdr (the dog of the giant, Orion). In
his catalog, al-Sufi says that the brilliant star in
the jaw, Sirius, is called al-kalb (the dog), follow-
ing Ptolemy. In his discourse on the constellation
al-Sufi {Suwar, 288) presents the traditional Be-
douin imagery for this region in which two Sirii
were seen. The traditional Arabic word used for
a Canis Majoris was shi"rd, which apparently
comes from a source, probably Babylonian, that
it shared in common with the Greek term aelpios
(see Hommel, 597-598, 619; Kunitzsch [1959],
117, 118, notes 1, 3). In Arabic the word has
little meaning that seems appropriate, except
perhaps as interpreted as "to have a distinctive
mark," i.e. to be distinctive, or possibly shaggy,
unkemp hair, which Drechsler (13) suggests
could apply to the radiant appearance of Sirius
(see section on Classical Greek and Pre-Islamic
Sources for the traditional legend of the two
Sirii, Sirius, the southern shi"rd, and Procyon,
the northern shi"rd in the Lesser Dog, who were
sisters of Canopus, suhayl, who had married al-
jawzd", and the related terminology). In Figure
79 Sirius is labeled si"rd yamdni. This is no doubt
an error (occurring also on globe No. 11), both
in agreement of the adjective and diacritical
points, and ought to read shi"rd yamdniyah (the
southern shi"rd). The reading on the Smithson-
ian globe for this star is from the root s"r, which
means to burn or be in flames. Thus it appears
to be more in alignment with the Greek aelpios,
meaning "scorcher."
  The large star on the upper front paw, no. 9,
is labeled mirzam and is fi Canis Majoris (Almir-
zam or Murzim). This is one of three stars on the
globe bearing the label mirzam, used in the sense
of companion, the other two being in Orion and
Canis Minor (see section on Classical Greek and
Pre-Islamic Sources).
  Al-Sufi {Suwar, 289) presents some additional
Bedouin star names for this region, which also
antedate the introduction of Ptolemaic astron-
omy. The two stars on the rump, the one on the
upper thigh, and the one on the tail [nos.
12,14,15,18; o^ber] CMa] were together called al-
"adhdra (the virgins) and also *" udhrat al-jawzd"
(the hymen of al-jawzd"), al-jawzd" being a large

feminine figure seen in the area occupied by the
Ptolemaic Orion, The Arabs also called the star
on the end of the rear foot (no, 17) and all of the
10 nearby external stars except the two brightest
(no. 9 and 10), al-qurud (the baboons) and also
al-aghrub (the crows; cf. Ideler, 248),
  The constellation of Canis Major is comprised
of 18 formed stars and 11 unformed ones. In
Figure 79 there is one formed star missing (no,
8), which should be next to the small star on the
leg immediately under the chin. Of the 11 un-
formed stars, 10 are to be seen below and to the
southwest of the dog. The eleventh external star,
actually no. 1 of the unformed stars in the cata-
logs, is north of the head of the dog on the other
side of the equator at 105° along it, and lies
between Canis Major and Canis Minor. The star
cannot be seen in Figure 79, but is visible in
Figure 80 of Canis Minor.
Constellation 39.    The Lesser Dog
[Canis Minor]
FIGURE 80
  The constellation of Canis Minor consists of
only two formed stars and is placed beneath
Gemini and Cancer, which can be seen to the
north. The dog lying north of the celestial equa-
tor, like the greater dog that is directly south of
the equator, follows the constellation of Orion
and was considered one of the hunter's compan-
ions. In the Homeric literature neither this con-
stellation nor either of its stars is mentioned,
while Aratus (line 450) refers only to the larger
of the two stars, calling it irpoKvoiv, which shines
brightly beneath the twins; the name literally
means "before the dog" and refers to the fact
that this star rises before the dog-star Sirius in
Canis Major. The Greek word is the source of
the modern name of the star Procyon. By the
time of Ptolemy the star had, with the nearby
star, become part of a recognized constellation,
itself having the form of a dog.
  In the Islamic world the constellation was com-
monly given the name al-kalb al-asghar (the lesser
dog) from the Ptolemaic conception of the con-
  
198

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 80.—The Lesser Dog [Canis Minor]. (Photo: Smithsonian no. 72-8591)

NUMBER 46

199



stellation. In Figure 80 it is labeled in front of
the figure siirat kalb saghir, which is poor gram-
mar for "the constellation of the small dog."
  The Arabic terminology for the two stars re-
flects the earlier traditional Bedouin ideas of the
heavens. The larger star, Procyon, the eighth
brightest of the skys, is labeled in Figure 80 si"rd
shdmi, which ought probably to be corrected to
al-shi" rd al-shdmiyah (the northern Sirius). Al-
SQfi {Suwar, 293) says it is called "the northern"
because it sets in the northern half The "south-
ern Sirius" is the Sirius of Canis Major (for a
discussion of the meaning of the word shi"rd see
Constellation 38, and for the legend of the two
Sirii who were considered sisters of Canopus see
the section on Classical Greek and Pre-Islamic
Sources),
  The smaller star in the neck of the dog is
labeled mirzam [fi Canis Minoris]. This is one of
three stars termed mirzam, probably meaning
companion, the other two being in Orion and
Canis Major (see p. 117 for a discussion of how
this term relates to the Bedouin legends given
this region of the heavens).
  Al-Sufi {Suwar, 293) adds that the Arabs called
the two stars together dhird" al-asad al-maqbudah
(the drawn up foreleg of the lion). He continues:
And they call it maqbudah because of its remaining behind
[rising after] the other forearm of the two stars which are
along the head of the Twins; and it is in a line with the one
on the shoulder of Orion [has the same celestial latitude].
He adds that many say the two stars form the
seventh lunar mansion, but that actually the sev-
enth lunar mansion is composed of the two stars
on the head of Gemini (see the section on Lunar
Mansions).
Constellation 40. The Ship
[Argo Navis]
FIGURE 81
  Aratus (lines 342-352) described the constel-
lation as being the ship Argo belonging to the
legendary Jason, who sailed in it with the Argo-
nauts from lolchis in Thessaly to Colchis at the
farthest end of the Black Sea to seek for the

Golden Fleece; when the voyage was over the
ship was placed in the heavens by Athena, The
constellation of Apyco sails with its stern foremost
or backwards, Aratus says:
For not hers is the proper course of a ship in motion, but
she is borne backwards, reversed even as real ships, when
already the sailors turn the stern to the land as they enter
the haven, and every one backpaddles the ship, but she
rushing sternward lays hold of the shore ([trns.] Mair, 235).
  The high-curving stern, which lies very close
to the tail of Canis Major, was called by Ptolemy
XV^'i-o^Kos, the usual term for such sterns of Greek
ships, since they frequently turned up like a
goose's neck {xw)- The ship had one mast and
two rudders or long steering oars. The prow of
early Greek ships of war was formed by a low
beak or ram, epfioXos, which projected from the
keel and which was driven into hostile ships. It
was not until later that the prow rose as high as
the stern. Thus the older style ship may have
given rise to the notion of its being a ship cut in
half, as Ptolemy seems to indicate in his place-
ment of the stars in his catalog when he speaks
of the cutting off {airoropr]) of the deck (see
Chapter 4 of this study for a comparison of the
ship on the Smithsonian globe with that on other
globes).
  In the Islamic world this constellation was
called surat al-safinah (the constellation of the
ship), as it is written between the flag on the stern
and the mast in Figure 81. There is some evi-
dence (Kunitzsch [1961], no. 259) that the early
Arabs viewed a ship in another area. However,
the only accounts of it place the ship from the
area under the stars forming al-dalw (the bucket
[byfia Pegasi]) to sa"d al-su"ud [fi^ Aquarii and c'
Capricorni] with the bow on "the anterior frog"
[a Piscis Austrini] and its stern on "the following
frog" [fi Ceti]; this appears an impossible arrange-
ment and is nowhere near the Ptolemaic Argo.
Al-SQfi {Suwar, 303) in his discussion of Argo
reports this tradition but dismisses it by saying
"but those who say this knew neither al-safinah
(the ship) nor al-su"ud nor the 2 frogs. But Allah
is wisest and knows best." The constellation of
Argo was considered to have 45 formed stars
and no unformed ones. It was mentioned as early
  
200

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 81.—The Ship [Argo Navis]. (Photo: Smithsonian no. 72-3102)

NUMBER 46

201



as Aratus as well as by al-SQfi that the Milky Way
passed through part of the stern of the ship. The
area covered by the classical constellation of
Argo Navis is today most frequently divided into
four constellations: Carina (the keel), Puppis (the
stern), Vela (the sail), and Pyxis (the mariner's
compass).
  The large star in the lower rudder is the sec-
ond brightest star in the heavens [no. 44; a
Carinae]. Ptolemy gave it the name Kav(x)fios, as
had his predecessors Hipparchus and Eratos-
thenes. The word Kav(x)fios, which the modern
name Canopus transliterates, was given to the
star, according to legend, from the name of the
chief pilot of the fleet of Menelaus, who while
returning from the Trojan War stopped at
Egypt. The pilot died a few miles from Alexan-
dria, where a monument was built in his honor,
a city named for him, and this star named after
him. Eratosthenes also called the star ireplyeios
(along the earth). The Arabic word for Canopus,
as can be seen in figure 81, is suhayl, which comes
from a root having to do with a flat ground or a
plain. It should be remembered that the star is
very noticeable, being the second brightest in the
heavens, and can be seen close to the horizon in
the lower latitudes (it is not visible north of the
thirty-seventh parallel), thus perhaps giving rise
to a term serving as a common source for all the
names associated with the earth that have been
given to it (see Schoy; Kunitzsch [1959], 208-
210 for Babylonian origins of the word Suhayl).
  Canopus is the only star labeled in Figure 81.
Al-SQfi {Suwar, 301) says that the traditions dif-
fer with regard to it, for some say it is called only
suhayl (i.e., without an adjective, "aid al-itldq)
while the stars of second magnitude near it—the
star in the center of the oar-top, the star at the
base of the easternmost flag, and the large star
to the east of the oar-top above the hull line [nos.
17, 31, 35; ^ Puppis, Xy Velorum]—are called
suhayl bulqin (?), suhayl haddr (an earthy suhayl),
suhayl al-wazn (the suhayl of the weight), and
suhayl al-muhlifwa al-muhnith (?). The meanings
of all these names are obscure (Kunitzsch
[1961], nos. 272,273). Another indigenous Arab
word for Canopus used by other authors (Kun-

itzsch [1961], no. 84) is al-fahl (the male camel).
Ibn Qutaybah cites a verse using this term and
adds afterwards that the poet "compared it to
a male camel who is no longer able to mate
and has isolated himself [from the rest of the
camels]" (Kunitzsch [1961], no. 84).
Constellation 41.  The Serpent
[Hydra]
FIGURE 82
  Aratus (lines 443-448) calls this constellation
vbpa, a hydra or water-serpent, whose head is
beneath the middle of the crab, its coil below the
lion (Leo), and its tail above the centaur. About
midway in its form is set the constellation of the
jar (Crater) and close to the end of the tail is the
form of a raven (Corvus) which appears to peck
at the Hydra. The Islamicate representation fol-
lows the Ptolemaic; a detailed view of the lower
half of Hydra with the constellations of Crater
and Corvus can be seen in Figure 83. This aster-
ism of a serpent with a bowl and raven may
possibly be of Babylonian origin and it has been
suggested that the serpent and the raven are
symbols of darkness and evil, with the huge jar
symbolizing the "vault of heaven wherein at times
storm, wind, clouds and rain are chaotically
mixed" (Brown, 1:107). Greek mythographers
(e.g., Hyginus, II, 40) identified the Hydra as
being the Lernaean hydra of the Hercules leg-
end. The raven was said to be a servant of Apollo
who was sent to fetch a cup of pure water for a
sacrifice; the raven encountered a fig-tree en
route and waited for the figs to ripen; when he
returned to Apollo he brought a snake saying
the snake had prevented him from getting the
water from the spring. Apollo punished the bird
so that when the figs are ripe, it cannot drink
water. The constellations were placed, according
to Hyginus, so that the crow seems to peck at the
serpent to be allowed to go over to the bowl of
water.
  In Arabic Hydra is called al-shujd" "the large
snake or serpent," as in the title written on the
Smithsonian globe beneath the coil of the snake.
  
202

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

N


w

FIGURE 82.—The Serpent [Hydra]. This view shows the area around the autumnal equinox.
(Photo: Smithsonian no. 72-3083)
The constellation is composed of 25 formed stars	130° and is beneath the head of Hydra. In Figure
and 2 unformed ones. One of the latter is located	83, which shows the lower half of Hydra, the star
to the north of the equator near the top of the	in the tail of the raven (Corvus) is misplaced, for
coil at 150° of the equator. The other unformed	it should be in the tail of Hydra as no, 24 of the
star is in the graduations of the equator at about	constellation. There are two stars named in Hy-

NUMBER 46

203



dra on the Smithsonian globe. The first one, the
anterior of the two on the nose of the serpent,
reflects the Ptolemaic outline. This star, no. 1 of
the constellation, is labeled mankhar al-shujd"
(the nose of the serpent [a Hydrae]).
  The other labeled star is the larger of the two
contiguous stars beneath the coil, no. 12 [a Hy-
drae]. This star is labeled/arc? al-shujd" (the iso-
lated one of the Serpent) and it is from this name
that the modern name of a Hydrae, Alphard,
arises. Al-Sufi {Suwar, 309) also calls this star,
which he says is written on astrolabes, ^ unq al-
shujd" (the neck of the serpent), a name it bears
on certain other globes (e.g., Nos. 4, 26, and 31).
Fard (the solitary) is a traditional Arab name for
the star, according to al-SQfi {Suwar, 312), "be-
cause of its seclusion from things similar to it and
its turning toward the south."
  In the Bedouin tradition the stars of Hydra
between fard al-shujd" and the stars of Corvus
the Raven, that is nos. 13 (the first he\ow fard)
through 23 (in the triangle to the south of Cor-
vus), are called al-shardsif (the cartilages of the
ribs). Kunitzsch ([1961], no. 284) interprets the
word to mean "the fettered camels" and identi-
fies the stars as KV^'^P^V Hydrae, fi Crateris and
x'^o/3 Hydrae. Al-SQfi {Suwar, 313-314), also
gives several other traditions for the placement
of al-shardsif, but concludes that all others are
incorrect.
  Al-SQfi mentions several stars overlooked by
Ptolemy, including one north of fard. He men-
tions a row of stars between Hydra and the stars
of the Lion which inclines far under the foot of
the Lion toward the northeast until it reaches
the fifth star on the southern shoulder of Virgo
(the star that is the first lunar mansion).
Al-SQfi {Suwar, 314) also states:
The second unformed star of Hydra, between Hydra and
Leo, along with the bright stars of Leo which fall in a line,
along with the remainder of the constellation of Hydra are
called al-khayl (the horses). And the small stars which are
around them are called afld^ al-khayl (the colts of the horses);
and near them is the constellation of the Jar, between al-
fard and the constellation of the Raven, and it is called al-
ma"laf{the manger).

Constellation 42. The Jar
[Crater]
FIGURE 83
  Aratus (line 448) mentions that on the back of
Hydra there is a Kparrip, a large bowl or vessel in
which wine was mixed with water and from which
the cups were filled. In some Greek writings,
such as the histories of Polybius (XXXIV, 11.12),
the word is used for a mouth of a volcano. Thus
the latinized form of the word. Crater, has gen-
erally the latter connotation in English and not
the early idea of a large earthenware bowl or
vessel, which is intended in the name of the
constellation. The constellation of Crater is in-
separably bound with the raven (Corvus) and
Hydra, both in its ultimate Babylonian origin
and in most later Greek legends associated with
it (see the discussion of Hydra and Figure 82).
One legend (Hyginus, 11,40) connects Crater
with the cup of Ikarios to whom Dionysus gave
the vine and who was translated into the skies as
the constellation Bootes.
  In the Islamic world Kparrip was translated as
bdtiyah (ajar), which can be observed to be writ-
ten across the decorative vessel in Figure 83.
None of the seven formed stars comprising the
constellation are given individual names. The
traditional Bedouin term for this circular group
of stars was al-ma"laf (the manger) and was as-
sociated with the nearby stars of Hydra, which
were called "horses" and "colts' (al-SQfi {Suwar,
313, 318).
Constellation 43. The Raven
[Corvus]
FIGURE 83
  A figure of a raven, Kopa^ in Greek, was de-
scribed by Aratus (line 449) and later writers as
seeming to peck at the Hydra. According to the
mythographers the Raven, who was a servant of
Apollo, was placed in the skies along with the Jar
and the Snake as punishment for having lied to
Apollo when he blamed the snake for his delay
in fetching water instead of admitting he had
tarried to wait for a fig tree to ripen.
  
204

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY






FIGURE 83.—The Jar [Crater] and the Raven [Corvus]. (Photo: Smithsonian no. 72-3104)

  In the Arabic world the Ptolemaic constella-
tion was translated as al-ghurdb (the raven). A
formal title is not given to the constellation in
Figure 83.
  
The names given the individual stars in Figure
83 all reflect the Greek-Ptolemaic conception of
the asterism as translated into Arabic. The star
in the bill of the raven is labeled minqdr al-ghurdb
  
NUMBER 46

205



(the bill of the raven) and is the first star of the
constellation [a Corvi]. The fourth star, the one
on the wing closest to the jar (Crater) is called
jindh al-ghurdb al-ayman (the right wing of the
raven [7 Corvi]). The larger and anterior star of
the two on the other wing, no. 5, is termed jinaA
al-ghurdb al-aysar (the left wing). This star, b
Corvi, is today called Algorab from the Arabic
word for raven.
  In the Bedouin tradition according to al-SQfi
the stars comprising this constellation were called
al-khibd" (the tent) and al-ajmdl (the camels; al-
SQfi [Suwar, 321] reads ahmdl, which seems un-
likely). They were also called "arsh al-simdk al-
a"zal (the throne or pavillion of the unarmed
simdk), which is a reference to the large star Spica
in Virgo (Kunitzsch [1961], no. 40). This star in
Virgo was viewed as one of the back legs of a
large lion by the early Arabs. Spica can be seen
in the hand of Virgo to the north of the tail of
Corvus in Figure 83. The stars were also called
"ajz al-asad (the buttocks of the lion). Al-SQfi
adds that it is asserted that sometimes the moon
falls short and so resides with "ajz al-asad—that
is, that sometimes it is a lunar mansion.
  The constellation of Corvus consists of seven
formed stars and no unformed. In Figure 83,
eight stars can be seen within the form of the
Raven; however, the one on the tail is misplaced,
and should be directly to the south in the body
of Hydra as the twenty-fourth star of Hydra,
Constellation 44.    QintHrus
[Centaurus]
FIGURE 84
  In the Greek conception of the sky this con-
stellation is the second centaur, the other being
Sagittarius who was half human and half horse.
This centaur was called by Aratus and later Hip-
parchus and Ptolemy, simply Kevravpos (the cen-
taur). Aratus (lines 436-443) says the constella-
tion of Centaurus is placed beneath Scorpio and
the Claws (Libra), but Hipparchus {Comm. arat.,
1,8.21) corrects him by saying that the constella-
tion is nearly all under Virgo, which is north of
Hydra and Corvus seen above his head. Aratus

also mentions that in his right hand he grasps a
wild-beast, dr)plov. This animal was variously in-
terpreted as several different wild animals, and
even at times confused with the constellation of
the Hare or Cetus the sea-monster, but most
frequently he was regarded as a wolf; it can be
interpreted as representing another type of dark-
ness. Writers subsequent to Aratus (e,g,, Erat.
Catast, 40; Hyginus, 11,38) stated that the Cen-
taur carried in his other hand a Svpavs, a thyrsus
or wand carried by devotees of Dionysus (Bac-
chus), which was a phallic symbol consisting of a
stick wreathed in ivy and vine leaves with a pine
cone on top. The Greek legend most frequently
associated with Centaurus was that of the wise
Cheiron who reared and taught Asclepius and
Achilles and surpassed all other centaurs and
men injustice. The scholiasts of Aratus, German-
icus, and the author of the pseudo-Eratosthenes
Catasterismi even called the constellation x^'^P^^^
Cheiron, Others associated it with Pholus, the
most skilled of all centaurs in augury and hence
say that it is approaching the altar (the constel-
lation Ara seen in front of him) with a sacrificial
victim (the wild beast).
  In Figure 84 the centaur is bearded and holds
some vine leaves in one hand and an unidentified
type of animal in the other (the forty-fifth con-
stellation Lupus). In the Arabic world the con-
stellation of the centaur was called qinturus, a
transliteration of the Greek word for centaur
(see Kunitzsch [1974], 200-202], The figure is
titled over his back Surat al-Qintdrus (the con-
stellation of the QTntQrus). Al-SQfi explains that
it is an animal whose foresection is that of a man
and whose back quarter is that of a horse. He
called {Suwar, 334) the object he carried the
qadib al-karm (the rod of the vine). Al-SQfi adds
{Suwar, 323), "Ptolemy said there are 37 stars,
but there are 36 stars, the thirtieth being incor-
rect." Further on al-SQfi gives the coordinates of
the thirtieth star as given in the Almagest, but
adding "but there is no star there which vision
can perceive." Nonetheless, al-SQfi left it in his
catalog that followed at the end of his discussion,
and five centuries later Ulugh Beg also kept it in
his catalog although noting, as had al-Sufi, that
  
206

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 84.—Qinturus [Centaurus]. (Photo: Smithsonian no. 72-3105)

NUMBER 46

207



it could not be observed. The star appears on all
globes studied and is the small, lower star of the
three forming a triangle on the belly of the
centaur. In Figure 84, 37 stars make up Centau-
rus; the lower star in the hand holding the con-
stellation Lupus is to be counted with the wild
animal and not with the centaur. There are no
unformed stars of Centaurus, the one under the
belly being counted as formed. The only star
labeled in Figure 84 is the thirty-fifth star of the
constellation labeled rijl al-qintdri (probably an
engraving error for al-qintdrus), meaning "the
foot of the centaur." This star is the third bright-
est in the firmament and the nearest to us. Its
name today is Rigil Kent, obviously derived from
the Arabic title.
  In the Bedouin tradition no. 35 [a Centauri]
on the extended front leg and no, 36 [fi Cen] on
the other front leg were together called haddr
wa al-wazn, whose meaning is somewhat obscure,
but could roughly be translated as "a poised
object and the weight" and has been compared
to the sun and the earth, Al-SQfi {Suwar, 333)
says he does not know which is haddr and which
al-wazn; however, no. 36 he thinks resembles
haddr, a poised object, because it rises before no.
35, but he notes that they always write haddr
first and al-wazn second (cf. Kunitzsch [1961],
no. 118). The name haddr wa al-wazn is written
between these two stars on globes Nos. 4 and 5.
The "modern" name of fi Centauri, no. 36, is
Hadar. Al-SQfi also remarks that the two stars
were called mahlifan wa mahnithdn as well, "the
two swearings of oaths and the two violations of
oaths" because when the first one rises a person
would swear it is suhayl (Canopus), but when he
sees the second one rise he realizes it was not
Suhayl and that he perjured himself in swearing
that it was.
Constellation 45.    The Wild Beast
[Lupus]
FIGURE 85
  The animal held by the centaur constitutes a
separate constellation. In Figure 85, which pre-
sents an overall view of the area around the
winter solstice on the Smithsonian globe, this

constellation can be seen in the lower left-hand
corner (for a detailed view see Figure 84). On
the Smithsonian globe the constellation is titled
surat sab" (constellation of the wild beast), being
a direct translation of the Greek drjplov. Al-Sufi
{Suwar, 329) says of this constellation "there are
18 stars, but Ptolemy says there are 19." The
star numbered 11 in the constellation, he contin-
ues, is according to the Almagest the southern of
the 3 in the end of the tail, "but there is no star
there which the vision perceives." Yet here again
al-SQfi included the eleventh star in his catalog,
noting that he could not observe it. Similarly,
Ulugh Beg stated in the introduction to his cat-
alog that he was going to eliminate the eleventh
star of Lupus, but in fact he included it in the
catalog just as his predecessor al-Sufi had done.
The star appears on all studied globes.^^
  On the Smithsonian globe there is one extra
star in the constellation of Lupus: the one on the
inside back thigh (marked by an arrow in Figure
84), that is the one above the north of the four
stars forming a quadrilateral on the other thigh.
This seems to be an error on the part of the
maker, for no star appears there in the catalogs,
although al-SQfi does mention an extra one un-
der the middle star at the end of the tail (but not
south where no. 11 is). No stars in Lupus are
labeled on the Smithsonian globe.
  In the Bedouin tradition the constellations of
the Centaur and the Wild Beast were viewed
together as one, and indeed al-SQfi treats Cen-
taurus and Lupus together in his discussions. Al-
SQfi {Suwar, 333) states that the Arabs tradition-
ally called the stars of Centaurus and Lupus
collectively al-shamdrikh (the vine branches
loaded with fruit, or a cluster of grapes or dates).
He adds that they resemble al-shamdrikh "be-
cause of their multitude and the thickness of all
of them." This label al-shamdrikh is written across
the two constellations on globes Nos. 4 and 5.
Constellation 46.    The Censer
[Ara]
FIGURE 86
  This constellation, which hangs in the sky im-
mediately beneath the tail of Scorpio between
  
208

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

N

•4 {«" ''te'Jft i{

- >

FIGURE 85.—The Wild Beast [Lupus]. This view also shows the area around the winter solstice.
See also Figure 84. (Photo: Smithsonian no. 72-3081)

Lupus and Sagittarius, is the only one about
which there is no Bedouin tradition: that is, the
seven stars comprising the constellation do not
seem to have been recognized in the Arab world
before the introduction of Greek astronomy. In

the Greek world the constellation, although a
very small one, seems to have been an important
one, for Aratus (lines 402-435) devoted an un-
usual amount of space to it in his poem, where
he discusses the implications for storms and winds

NUMBER 46

209



at sea which the sighting of these stars can have
for the sailor under various conditions. Aratus,
pseudo-Eratosthenes, Cicero, Manilius, and Hy-
ginus clearly viewed the form of the constellation
as a sacrificial altar, calling it a dvrrjpiov in Greek
or Ara in Latin. Hyginus (11,39) gives the legend
that the gods first made offerings and formed an
alliance to oppose the Titans on this altar which
had been made by the Cyclopes. Some (Erat.
Catast., 39) add that the fire was covered so that
the Titans could not see the power of the thun-
derbolt. Hyginus (and Erat. Catast, 39) has only
four stars in the constellation.
  On the other hand, Eudoxus, Hipparchus
{Comm. arat., 1,8.14 and 1,11.9), Geminus, and
Ptolemy thought of the constellation as a
censer or vessel for burning incense, calling it
dvpiarripiov in Greek and Thuribulum in Latin. It
is this latter tradition which dominated in the
Islamic world where the constellation is called al-
mijmarah (the incense burner); and on all Arabic
globes studied, but one, it is depicted as a censer
upside down with the flames going toward the
southwest. The one exception is the earliest ex-
tant globe. No. 1, where it is drawn to look like
a covered furnace with a bellows, perhaps a kiln.
  No stars bear individual names in Figure 86,
nor do they in the texts of al-SQfi or the catalog
of Ulugh Beg. Al-SQfi states that the Arabs,
meaning the Bedouins, say nothing about these
stars. Yet in his enumeration of the stars al-SQfi
{Suwar, 339) says that behind the second star of
the constellation (the large star toward the east
side of the base of the censer) there is a star not
mentioned by Ptolemy of the fourth magnitude,
which is actually doubled because right near it is
a star of the sixth magnitude; moreover, he says
that between this star and the second star of the
constellation is another star of fifth magnitude
not mentioned by Ptolemy. Such comments,
which by his own statement could not have been
drawn from the anwd" tradition, in addition to
his constant correction of Ptolemy's magnitudes
and coordinates, strongly suggest that al-Sufi
actually tried to observe many of the stars him-
self

Constellation 47.    The Southern Crown
[Corona Australis]
FIGURE 86
  The decorative device to the east of the censer
[Ara] in Figure 86 lies between the front legs of
Sagittarius (see Figure 71 for the entire constel-
lation). It is labeled iklil janubi {southern crown),
an Arabic translation of the Greek oTetpavos
voTios, used by Ptolemy and Geminus before him.
Neither Hipparchus nor Aratus used the term
areipavos, but the constellation may have been
referred to by Aratus (line 401) by the phrase
"the stars turned in a circled ring" {bivu^rol
KVKXW).
  Al-SQfi {Suwar, 344) says the Arabs differed
on the constellation: some called it udhial-na"dm
(the nesting place of the ostriches), with obvious
reference to the ostriches seen in Sagittarius;
others called it al-qubbah (the tent); and still
others saw it as part of the Scorpion's tail.
  No stars in this constellation are labeled on the
Smithsonian globe. According to the star catalogs
the southern crown consisted of 13 formed stars.
The Smithsonian globe has only 12 stars in this
constellation, however. The star that is inexplic-
ably missing is no. 10 of the constellation [v
Coronae Australis] and would have been a com-
panion star to the one that lies immediately to
the northwest of the double star pictured in the
constellation on the Smithsonian globe (see Fig-
ure 71). On the Islamicate globes studied this
constellation is drawn either as a shapeless form
(e.g., Nos. 1 and 3) or as a decorative device
(e.g., Nos. 4 and 5, and the Mughal globes, Nos.
26 and the Smithsonian globe).
Constellation 48.    The Southern Fish
[Piscis Austrinus]
FIGURE 87
  Aratus (lines 385-387) says that below Capri-
corn "before the blasts of the South wind swims
a Fish, facing Cetus, alone and apart from the
  
210

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY


FIGURE 86.—The Censer [Ara] and part of the Southern Crown [Corona Australis]. See also
Figure 71. (Photo: Smithsonian no. 72-3103)

NUMBER 46

211


FIGURE 87.—The Southern Fish [Piscis Austrinus]. (Photo: Smithsonian no. 72-3111).

212

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



former Fishes; and him men call the Southern
Fish {ixOvs voTios)" ([trns.] Mair, 237). Mythog-
raphers (e.g., Erat. Catast., 38) recounted a leg-
end in which Isis, when she was in labor, was
saved by a fish and as a reward for this kindness
she placed the fish and its young (the two fishes
of Pisces) in the heavens. Hyginus (11,41) adds
that for this reason the Syrians do not eat fish,
and worship the gilded images of fish as house-
hold gods.
  In Figure 87 the fish is labeled surat hut janubi
(the constellation of the southern fish), a trans-
lation of the classical Greek title, though with
rather poor grammar. The fish itself contains 11
formed stars, not counting the large one in the
mouth at the end of the water, and such is the
numbering given by al-SQfi and Ulugh Beg in
their catalogs. Ptolemy, however, counted the
one in the mouth in with those of the Southern
Fish, even though he had already numbered it

with those of Aquarius. Thus al-Sufi lists 11,
while Ptolemy had 12. Ptolemy also recorded six
unformed stars. Al-SQfi {Suwar, 347) stated that
he could not observe them, but apparently not
wishing to overthrow completely the authority
of Ptolemy, he nonetheless included them in his
catalog of stars at the end of his discussion of the
Southern Fish. He clearly labeled this section of
the catalog with the statement that these six stars
were-taken from the Almagest. Ulugh Beg on the
other hand went ahead and eliminated them
from his catalog. These six stars are not shown
on the Smithsonian globe or on any globes made
after the fifteenth century.
  One star is labeled in Figure 87, the tarnished
star near the end of the tail, no. 11 of al-SQfi's
listing. It is titled dhanab al-hut, (the tail of the
fish [K Piscis Austrini or 7 Gruis]), and seems to
be labeled only on the later globes (those by the
Lahore workshop and the later No. 31).
  
Extant Islamicate Celestial Globes
6.    Descriptive Catalog
It is He who has appointed for you the stars,
that you might be guided by them
in the darkness of the land and sea.

Introduction
  Some preliminary remarks regarding the for-
mat of this catalog of extant globes and com-
ments on the terminology and styles of the sig-
natures and dates are necessary. The globes are
divided into three major classifications: Class A
globes (Nos. 1-58) are those with the full com-
plement of stars (1018-1025) and the outlines of
the 48 classical constellations; Class B globes
(Nos. 59-90) are those with only a small number
of stars, usually those occurring on astrolabes,
and having no constellation outlines; and Class C
globes (Nos. 91-124) have no stars and no con-
stellation forms. At the end of these three group-
ings are two globes that could not be classified
for lack of sufficient information. There are a
total of 126 globes in the catalog. Within each of
these three classes, the globes are organized so
that those that are signed or dated appear first,
in chronological order, followed by those that
are unsigned and undated. Three globes having
the full set of constellation figures but no stars
do not fit well into the three basic classifications,
although they have been placed in Class A. Sim-
ilarly, four globes having only the twelve zodiacal
figures along with some stars have been placed
in Class B, For further discussion of these globes
see Chapter 1 and the Tables in Chapter 7,
  The basic technical aspects and differences
among the globes are compared in the Tables of
Basic Characteristics of the Globes, which follow
the catalog. The six globes listed in the Adden-
dum to the catalog have not been included in the
following comparative discussion of the globes
nor in the Tables of Basic Characteristics, The
Tables (in Chapter 7) include a breakdown of
globes by methods of construction, when known.

Qur^dn, VI,97
including the metal globes which are obviously
in hemispheres, and those in which a seam is not
visible but plugs are evident. Whenever a globe
does not have a visible seam and also lacks evi-
dent plugs, and further examination was not
possible, it has been placed in the unknown cat-
egory. The Tables also include an attempted
classification of globes based on the technical
accuracy and precision evident in the execution
of the circles and graduations. There are a num-
ber of globes, which, although maintaining the
basic form and design of one of the three basic
classes, have been executed by an artisan who
was not a professional instrument maker, with
the result that the graduations are uneven, the
circles not precisely delineated, and other tech-
nical features are inaccurate, so that the globe
cannot function as a true astronomical instru-
ment. Occasionally these technically inaccurate
and non-functional globes are outstanding ex-
amples of fine metal-working or calligraphy.
  Only two globes with dates prior to the six-
teenth century are inscribed in naskhi script; the
other early globes are engraved in Kufic script.
The globe made by Qaysar (No. 3) in 622 H/AD
1225-1226 for the nephew of Saladin has both
labels and the signature in naskhi script. This
globe is unique in that the constellations bear
engraved Latin titles as well as Arabic ones, and
the ecliptic and equator are numbered with the
European form of Arabic numerals as well as the
abjad letters of the alphabet commonly used in
the Near East; such inscriptions were possibly
engraved well after the globe was actually con-
structed. The other globe dating prior to the
sixteenth century to be inscribed in naskhi script
is a globe bearing the date of 718 H/AD 1318-
1319 made by "Abd al-Rahman ibn Burhan al-
  
213

214

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



MawsilT (No. 60); the authenticity of this date is,
however, questionable.
  Signatures were usually placed around or near
the south poles for the simple reason that space
was to be found there since the southernmost
constellations had not yet been mapped and were
not represented on Islamicate globes, even those
made as late as the nineteenth century. A tran-
scription of all the imprint inscriptions on the
globes will be found in Chapter 8.
  On the globes possessing makers' names, the
word used most frequently as the equivalent of
the Latin opus is san"ah. Interpreting the word
as the noun san"a(t) (product), in construct with
the maker's name, rather than the verb sana"ahu
(he made it) seems preferable, since the Arabic
word for globe, kurah, is always feminine and
would require a feminine object, thus sana"ahd
rather than sana"ahu. Furthermore, on several
globes (e.g., Nos. 11-14) there are clearly dots
over the final letter, indicating that the maker
intended the final letter to be a td" marbdtah and
hence the word to be a noun rather than a verb.
The word san"ah is used on all but two of the
globes dated prior to the sixteenth century as
well as on globes by the seventeenth-century
metal-workers Muhammad Zaman (Nos. 16 and
64) and Lutf Allah ibn "Abd al-Qadir al-Muhibb
(No. 67), and the globes made in Mughal India
by Qa^im Muhammad (Nos. 11-14), his brother
Muhammad MuqTm (No. 15), and his nephew
Hamid (No. 68) of the Lahore workshop. The
very earliest Islamicate globe known today, made
in Moorish Spain in 473 H/AD 1080 employs the
verb from this same root reading sana"a hddhihi
al-kurah (Ibraham ibn Sa " Td . . . made this
globe), as does globe No. 10 by "AIT KashmTrT ibn
Luqman which reads sana"hd (he made it [the
globe]).
  The most prolific member of the Lahore work-
shop, Diya^ al-DTn Muhammad the son of Qa^'im,
employed on his earliest globes (Nos, 18 and 19)
the word san"at, where he wrote the td" marbiitah
as a simple final td". On all of his other globes,
however, Diya^ al-DTn used the word "amal in the
sense of "work" or opus (Nos. 20-24, 26-28, 30,
66,69, 71).
The use of the word "amal is usual with Diya^

al-DTn of the Lahore workshop as well as later
makers such as Muhammad Salih TatawT (Nos.
25 and 29) of the seventeenth century, Ridwan
who constructed his globe in 1112 H/AD 1701
(No. 31, where the term occurs in the expression
kumila "amal hddhihi al-kurah [the construction
of this globe was carried out]), and Muhammad
KarTm (No. 74) of the first half of the nineteenth
century. The word "amal occurs only twice on
globes before the seventeenth century: on the
sixteenth-century globe by Jamal al-DTn Muham-
mad ibn Muhammad al-HashimT al-MakkT (No.
9) which is clearly a poorly constructed and non-
functional globe, and that dated 718 [H]/AD
1318-1319 by ^Abd al-Rahman ibn Burhan al-
MawsilT (No. 60) which is non-functional and also
of questionable authenticity.
  Very few other words are employed to de-
scribe various aspects of the construction or de-
sign. On globe No. 72 the noun rasmah (inscrip-
tion) is found in construct with the maker's name
Muhammad Ashraf TuqadT Zadah. I prefer the
reading of it as a noun rather than as a verb
rasamahu for the same reason that 1 prefer san"ah
to sana"ahu. In the context of this particular
globe, it is likely that the maker intended to
indicate by rasmah that he inscribed the surface
design of the sphere, the sphere itself having
actually been fashioned by another workman.
Since he also signed the stand "the work of
{"amal) Muhammad Ashraf," he may actually
have constructed the stand as well as incised the
graduations on it, while on the sphere he may
have drawn the surface design onto a sphere that
was made by another craftsman out of wood or
possibly ivory. The design and graduations of
this particular globe and stand were executed
with impressive precision.
  The passive verb rusimat occurs when refer-
ring to the positioning of the stars on the globe
by Muhammad ibn MahmQd al-TabarT (who also
uses the noun kutub in the sense of "inscription")
in 684 H/AD 1285-1286, (globe No, 6), and on
the products of the fourteenth-century astrolabe
and globe maker Ja"far ibn "Umar and his son
Muhammad (Nos, 7, 8, and 62) as well as one
anonymous seventeenth-century globe (No. 65).
The globe made in 1069 H/AD 1659 by Hamid
  
NUMBER 46

215



of the Lahore workshop (No. 68) employs the
verbal noun tahrir, meaning the act of composing
or writing, used here apparently in the sense of
engraving the words on the globe. The globe
made in 1221 H/AD 1806 by Muhammad "AlT al-
HusaynT (No. 73) uses the modern Arabic con-
struction of a passive with agent, tummat "aid yad
(it was finished by the hand of), while the Otto-
man Turkish globe made in 1229 H/AD 1882 by
Husayn Husnu KanqarIT (No, 75) refers to the
construction {ikhtird") of the globe, a similar word
being used on the earlier sphere (No. 30) made
in 1090 H/AD 1679-1680, which states that the
sphere was constructed {ikhtird"i) by Diya^ al-DTn
Muhammad, here used perhaps more in the sense
of invention since the design of this sphere is
quite unique.
  The expression bi-rasm is used in two senses
by Qaysar (No. 3, dated 662 H/AD 1225-1226);
first, in the sense of "by the design of Qaysar
ibn ..." and then secondly to mean "by the order
of" the patron. The expression bi-rasm is used in
a similar manner to indicate the patron on the
eighteenth century globe by Ridwan (No. 31)
and the somewhat questionable fourteenth-cen-
tury globe by "Abd al-Rahman ibn Burhan al-
MawsilT (No. 60), which also seems to have the
curious expression matwd bar nazar (undertaken
at the instigation of) and the name of yet another
patron. In a similar way the Persian expression
hasb al-amr is used to express the patronage of
Shah "Abbas the First (No. 63), while on an early
eighteenth-century globe (No. 32) the dedication
to a patron seems to begin bar hasb. On two
Mughal globes (Nos. 25 and 30) the Persian bi-
farmdyish indicates the patron. The globe made
in the Punjab in AD 1842 by Lalah BalhQmal
LahurT employed bi-tajwiz to indicate patronage
(No. 33).
  An unsigned and undated globe gives the
name of a patron who requested that it be made
and figured (No. 37), istasna"ahu wa istaswarahu
using a specific verb to indicate the engraving of
the full set of constellation figures upon the
globe. On the earliest globe made by the Lahore
workshop (No. 11), the proprietor's name, in this
case probably also the patron, is indicated by the
formula sdhibuhu nawab, while a similar expres-

sion, mdlikuhu nawdb is found on a nineteenth-
century globe (No. 92). A nineteenth-century
owner of one globe (No. 59) indicated possession
by the expression min mumtalakdt.
  The names of the makers are frequently pre-
ceded by self-deprecating formulae such as "the
one who is in need of God the Merciful" (No, 4)
or the expressions so common on the globes
produced by the Lahore workshop, aqall al-"ibdd
(the least of the servants [of God]) and ad"af al-
"ibdd (the weakest of the servants). Particularly
elaborate formulae were used by the Safavid
Persian maker Muhammad Zaman, who refers
to himself as "the dust of the threshold of Riza
{turdb "atabat al-rizd; No, 16), a reference to the
sanctuary of an Imam in Mashhad where he
worked, and the Persian expression khak rdh dn
(the dust of the road of life; No, 17).
  The basic language for all Islamicate celestial
globes is Arabic, in that the technical terminol-
ogy and star names are written in Arabic, with
the exception of the two Sanskrit globes (Nos.
29 and 54), one of which has the constellation
names written in both Arabic and Sanskrit. Some
of the signature inscriptions and statements of
patronage on the later globes are in Persian, or
Indo-Persian, or in one instance in Ottoman
Turkish, and many of the later signatures employ
one or two words of Persian. Only very occasion-
ally do Persian words occur in the technical vo-
cabulary itself (for example, on Nos, 30 and 94),
and then usually only prepositions or pronouns.
The grammatical structure and orthography of
the Arabic on the later globes, however, fre-
quently reflect a maker whose main language is
Persian or Indo-Persian (a form of Persian com-
mon in parts of India such as the Punjab) or
other eastern dialects of Persian such as Tajik,
which is spoken in southern Turkestan.
  Occasionally the date of the globe is given in
several eras besides that of the Hijrah, such as
that of the Yazdijird era (Sassanian solar calendar
reckoned from AD 632), the Alexandrian era
(Seleucid era calculated from 312 BC), the Year
of the Flood (a Coptic calendar) and the Malik-
shah era (JalalT), an eleventh-century reform of
the older Persian solar calendar. The year of
construction given in terms of the year of the
  
216

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



reign of one of the Mughal emperors, Akbar,
JahangTr or Shahjahan, occurs on several Indo-
Persian products. And on a nineteenth-century
product of Lahore (No. 33) the date is given in
terms of the Vikrama era still current in India
today as well as in the Christian era (see Table 5
in Chapter 7; for details on various calendars see
al-BTrQnT [CAron. [trns.], 16-36, 136-380], Carra
de Vaux, and Taqizadeh),
  Generally all the globes made before the six-
teenth century express the year in abjad letters,
while all those made in the sixteenth century and
later employ standard numerals for the dates.
The exceptions to this pattern are the thirteenth-
century globes by Qaysar (No. 3) and Muham-
mad ibn Mahmud al-TabarT (No. 6), both of
which use standard numerals, while the anony-
mous globe made in 1056 H/AD 1646-1647 uses
abjad numerals (No. 65). The globe by "Abd al-
Rahman ibn Burhan al-MawsilT (No. 60, dated
718 H/AD 1318-1319 writes the dates out in
words, as does the nineteenth-century one by
Lalah BalhQmal (No. 33).
  Each globe in the catalog is described in as
much detail as possible, either from personal
examination of the globe or from photographs
or published descriptions. Unless otherwise
stated, the entry is based on personal examina-
tion. Following the description of the globe and
statements as to the provenance of the globe if
known, citations are presented to studies, repro-
ductions, or references to the globe (a full list of
the abbreviations will be found in the bibliog-
raphy). For the better known globes only the
most important or very recent citations are listed,
the other citations being referred to in the places
cited, so that if all the items listed are consulted
all known reproductions or descriptions of the
globe can be located.
  A full transcription of all the signature inscrip-
tions and other major inscriptions is to be found
in Chapter 8 where the numbers correspond to
the catalog numbers of the globes. The inscrip-
tions on all the signed globes will also be de-
scribed in the Repertoire des facteurs d'astrolabes
et de leurs oeuvres. Premiere partie: Islam, by A,

Brieux and F.R, Maddison, who generously gave
me access to their files while I was preparing this
catalog. Since the Repertoire has not yet been
completed and published, however, I have not in
each instance given a citation to the forthcoming
Repertoire.
  Since the precise number of stars on a globe
very frequently differs from the number given
in standard catalogs and even, on occasion, from
the total specifically stated on a globe, usually
because the maker inadvertently adds or omits
some stars or sometimes purposefully adds an
extra star such as the "overlooked star" in Ursa
Major, in the present catalog the number of stars
is always qualified by "approximately" unless the
specific number has actually been counted. The
approximate number given for Class A globes
depends on which star catalog the maker was
likely to have been using (i.e., Ptolemy with 1025
stars, al-SQfi with 1024, or Ulugh Beg with
1018).
  Unless otherwise stated a globe has holes
drilled at the two (celestial) equatorial poles, so
that the axis of rotation could pass through the
celestial poles. The term "equator" on a globe,
of course, refers to the celestial equator. In re-
gard to the rings, the phrase "non-consecutive
segments of 90°" is used to describe a labeling
such as that illustrated in Figures 31, 32, and
33—that is, the ring is divided into four quad-
rants, but the graduations are arranged so that
there are only two points, diametrically opposite,
that serve as zero points, the other two points
separating the quadrants being marked 90°. The
numbering on the ecliptic, equator, and rings is
with abjad numerals unless otherwise specified.
  In the descriptions of some globes I have re-
ferred to there being inscriptions in double-out-
lined writing. This term is intended as an aid in
grouping together globes that have the common
characteristic of labeling, most frequently of the
zodiacal names, with each letter formed by par-
allel lines producing a wide brush-like stroke
rather than the usual engraving with letters
formed by single lines. Examples can be seen in
Figures 7 and 20, the former showing a narrow
  
NUMBER 46

217



form of double-outlined writing.
  When a metal globe appears to be seamless, it
has been stated in this catalog that it has "no
seam observable" rather than stating that it is a
seamless cast globe, because only examination of
the inside or use of radiography can determine
with certainty whether such a globe was raised
or cast. Whenever plugs are visible, it is also
noted. In the case of the Smithsonian globe (No.
38), tests of course have confirmed that it was
cast by the cire perdue process. No attempt has
been made to distinguish the copper alloys, brass
and bronze, or the quaternary alloy commonly
used in cast objects, for the actual composition
of the alloys varied greatly and only laboratory
analysis can determine the nature of the alloys
employed.
  See Chapter 2 for a discussion of the methods
of construction and the design and function of
the three basic types of Islamicate celestial globes
and Chapter 1 for a discussion of the individual
known globe makers.
The Catalog
CLASS A—CELESTIAL GLOBES WITH
CONSTELLATIONS
Signed or Dated
1.
  Location: Florence, Istituto e Museo di Storia della
Scienza; Inventory No. 2712.
  Date: The first of Safar of the year 473 H [22 July AD
1080] (or the first of Safar 478 H/28 May AD 1085).
  Maker: IbrahTm ibn Sa"Td al-SahIT al-Wazzan with his son
Muhammad.
  Metal; in hemispheres with seam along equa-
tor. Diameter 209 mm. Forty-seven constellation
outlines; Crater is not drawn or labeled, though
the stars are present. Has 1015 stars, each indi-
cated by a dot inside an engraved circle. Ecliptic
latitude circles; southern equatorial polar circle,
around which there is the inscription giving the
maker. Ecliptic and equator graduated by single
degrees with every fifth labeled; ecliptic repeats
every 30°, equator is labeled continuously from

vernal equinox. Zodiacal names engraved along
ecliptic; equatorial poles, constellations, and
many stars labeled; MaghribT KQfic. Very uni-
form and precise globe. Signature reads:
"Ibrahim ibn Sa"Td al-SahiT al-Wazzan in Valencia together
with Muhammad his son made {sana"a) this globe with stand
{al-kurah dhdt al-kursi) for the holder of the dual office of
wazlr, the supreme commander-in-chief (^a^irf) Abu "Isa ibn
Labbun, may God prolong him and sustain him. And the
fixed stars were placed on it in proportion to their magni-
tude and sizes. It was completed the first of the month of
Safar in 473 H [or can be read as 478 H]. God bless him [the
prophet] and give him secure peace."
[Dhu al-wizdratayn means he was holder of two
offices of wazir, i.e., the positions of wazir of war
and wazir of peace {qalam). See Meucci, 11.]
  Modern wooden pedestal stand having four
supports for horizon ring; ungraduated meridian
ring. The rings are recent replacements and not
the originals mentioned in the inscription.
Entry based on published accounts.
Citations
Meucci. Presents a lithograph reproduction of globe and
analysis of terminology; says date can be read either as
473 or 478 but prefers 473.
Rep. epig. arabe, 7: no. 2121. Reproduces inscription.
Fiorini, 38-39.
Stevenson, 1:28-29.
Destombes [1957], 319, no. 2; reads date as 473.
Destombes [1958] 305-306; reads date as 478.
Mayer [ 1959], 51; reads date as 478.
Catalogo degli strumenti du Museo di Storia della Scienze,
Florence, 1954, No. 2712.
Righini Bonelli, 156; plate 22 is color photograph.
2.
Location:    Paris, private collection of M. Destombes.
Date:    539 H [AD 1144-1145].
Maker:    Yiinus ibn al-Husayn al-Asturlabl.
   Metal; hemispheres joined together along
roughly the path of the Milky Way (which is not
otherwise indicated on the globe); later resold-
ering has marred some of the engraving. Diam-
eter 175 mm. The walls are quite thick (6-7 mm)
and the globe is very heavy; probably cast hemi-
spheres. Modern varnish on globe; at every 90°
along the seam there is a nail of different alloy.
   
218

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Full set of constellation figures; about 1025 stars
indicated by inlaid silver points varying in size
corresponding to magnitude and numbered
within each constellation. Ecliptic latitude circles;
ecliptic and equatorial polar circles. Ecliptic and
equator graduated by single degrees with every
fifth labeled; ecliptic repeats every 30°, equator
is numbered continuously from vernal equinox.
Names of zodiacal signs engraved along ecliptic;
equatorial and ecliptic poles, constellations, and
72 stars and 28 lunar mansions labeled; KQfic.
Very uniform and precise globe. The signature
near the south equatorial pole reads: "This globe
includes all the stars mentioned in the book of
the Almagest after modifying them in proportion
with the interval between the calculations of Pto-
lemy and the year 540 H, i.e., 15° 18'. The work
of {san"at) Yunus ibn al-Husayn al-Asturlabi [in
the] year 539".
  Meridian ring missing. Metal quadruped stand
on ring base with horizon ring is more recent,
possibly added in eighteenth or nineteenth cen-
tury during a restoration. The horizon ring is
graduated unevenly by single degrees with every
fifth labeled in non-consecutive segments of 90°;
however, there are no notches for the meridian
ring and no undersupport for the globe. The
stand is non-functional, although it may have
been copied from a stand that would have been
outfitted with a stationary meridian ring, which
would rest on the horizon ring.
  Acquired about 1930 in Bombay and then sold
at a public sale on 17 April 1958 at Hotel des
Commissaires-Priseurs, rue Drouot. Exhibited
April-August 1976 at the exhibition "Science and
Technology in Islam" held at the Science Mu-
seum, London, as part of the Festival of the
World of Islam.
Citations
Destombes [1958], 300-312, 2 photographs.
Destombes [1959], 447-452. Concludes the maker was a
   Persian working in Isfahan or Rayy.
Muris 8c Saarmann, 36.
Mayer [1959], 296.
Guye & Michel, 210; plate 198 gives an overview of globe.
Maddison Sc Turner, 77.

3.
Location:    Naples, Museo Nazionale.
Date:    622 H [AD 1225-1226].
  Maker: Qaysar ibn Abi al-Qasim ibn Musafir al-Ashrafi
al-Hanafi.
  Metal; hemispheres with seam on a plane
through equatorial poles, near but not on solsti-
tial colure. Diameter 221 mm. Full set of con-
stellation figures engraved and damascened with
copper. About 1025 stars indicated by inlaid
silver points within small circles, in five different
sizes to indicate magnitude. Ecliptic latitude cir-
cles; equatorial polar circles. Zodiacal names en-
graved along ecliptic; constellations and many
stars labeled; the names of the constellations and
a considerable proportion of the star names are
damascened with silver; naskhi. Ecliptic and equa-
tor graduated by single degrees with every sixth
labeled with abjad numerals; the equator is num-
bered continuously beginning at the vernal equi-
nox while the ecliptic is numbered in intervals of
30°. Very uniform and precise graduations and
circles. Every 30° interval along the ecliptic has
also been labeled with the European form of
Arabic numeral and the equator labeled at 100°,
200°, 300°, and 360°; engraved at a later date
(?). The constellations also bear engraved Latin
titles (Pengraved later). Globe also has on it a
scale of the five sizes of inlaid points as a guide
to the stars of the first five magnitudes. It bears
two inscriptions engraved in naskhi script and
damascened with silver, both beneath the South-
ern Fish. The first reads: "By the order of (^i-
rasm) the treasury {khizdnah, i.e., library or cab-
inet of curiosities) of Mawlana al-Sultan al-Malik
al-Kamil, the learned, the just, the defender of
the world and of the faith {al-"dlim al-"ddil ndsir
al-dunyd wa al-din), Muhammad ibn AbT Bakr
ibn AyyQb, great be his victory." The second
reads: "By the design of {bi-rasm) Qaysar ibn AbT
al-Qasim ibn Musafir al-Ashrafi al-Hanafi in the
year 622 H, with an augmentation of 16°46'
over what is in the Almagest."
   A quadruped metal stand with legs concave in
profile which meet to form a pedestal support
for the meridian ring and semicircular arc hold-
  
NUMBER 46

219



ing the horizon ring. Horizon ring graduated by
single degrees with every sixth labeled in KQfic
script; ring has the four cardinal points and sea-
sons engraved near outside edge; numbering of
graduations is in non-consecutive segments of
90° beginning at the East-West points and end-
ing at the North-South points. The meridian ring
is graduated by single degrees with every sixth
indicated by a longer line; the European form of
Arabic numerals for 30, 40(?), and 60 have been
engraved along a 30° interval of the meridian
ring; added later (?), Entire meridian ring may
be a replacement. Incorporated into the stand
are two gnomons on the west and east sides each
above a 90° arc in the side of the leg directly
under it; the arc is graduated by single degrees
with every fifth labeled in Kufic abjad numerals.
The gnomons and scales allow a direct reading
of the elevation of the sun. The stand and hori-
zon ring may be contemporary with the globe.
  Formerly in the Museo Borgiano, Velletri. Ex-
hibited at exhibition "Science and Technology in
Islam" held at the Science Museum, London, as
part of the Festival of the World of Islam, April-
August 1976.
  Entry based on published accounts and pho-
tographs in the files of A, Brieux and F,R, Mad-
dison,
Citations
Assemani. Includes three lithographed drawings which pres-
ent a rather poor facsmile of the globe, and an analysis of
the terms which is not always reliable.
Fiorini, 31-33.
Stevenson, 1:29
Rep. epig. arabe, 10: no. 3924; reproduces inscription.
Mayer [1956], 80-81.
Destombes [1957], 319, no. 4.
Margotta, 109 (photograph); no discussion.
Sabra[1969], 8.
Maddison 8c Turner, 78.
4.
Location:    London. British Museum, Department of Ori-
ental Antiquities; Inventory No. 71.3.1.
Date:   674 H [AD 1275-1276].
Maker:    Muhammad ibn Hilal al-Munajjim al-MawsilT.
  
Metal; hemispheres joined with seam along
ecliptic. Diameter 240 mm. Full set of constella-
tion figures with about 1025 stars indicated by
inlaid silver points. Ecliptic latitude circles. Eclip-
tic and equator graduated by single degrees with
every fifth labeled; ecliptic repeats every 30°;
equator beginning at vernal equinox has three
segments of 100° and one of 60°, Very uniform
and precise graduation and circles. Zodiacal
names engraved along ecliptic; equatorial poles,
constellations and many stars labeled; KQfic, In-
scription below Southern Fish, inlaid with silver,
reads: "The work of {san"at) the one who is in
need of God, may He be exhalted, Muhammad
ibn Hilal al-Munajjim al-MawsilT in the year 674
H,"
  An eight-lobed base frame supports four legs,
concave in profile, which meet to support two
crossing hemispherical bars holding the horizon
ring. The latter is graduated by single degrees
with the names of the four cardinal points en-
graved near outer edge. Labeling on ring is in
naskhi scr\\)t, indicating it is probably not contem-
porary with the globe. On each side of one of
the hemispherical supports of the horizon ring
are two devices that look very similar to the two
gnomons on globe No. 3, They are positioned
over the curved legs in just the same manner as
in No, 3, but the concave surface of the leg is
not graduated. Meridian ring missing.
  Formerly in the collection of Sir John Malcolm
(1796-1833), who was presented with the globe
by the leader of the Bohara community whose
ancestors had brought it to India from Persia. It
was on deposit with the Royal Asiatic Society
until 25 February 1871, when it was purchased
by the British Museum from General George
Malcolm.
Citations
Dorn [1830]. Includes two lithographed plates representing
the constellations in each hemisphere and a detailed study
of the globe.
Rep. epig. arabe, 21: no. 4708; reproduces inscription.
Fiorini, 37.
Stevenson, 1:29-30.
Mayer [1956], 68.

220

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Lehner,   122-123.   Reproduces  lithographed  plates  from
  Dorn.
Muris it Saarmann, 36-37.
Pinder-Wilson BM. Presents detailed photographs of each
  constellation.
Destombes [1957], 315 photograph; 319 no. 3, discussion.
5.
  Location: Dresden. Staatlicher mathematisch-physikal-
ischer Salon.
  Date: None given; estimates range from AD 1278 to AD
1310.
Maker: Muhammad ibn Mu^ayyad al-"UrdT.
  Metal; in hemispheres with seam at ecliptic.
Diameter 146 mm. Full set of constellation fig-
ures with about 1025 stars indicated in silver
inlaid points. Ecliptic latitude circles damascened
in silver at the zodiacal divisions. Ecliptic and
equator graduated by single degrees with every
fifth labeled; ecliptic repeats every 30°; equator
beginning at vernal equinox has three segments
of 100° and one of 60°. The two lines of the
equator are inlaid with gold, while all other cir-
cles are inlaid with silver. The names of the
zodiacal signs are inlaid alternately in gold and
silver along the ecliptic; equatorial and ecliptic
poles, constellations, and 130 stars are labeled;
inscription inlaid in silver and gold; KQfic. Near
equatorial pole is the inscription: "The work of
{san"at) Muhammad ibn Mu^ayyad al-"UrdT."
   Possesses a horizon ring, a meridian ring, and
a zenith ring and a suspensory device by which
it may be hung. Is unique among the extant
globes in not having the horizon ring attached to
a stand. All the rings are graduated by single
degrees with every fifth labeled; metal horizon
ring has the two directions (east and west) near
outer edge of ring and the numbering in non-
consecutive segments of 90° begins at the east-
west points; metal meridian ring also numbered
in non-consecutive segments of 90° beginning at
points of attachment to the horizon ring.
  It appears that the metal zenith ring can pivot
around the globe, and it appears to have a slit
the length of the upper quadrant parallel to the
face of the ring. It is possible that a needle-like
gnomon could have been inserted through this
slit, similar to that described by al-Battani (see
Chapter 2).  There  is a  more  recent wooden

quadruped stand on a circular base with a central
support in which globe and rings can rest; not a
functional part of the globe.
  Purchased in AD 1562 by the Elector August
of Saxony from Nicholaus Valerius, mathemati-
cian of Coburg. Preserved in the court art collec-
tions in Schloss Moritzburg until the early eight-
eenth century, when scientific instruments were
placed in the Mathematisch-physikalischer Salon
in the Zwinger palace. Exhibited at exhibition
"Science and Technology in Islam" held at the
Science Museum, London, as part of the Festival
of the World of Islam, April-August 1976.
Entry based on published accounts.
Citations
Drechsler. Fight lithographs reproduce the constellations;
terminology analyzed; estimates date AD 1278-1279.
Harari. Four good photographs on plate 1403, volume 6.
Rep. epig. arabe, 13: no. 4699 reproduces inscription.
Hartner [1950], 190.
Hartner [1955], 8, photograph.
Fiorini, 37.
Stevenson, 1:30-31, and photograph in figure 14.
Destombes [1957], 320 no. 6, where he estimates a date of
AD 1304 a.ssuming the maker used a value of 1 ° per 66
years, as did al-SufT, for the precession, or at AD 1310 if
he used 1 ° per 70 years as advocated by Nasir al-DTn al-
Tu.sT, founder of the Maragha observatory.
Mayer [1956], 72-73; says merely last third of thirteenth
century.
Needham, 382, and photograph in figure 174, plate LIX.
Maddison ^- Turner, 79.
Gr6tz.sch, plates 108 and 109.
Zick-Nissen, plates 47 figure 4, 49 figure 3, 50 figure 5.
6.
  Location:    Paris.  Mu.see du  Louvre,  Section   Islamique;
Inventory No. 6013.
Date:    684 [H] [AD 1285-1286].
Maker:    Muhammad ibn Mahmud al-TabarT.
  Metal; no seam observable; two plugs ob-
served; rattles slightly. Diameter 130 mm. Full
set of constellation figures with about 1024 stars
indicated by inlaid silver points. Ecliptic latitude
circles. Ecliptic and equator graduated by single
degrees with every fifth labeled; ecliptic repeats
every 30°; equator beginning at vernal equinox
has three segments of 100° and one of 60°.
Zodiacal names engraved along ecliptic; constel-
lations and many stars labeled; Kufic. The mak-
  
NUMBER 46

221



er's name is given near the north pole: "The
work of {san"at) Muhammad ibn Mahmud al-
TabarT." A second inscription near the south
equatorial pole reads: "These stars were placed
{rusimat) according to the Book of Constellations
[written] by AbQ al-Husayn al-SQfi after increas-
ing their longitudes 5° [to] our epoch, while the
correction of whatever errors occur and the pro-
duction of variant readings* [is left] to those who
are expert in language"!", ^"d that is in the year
684. Inscriptions:]: of Muhammad ibn MahmQd
al-AsturlabT."
  The globe has holes bored at the ecliptic poles
as well as the equatorial poles. The Kufic script
is not very angular and at times tends toward
naskhi. The great circles wobble slightly and the
graduations are uneven. The crossing of the
equator and ecliptic at the equinoxes is poorly
executed. The stars themselves are so poorly
positioned on the globe that the epoch to which
they conform varies from the tenth to the six-
teenth century AD. It appears that the stars were
positioned after the constellation outlines were
engraved. The constellations, however, are quite
nicely engraved. The imprecision suggests that
the globe was not made by a professional instru-
ment maker, despite the maker signing himself
as an astrolabe maker. This globe is outstand-
ingly attractive visually, but is in fact imprecise
and technically non-functional.
  An eight-lobed base with center cross pieces
supports four legs (which are clearly modern)
attached by screws to the horizon ring. Four 90°
arcs are attached to the base and meet to form a
pedestal support for the meridian ring and a
semicircular arc supporting the horizon ring.
The horizon ring is graduated by 2° intervals
with every tenth labeled; the numbering repeats
every 90° proceeding clockwise. The meridian
ring is ungraduated on the east side, as presently
mounted, but graduated by 2° intervals on the
other, in non-consecutive segments of 90° begin-
ning at the points opposite the celestial equator
and ending at the pole. Two of the 90° arcs
attached to the base are graduated by 5° inter-
vals with each labeled in Kufic abjad letters from
the top of the arc. Two gnomons are attached to
two of the semicircular arcs supporting the globe,

but the gnomons are not aligned over the grad-
uated arcs as one would expect. The general
design of the stand is strikingly similar to No. 3
(made in AD 1225-1226), except that the gno-
mons are not aligned properly (a failure of the
maker to realize the function of the gnomons, or
perhaps a poor job of repairing the globe). While
the alloy, finish, and engraving of the stand seem
to be contemporary with that of the globe, the
construction of the stand seems rather recent,
which might be due to the stand having been
severely and incorrectly repaired and recon-
structed in recent times.
See Figure 6 for an illustration.
  The globe was acquired by the Mission Ar-
cheologique Francaise in Cairo and since 1892
has been in the Louvre. From June to August
1971 it was exhibited in the "Arts de I'lslam des
origines a 1700 dans les collections publiques
fraiifaises" at the Orangerie des Tuileries in
Paris.
Citations
Casanova. A detailed study of the terminology on the globe;
contains no reproduction of the globe
Harari, 5:2518.
Rep. epig. arabe, 13: no. 4864; reproduces inscription.
Destombes [1957], 316 photograph; 320 no. 5, discussion,
where maker is listed as Muhammad ibn Muhammad al-
AsturlabT.
Destombes [1958], 307, where maker given as Muhammad
ibn Muhammad ibn "AIT al-TabarT.
Mayer [1956], 72.
Orangerie des Tuileries, plate 190; text p. 42.
Sabra[1976], 194, photograph.
   * Reads tashtf meaning "the misplacement of diacritical
|5oints" or "the production of alternative readings."
  "f" Reading mutarjamin; or one could read mutarahhimln
(omitting the diacritical point) meaning "those who are sym-
pathetic or understanding." Casanova (3 19) prefers the latter
reading.
  J Reads kutub, in the sense of inscriptions; see Dozy 2:
450.
7.
  Location: Oxford. Museum of the History of Science,
Lewis Evans Collection; Inventory No. 2900.
   Date: 764 H [AD 1362-1363], 732 Yazdijird, and 1674
Alexandrian era.
Maker:   Ja"far ibn "Umar ibn Dawlatshah al-KirmanT.

222

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



  Metal; hemispheres joined at equator with sol-
dering evident along seam. Diameter 165 mm;
weight 1814.4 gr. Full set of constellation figures
with about 1025 stars indicated by inlaid silver
points with a depression in the center of each.
Ecliptic latitude circles. Ecliptic is graduated by
single degrees with no numbering; equator is
graduated by single degrees with every fifth la-
beled in three segments of 100° and one of 60°.
Zodiacal names engraved along ecliptic; ecliptic
and equatorial poles, constellations and about 50
stars labeled; KQfic. Holes have been drilled at
the equatorial poles, the ecliptic poles and at a
position as far from the equatorial poles as the
ecliptic poles, but in the opposite direction along
the solstitial colure (the purpose of the latter
holes is unknown). A removeable pin has been
inserted at both the northern and southern eclip-
tic poles. The KQfic inscription beneath the
Southern Fish reads: "The stars were placed {ru-
simat) according to the Book of Constellations
[written] by AbQ al-Husayn "Abd al-Rahman al-
SQfi after increasing their longitudes by 6°3' to
our time in the year 764 H, 732 Yazdijird, and
1674 of the Alexandrian era. The work of
{san"at) ]a"far ibn "Umar ibn Dawlatshah al-Kir-
manT." The inscription is engraved over an older,
rubbed out inscription, which seems to record
that the globe was made by Ja"far for a patron
named Muhammad ibn AsTl.
  There is a quadruped metal stand supporting
two semicircular arcs holding a horizon ring
bearing the four cardinal points near the outside
edge; the ring is graduated by single degrees
with every fifth labeled in non-consecutive seg-
ments of 90° beginning at the east-west points
and ending at the north-south points where there
are notches. The stand, which seems to be orig-
inal with the globe, is curious in that it allows no
room for a meridian ring, which is usually an
integral part of a complete celestial globe with
stand. There are notches at the north-south
points, but not of a size to accommodate a regular
meridian ring, for the semicircular metal arc
attached to the underside of the horizon ring at
the north-south points will not permit a ring to
pass through the notches. This semicircular arc
essentially serves to mark the nadir of the merid-

ian ring and has a series of holes at 10° intervals
for the insertion of a pin. Thus if the moveable
pin at the south ecliptic pole were to be placed
in the equatorial southern pole of the globe it
could then be inserted in one of the holes of this
semicircular arc and the globe thus adjusted to
different terrestrial latitudes by 10° increments.
Perhaps a semicircular arc (now lost) was in-
tended to be placed over the globe and inserted
into the notches at the north-south points to
serve as the top of the meridian ring and yet be
removable to allow readjustment of the globe.
See Figure 9 for an illustration.
  Bought by Lewis Evans in 1922 from Percy
Webster of London. Exhibited at the exhibition
"Science and Technology in Islam' held at the
Science Museum, London, as part of the Festival
of the World of Islam, April-August 1976.
Citations
Gunther [ 1923], 2:247-249, with a photograph between pp.
248 and 249. Lists the stars labeled on globe; does not
  mention older inscription.
Mayer [1956], 53; with photograph of inscription in plate
  XI.
Destombes [1957], 321 no. 9.
Destombes [1960], 205-206; gives inscription.
Maddison [1961], photograph plate 153B.
Harari, 5:2518.
Muris & Saarmann, 38.
Ghirshman et at., 135; large color photograph. No location
or source is given. Plate is labeled "Persian celestial globe,
with  stars  inlaid  in  silver dated  764."   Photograph  is
  printed as a mirror image.
Nasr [ 1976], 123, plate 80. Labeled as a spherical astrolabe,
  on the plate; correction given in acknowledgements.
Maddison & Turner, 80.
8.
Location:    Istanbul. Kandilli Observatory, History of
Science Museum; Inventory No. 763.
Date:    785 H [AD 1383-1384].
Maker:   Ja"far ibn "Umar ibn Dawlatshah al-Kirnianl.
  Metal; hemispheres joined at equator, A band
of metal of a different alloy has been placed along
the seam, with irregular hatch-marks or gradua-
tions (probably added later, for the graduations
elsewhere on the globe are quite precise). Diam-
eter 114,3 mm. Full set of constellation figures
with about 1025 stars indicated by inlaid silver
  
NUMBER 46

223



points with a depression in the center of each,
differing in size apparently according to magni-
tude. Ecliptic latitude circles. Ecliptic is gradu-
ated by single degrees with no numbering; equa-
tor is graduated by single degrees with every
fifth labeled, in three segments of 100° and one
of 60°. Zodiacal names engraved along ecliptic;
ecliptic and equatorial poles, constellations and
about 50 stars labeled; Kufic. Holes have been
drilled at the equatorial poles, the ecliptic poles
and at a position as far from the equatorial poles
as the ecliptic poles, but in the opposite direction
along the solstitial colure (the purpose of the
latter holes is unknown). The inscription beneath
the Southern Fish reads: "These stars were
placed {rusimat) with an increase of 6° [lacuna]
minutes according to [the Book of] the Constel-
lations [by] "AbQ al-Husayn "Abd al-Rahman al-
SQfi in the year 785 H. The work of {san"at)
Ja"far ibn "Umar ibn Dawlatshah al-KirmanT."
  The increment ought to read 6° 23' to give a
value for the precession equivalent to that em-
ployed by the maker of the previous globe (No.
7) and by his son (globe No. 62). This globe is
identical in nearly every respect to globe No. 7.
Stand and rings missing.
  Entry based on published descriptions and
photographs.
Citations
Destombes [1960], 206 note 2; gives inscription.
Dizer Sc Meyer; present six photographs.
9.
  Location: Paris. Bibliotheque Nationale, Departement
des Cartes et Plans; Inventory No. Ge.A.326.
Date:    981 H and 952 Yazdijird [AD 1573-1574].
  Maker: Jamal al-DTn Muhammad [ibn] . . . al-DTn Mu-
hammad al-HashimT al-MakkT.
  Metal; appears to be seamless; no seam observ-
able from inside or outside and four possible
plugs visible of same alloy; the signature is dam-
aged by a plug of lead 14 X 17 mm and a
rectangular hole 13 X 16 mm. Diameter of globe
150 mm; wall thickness 2.5 mm; weight 850
grams. Full set of constellation figures rather
poorly drawn. About 900 stars indicated by large
silver inlaid points, quite poorly and inaccurately

placed. Ecliptic latitude circles. Ecliptic gradu-
ated into unlabeled irregular intervals; equator
graduated also into irregular intervals with every
fifth, more or less, indicated by a longer line and
labeled. Zodiacal names engraved along ecliptic;
constellations and some stars labeled; naskhi.
Stars appear to have been added after constella-
tion outlines engraved. Has no holes at equatorial
poles. There is a cord at the north ecliptic pole
by which it could be suspended. The inscription
beneath the Southern Fish in crude naskhi is
badly damaged by a large rectangular hole and
a plug of lead. It appears to read: "Here is the
work {"amal) of the least [of the servants] . . . al-
A"azz Jamal al-DTn Muhammad [ibn] . Din
Muhammad al-HashimT by ancestry {nasaban),
al-Makki by birth {mawlidan) ... in the most no-
ble of cities . . . may [God] increase its great-
ness . . . with high esteem and honor in the year
981 H and 952 Yazdijird."
Stand and rings missing.
  Entry based on photograph supplied by A.
Brieux and F.R. Maddison, published accounts,
and information obtained from M. Destombes
and A.J. Turner.
  Formerly in the collection of M. le Comte de
Viel Castel; acquired by Bibliotheque Nationale
in 1857. Exhibited in Paris at the UNESCO
exhibition "L'Islam et les Sciences," 8-21 July
1981.
Citations
Vallee, 47 no. 292.
Stevenson, I: fig. 15, opposite p. 33. States globe is by
Diemat Eddin Mohammed and dated 981 H/AD 1573,
information furnished him by Vallee. This is apparently
an attempt to attribute it to Diya^ al-DTn Muhammad of
Lahore of the seventeenth century.
Fiorini, 47. Here it is also attributed to Diemat-Eddin-
Mohammad.
Muris & Saarmann, 38; again gives maker as Diemat Eddin
Mohammad.
Mayer [1956], 73; gives correct maker.
10.
Location:    London. Private collection.
Date:    998 [H] and thirty-fourth year of reign of Akbar
[AD 1589-1590].
Maker:    '^All Kashmiri ibn Luqman.

224

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



  Metal; no seam observable; four small round
plugs visible in southern hemisphere. Diameter
unknown. Full set of constellation figures with
about 1018 stars indicated by inlaid silver points.
Ecliptic latitude circles. Ecliptic and equator in-
dicated by single lines on which are small dots at
2° intervals without numbering. Zodiacal names
are engraved along ecliptic in small inconspicu-
ous labels and do not extend to the width of each
house as is customary. Constellations, equinoxes,
solstitial colure, equatorial poles, and major stars
labeled; naskhi tending toward ta"liq, written
without diacritical points. Beneath the Southern
Fish is an inscription in Arabic without diacritical
points which seems to read: "The one in need of
God the Merciful, "AIT KashmTrT ibn LQqman,
[in] the year 998 corresponding to 34 [of the
reign] of Akbar affixed to the globe the observ-
able stars after the calculation of their position
{taqwim), and he made it {sana"ahd) in a manner
useful for the knowledge of all the requirements
of astrolabe makers, as an aide-memoire to their
craft."
  The word for "of Akbar" {Akbari) is not at all
clear on the globe, but if the year 998 is inter-
preted as a Hijrah date, then the thirty-fourth
year of Akbar's reign would correspond as stated.
Stand and rings missing.
See Figures 11 and 69 for illustrations.
  Entry based on photographs in the files of A.
Brieux and F.R. Maddison.
11.
  Location: .Stonvhurst College Library near Blackburn,
Lancashire, Englancl.
  Date: The eighteenth year of the reign of JahangTr [AD
1622-1623].
  Maker: Qa^im MLihammad ibn "l.sa ibn /Mlahdad Astur-
labT LahurT HiMiia\ iiiiT.
   Metal; no seam observable; three plugs ob-
served including one large round one near in-
scription; rattles. Diameter 188 mm. Full set of
constellation figures with about 1018 stars indi-
cated by inlaid silver points. Ecliptic latitude cir-
cles. Ecliptic and equator graduated by single
degrees, with every fifth labeled; both ecliptic
and equator are continuously numbered from 5°

to 360°, but the ecliptic bears an additional
numbering of 5 to 30, repeated for 12 segments.
Zodiacal names engraved along ecliptic; the
names of the 28 lunar mansions are also engraved
along ecliptic on opposite side from zodiacal
names; Constellations, celestial poles, and major
stars labeled; naskhi. Between the two south poles
and the end of Eridanus an inscription in Arabic
reads: "Its owner {sdhibuhu) is Nawab (His Ex-
cellency, l"tiqad Khan. The eighteenth year of
the reign of JahangTr." Beneath in the same hand
it reads "The work of {san"at) the least of the
servants {aqall al-"ibdd) Qa^im Muhammad ibn
"Isa ibn Allahdad AsturlabT LahQrT HumayQnT."
  The ungraduated metal meridian ring and the
wooden pedestal stand on which it is mounted
are of very recent origin.
  See Figures 12, 13, 37, 41, and 46 for illustra-
tions.
Acquired in Lucknow in 1858.
12.
Location:    Paris. Private collection.
   Date: 1035 H and the twenty-first year of the reign of
JahangTr [AD 1625-1626].
  Maker: Qa^im Muhammad ibn "Isa ibn Allahdad Astur-
labT LahurT HumayunT.
  Metal; no seam observable; one round plug is
missing from globe in the constellations of
Aquilla and Serpens (see Figure 36 for illustra-
tion). Diameter unknown. Full set of constella-
tion figures with about 1018 stars indicated by
inlaid silver points. Ecliptic latitude circles. Eclip-
tic and equator graduated by single degrees with
every fifth labeled; equator numbered continu-
ously from vernal equinox, while ecliptic repeats
every 30°. Zodiacal names engraved along eclip-
tic; the names of the 28 lunar mansions are also
engraved along ecliptic on opposite side from
zodiacal names; constellations, celestial poles and
major stars labeled; naskhi. Around the two
southern poles is written the following inscrip-
tion: "The year 1035 H. Work of {san"at) the
least of the servants Qa^im Muhammad ibn "Tsa
ibn Allahdad AsturlabT LahQrT HumayQnT. The
twenty-first year of the reign of JahangTr."
The   ungraduated  horizon  ring  with  quad-

NUMBER 46

225



ruped stand on ring base is a recent replacement.
Meridian ring missing. The globe has been
mounted by a pin passing through the celestial
poles and fixed horizontally to the horizon ring.
  Formerly in the collection of Bahari in Teh-
eran,
  Entry based on photographs in the files of A,
Brieux and F.R. Maddison.
13.
  Location: London. The Victoria and Albert Museum,
Department of Metalwork; Inventory No. M.828-1928.
Marling Gift.
  Date: The twenty-second year of the reign of JahangTr
[AD 1626-1627].
  Maker: Qa^im Muhammad ibn "Isa ibn Allahdad Astur-
labT LahurT HumayiinT.
  Metal; no seam observable; one obvious round
plug. Diameter 156 mm. Full set of constellation
figures with about 1018 stars indicated by inlaid
silver points. Ecliptic latitude circles. Ecliptic and
equator graduated by single degrees with every
fifth labeled; ecliptic repeats every 30°, while
equator is numbered continuously from vernal
equinox. Zodiacal names engraved along ecliptic;
constellations, celestial poles, and major stars la-
beled; naskhi. Signature is inscribed near south-
ern ecliptic pole: "The work of the least of the
servants {san"at aqall al-"ibdd) Qa^im Muham-
mad ibn "Isa ibn Allahdad AsturlabT LahQrT Hu-
mayunT." Near the end of Eridanus the date is
given: "Twenty-second year of the reign of Ja-
hangTr."
  Has a metal quadruped stand on circular base
with central support; horizon ring graduated by
single degrees with every fifth labeled, with the
four cardinal points, and the four intermediate
ones, engraved near outer edge; the ring is num-
bered in non-consecutive 90° segments, begin-
ning at the east-west points and terminating at
the north-south points where there are notches
for the meridian ring. Ungraduated recent me-
ridian ring. There are also notches at the east-
west points of the horizon ring which may have
been for a semicircular ring (now missing) over
the top to mark the zenith and serve as the prime
vertical.
See Figures 38 and 42 for illustrations.

14.
   Location: Patna (Bihar), India. Khuda Bakhsh Oriental
Public Library; Inventory No. 1 127.
Date:    1047 H [AD 1637-1638].
   Maker: Qa^im Muhanmiad ibn "Isa ibn Allahdad Astur-
labT LahiirT HumayunT.
  Metal with silver inlaid stars. The inscription
giving the maker reads: "The work of {san"at)
the least of the servants Qa^im Muhammad ibn
"Isa ibn Allahdad AsturlabT LahQrT HumayQnT
[in the] year 1047 H." A second inscription in
Persian on the globe reads:
The whole of this complete globe contains 1022 stars, all of
which scholars and astronomers have observed to be 48
constellations, just as has been observed also in the observa-
tory of Mirza Ulugh Beg, and we have added 3 degrees to
the position {taqwim) of each Fixed star, according to the
calculations of the scholars and wise men of this science, up
to this date, the year 1047 H.
This inscription displays elements of the eastern
form of Persian spoken outside of Persia proper,
arid probably a form of Persian called Tdjik (see
Barthold EI).
  No further information available; no photo-
graph or drawing has been published. Entry
based on published accounts.
Citations
Nadvi [1935], 627 gives the two inscriptions.
Kliiber, 5.
Wiet ([1936], 98, no. 5).
15.
Location:    Kuwait, private collection.
Date:    1049 H [AD 1639-1640].
  Maker:    Muhammad MuqTm ibn "Isa ibn Allahdad Astur-
labT HumayiinT LahurT.
  Metal; no seam observable; rectangular plug
near south ecliptic pole. Diameter unknown. Full
set of constellation figures with about 1018 stars
indicated by inalid silver points. Ecliptic latitude
circles. Stars are numbered within a constella-
tion. Major stars labeled; naskhi. Near the end of
Eridanus the inscription reads: "The work of
{san"at) of the weakest of the servants {ad"af al-
"ibdd) Muhammad MuqTm ibn "Isa ibn Allahdad
AsturlabT HumayQnT LahQrT in the year 1049 H."
No further information available. Entry based

226

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



on photograph in the files of A. Brieux and F.R.
Maddison.
16.
Location:    London.  The Victoria and Albert Museum,
Department of Metalwork; Inventory No. M. 827-1928.
Date:    1050 [H] [AD 1640-1641 ].
Maker:    Muhammad Zaman.
  Metal; hemispheres joined at equator. Diame-
ter 179 mm. Full set of constellation figures with
about 1018 stars indicated by very small inlaid
silver points. Ecliptic latitude circles; equatorial
tropic circles and ecliptic polar circles; equinoc-
tial colure. Equator and ecliptic graduated by
single degrees with every sixth labeled; ecliptic
repeats every 30°, while the equator is numbered
continuously from the vernal equinox. Solstitial
colure graduated by single degrees with every
sixth indicated by a slightly longer line. Zodiacal
names in small writing engraved along ecliptic;
constellations, equatorial and ecliptic poles and
many stars labeled; small cursive naskhi. The
inscription in Arabic between the two southern
poles reads: "The work of {san"at) the dust of the
threshold of Riza {turdb "atabat al-rizd) Muham-
mad Zaman, 1050." The reference is to the
sanctuary of the Imam Riza in Mashhad, where
the maker worked. The date is not entirely clear
and can be read as 1051 or 1050. The maker
also made a globe without constellations. No. 64.
  The quadruped metal stand, possibly contem-
porary with globe, supports the globe with
semicircular metal arcs and has an horizon ring
graduated by single degrees with every sixth
labeled and numbered continuously in a coun-
terclockwise direction; the ring has two notches,
one at 90° and one at 270° to accommodate the
meridian ring. Meridian ring is missing.
See Figure 21 for an illustration.
  Formerly in the collection of Sir Charles Mar-
ling, K.C.M.G. Exhibited in the exhibition "Sci-
ence and Technology in Islam" held at the Sci-
ence Museum, London, as part of the Festival of
the World of Islam, April-August 1976.
                       Citations
Mayer [1956], 78; reads date as 1051.
Maddison Sc Turner, 81, read date as 1051.

17.
Location:    Kuwait, private collection.
Date:    1054 [H] [AD 1644-1645].
Maker:    Muhammad Zaman.
  Metal; seam along equator. Diameter un-
known. Full set of constellations figures engraved
with about 1018 stars indicated by small inlaid
silver points. Ecliptic latitude circles. Equator and
ecliptic graduated by single degrees with every
fifth indicated by a longer line. The Persian
inscription in naskhi script reads: "Dust of the
road of life {khdk rah dn) Muhammad Zaman
1054."
  A modern pedestal stand supports two crossed
semicircular bars which hold the horizon ring.
The globe has a rod passing through the celestial
poles, which is mounted horizontally to the ho-
rizon ring. Meridian ring missing.
  No further information available. Entry based
on photographs in the files of A. Brieux and F.R.
Maddison.
18.
   Location: New York City. Columbia University, Butler
Library, David Eugene Smith Collection; No. 27-198.
Date: 1055 H [AD 1645-1646].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn "Isa ibn Allahdad AsturlabT HumayiinT LahurT.
  Metal; dark finish; no seam observable; several
plugs visible including one quite large oval plug
from the Southern Fish to the southern celestial
pole and a round one in the northern hemi-
sphere. Diameter 170 mm. Full set of constella-
tion figures engraved with about 1018 stars in-
dicated by small inlaid silver points which have
tarnished. Ecliptic latitude circles. Ecliptic and
equator graduated by single degrees with every
fifth labeled; ecliptic repeats every 30°, while
equator is labeled continuously from vernal equi-
nox. Division between the 5° intervals indicated
by dotted lines on both ecliptic and equator.
Names of zodiacal signs engraved along ecliptic;
equatorial poles, constellations and major stars
labeled; naskhi. The constellations and the indi-
vidual stars within a constellation are numbered.
Near the end of Eridanus is the Persian inscrip-
  
NUMBER 46

227



tion: "This is the work of {in san"at) Diya"" al-DTn
Muhammad ibn Qa^im Muhammad ibn "Tsa ibn
Allahdad AsturlabT HumayQnT LahurT in the year
1055 H."
  The heavy three-legged metal stand holds an
horizon ring graduated by single degrees with
every fifth labeled in non-consecutive segments
of 90°, beginning at the two points of the attach-
ment of the semicircular ring over the top of the
globe and ending at the two notches for the
meridian ring. To the underside of the horizon
ring is attached a semicircular arc for supporting
the meridian ring and globe. The semicircular
ungraduated metal arc over the top of the globe
at right angles to the horizon and meridian rings
serves to mark the zenith and prime vertical and
is hinged at one point of attachment to the ho-
rizon ring and held with a pin at the other so
that the globe can be removed or readjusted; a
small ring is attached as a handle. The meridian
ring is graduated along the outside edge by single
degrees with every fifth labeled in non-consecu-
tive segments of 90°; the inside edges of the ring
are graduated by single degrees. One half of the
ring (from zero point to zero point) is twice as
thick as the other semicircle. Under the thinner
half of the ring there rotates a thin 90° arc
graduated by single degrees which is attached at
the midpoint, or 90° label, of the thinner half-
circle. On the outside edge of the meridian ring
at one of the zero points is a protruding lip. Two
opposite 90° arcs of the ring contain holes at
every 5° interval; the other two opposite quad-
rants have holes at approximately 18°, 23°, 27°,
32°, and 37°, Thus when set into the stand with
the thicker half of the ring below the horizon
ring and the protruding lip resting on the hori-
zon ring to hold it steady, the globe can then be
rectified for five northern geographical posi-
tions. The rotating arc attached at the zenith
could be used to measure altitude. It is likely that
the holes at every 5° for the southern settings
are simply for form. The rod serving as the axis
for the globe is missing. The horizon ring has
been repaired in many places. Rings appear to
be contemporary with globe.
See Figure 15 for an illustration.

Citations
Industrial Museum of New York, 59 (exhibition catalog).
Sanford, 17 (photograph incorrectly said to be of a "Hindu
  celestial sphere ca. 1590").
Yonge, 84.
19.
   Location: Leningrad, Musee asiatique (Muzej antropolo-
gii i etnografii im. Petra Velikogo).
Date:    1057 H [AD 1647-1648].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn "Isa ibn Allahdad AsturlabT HumayunT LahurT.
  The signature on globe reads: "This is the
work {in san"at) of Diya^ al-DTn Muhammad ibn
Qa^im Muhammad ibn Tsa ibn Allahdad Astur-
labT HumayQnT LahQrT in the year 1057 H."
  No further information available; entry based
on printed description.
Citations
Khanykov, 66-69; gives the signature and a few star names;
  no physical description of globe.
Dorn [1865], 1.
Wiet [1935], 98, no. 9.
20.
   Location: London. The Victoria and Albert Museum,
Department of Metalwork; Inventory No. M.507-1888.
Date:    1060 H [AD 1650].
  Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
Mulla "Isa ibn Shaykh Allahdad AsturlabT HumayunT La-
hurT.
  Metal; no seam observable; two plugs visible,
one of a different alloy from rest of globe. Dia-
meter 113 mm. Full set of constellation figures
with about 1018 stars indicated by inlaid silver
points. Ecliptic latitude circles. Ecliptic and equa-
tor graduated by single degrees with every fifth
labeled; ecliptic repeats every 30°, while equator
is numbered continuously from the vernal equi-
nox; divisions between 5° intervals marked by
dotted lines on both ecliptic and equator. Zodia-
cal names engraved along ecliptic; equatorial
poles, constellations and many stars labeled; nas-
khi. Constellations are numbered in order. In-
scription near end of Eridanus reads: "The work
of {"amal) the least of the servants {aqall al-"ibdd)
  
228

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Diya^ al-DTn Muhammad ibn Qa^im Muhammad
Mulla Tsa ibn Shaykh Allahdad AsturlabT Hu-
mayQnT LahQrT [in the] year 1060 H.'
  The quadruped metal stand on a ring base is
non-functional in that the top ring cannot serve
as a horizon ring; not contemporary with globe.
Meridian ring missing. The holes at the equato-
rial poles of the globe have been filled in; many
of the inlaid silver points are now missing.
Citations
Casanova, 318.
Wiet [1935], 17.
Wiet [1933], 60.
Wiet [1936], 98, no. 12.
Kliiber, 2.
Harari, 5:2518.

nox. The divisions between the 5° intervals are
indicated by dotted lines on both equator and
ecliptic. Zodiacal names engraved along ecliptic;
constellations and many stars labeled; naskhi.
Constellations are numbered in order. Signature
reads: "Work of {"amal) the least of the servants
Diya" al-DTn Muhammad ..."
Stand and rings missing.
  The globe was purchased in Bombay in 1882
and was part of the India Museum which was
formerly located where now the Imperial College
of Science is in London. The entire collection of
the India Museum was transferred to the Indian
Section of the Victoria and Albert Museum. In
1974 the globe was in a traveling exhibit from
the Victoria and Albert.
  
Citations

21.
Location:    Aligarh, India, Tibbiyah College.
Date:    1064 H [AD 1653-1654].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn Mulla "Isa ibn Shaykh Allahdad AsturlabT HumayunT
LahurT.
  The inscription reads: "The work of {"amal)
the least of the servants Diya^ al-DTn Muham-
mad . . . ."
No further information available.
                       Citation
Nadvi [1935], 628 no. 4, gives inscription.
22.
   Location: London. The Victoria and Albert Museum,
Indian Section; Inventory No. 2324-1883 (I.S.).
Date:    1067 H [AD 1656-1657].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn Mulla "Isa ibn Shaykh Allahdad AsturlabT HumayunT
LahurT.
  Metal; no seam observable; one large round
plug visible. Diameter 127 mm. Full set of con-
stellation figures and about 1018 stars indicated
by inlaid silver points. Ecliptic latitude circles;
equatorial tropic circles. Ecliptic and equator
graduated by single degrees with every fifth la-
beled; ecliptic repeats every 30° while equator is
numbered continuously from the vernal equi-

Casanova, 318.
Wiet [1933], 60.
Wiet [1935], 17.
Wiet [1936], 99 no. 16.
Gunther/IIV, 1:210.
Kliiber, 2 and 5.
23.
   Location: Cardiff. Welsh Industrial and Maritime Mu-
seum, Amgueddfa Diwydiant a Mor Cymru; Inventory No.
39.573.1.
   Date: 1068 H and the thirty-second year of the reign of
Shahjahan [AD 1657-1658].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn Mulla "Isa ibn Shaykh Allahdad AsturlabT HumayunT
LahurT
  Metal; no seam observable; one visible plug;
rattles; biased in weight. Diameter 113 mm. Full
set of constellation figures with about 1018 stars
indicated by small inlaid silver points. Ecliptic
latitude circles; equatorial tropic circles and
equatorial polar circles. Equator and ecliptic
graduated by single degrees with every fifth la-
beled; ecliptic repeats every 30°, while equator
is numbered continuously from the vernal equi-
nox. Divisions between every 5° interval indi-
cated by dotted lines on both ecliptic and equa-
tor. Zodiacal names engraved along ecliptic;
equatorial poles, constellations and many stars
labeled; naskhi. Constellations are numbered; in-
  
NUMBER 46

229



dividual stars are numbered within some of the
northern constellations. The inscription around
the south equatorial polar circle reads: "The
work of {"amal) the least of the servants Diya" al-
DTn , , . , [in the] year 1068 H, the thirty-second
year of the blessed reign {julus mubdrak) of Shah
Jahan."
  The metal pedestal stand on a circular base
with two crossing semicircular arcs holding a ring
is non-functional in that the ring cannot serve as
a horizon ring; the stand was made by Newton
8c Co., 3 Fleet Street, London, The meridian
ring, which is graduated by 5° with every 30°
interval indicated by a longer line, is also proba-
bly not contemporary with the globe.
See Figures 39, 44, and 48 for illustrations.
  Bought in Algiers in 1910 and given by donor
to museum in 1939.
24.
  Location: Cairo. Museum of Islamic Art; Inventory No.
15226.
   Date: [10]68 H and thirty-second year of reign of Shah
Jahan [AD 1657-1658].
  Maker: [pi]ya^ al-DTn Muhammad ibn Qa^im Muham-
mad ibn Mulla "Isa ibn Shaykh Allahdad AsturlabT Huma-
yunT LahiirT.
  Metal; no seam observable; one rectangular
plug and one very large plug constituting nearly
one-fourth of the globe. Diameter 100 mm. Full
set of constellation figures with about 1018 stars
indicated by small inlaid silver points. Ecliptic
latitude circles; southern celestial polar circle.
Holes at both ecliptic and celestial poles. Equator
and ecliptic graduated by single degrees with
every fifth labeled; ecliptic repeats every 30°
while equator is numbered continuously from
vernal equinox. Division between every 5° inter-
val indicated by dotted line on both ecliptic and
equator. Zodiacal names engraved along ecliptic;
equatorial poles, constellations and major stars
labeled; naskhi. The inscription around the south
equatorial polar circle reads: "The work of
{"amal) the least of the servants [Dijya^ al-DTn
Muhammad .... in the year [10]68 [H] and the
thirty-second year of the blessed [reign] of Shah
Jahan." The plugs interfere with the inscription.
  
Metal horizon ring supported by a metal quad-
ruped stand, 137 mm high on ring base with
semicircular undersupport for globe. Probably
not contemporary with globe. Meridian ring
missing.
Formerly Harari Collection, No. 157.
Citations
Wiet [1933], 60.
Wiet [1935],   16-17, no. 36, gives inscription in slightly
variant form.
25.
Location:    New  Delhi,  India.  Red Fort Archaeological
Museum; Inventory No. 40-415.
Date:    1070 [H] [AD 1659-1660].
Maker:    Muhammad Salih TatahwT [sic].
  Metal; no seam or plug observed in available
photographs. Diameter 210 mm. Full set of con-
stellation figures with about 1018 stars indicated
by silver inlaid points. Ecliptic latitude circles.
Holes at both ecliptic and celestial poles. Gradu-
ation of equator and ecliptic not evident from
photographs. Equatorial and ecliptic poles, eclip-
tic latitude circles, constellations, and major stars
labeled; naskhi. Near the south ecliptic pole is the
inscription reading: "By the order of {bi-farmdy-
ish) Shaykh "Abd al-Khaliq. The work of {"amal)
Muhammad Salih TatahwT [in the] year 1070."
  The name of the maker, TatahwT (perhaps
more accurately transliterated as Tatah-wT) is
written differently on this globe than on globe
No. 29, where it is spelled TatawT. By writing it
as Tatah-wT, though it is unusual orthography,
the maker probably wished to indicate the pro-
nunciation, for one might be inclined to pron-
ounce it as TatwT with the spelling on globe No.
29. It is unlikely that the maker was from Tatta
in the Sind (the delta of the Indus river) since
the name of the town is written with slightly
different letters and should more accurately be
transliterated as Thattha. The globe itself dis-
plays basically Mughal design of constellation
figures and exhibits considerable precision in
design.
  In the one published photograph a rod passes
through the globe by which it is mounted hori-
  
230

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



zontally to a stand that is 660 mm high. The
stand is clearly non-functional and probably not
contemporary with globe.
  Entry based on the published account and two
photographs in the files of A. Brieux and F.R.
Maddison.
Citations
Dhama, 143-144, plate LVIII.
Stone, 160.
26.
   Location: Berlin. Staatliche Museum zu Berlin, DDR;
Inventory No. 1101.
Date:    1071 H [AD 1660-1661 ].
  Maker: Diya^ al-DTn Muhammad ibn Muhammad
[Qa^im] Muhammad ibn Mulla "Isa ibn Shaykh Allahdad
AsturlabT HumayCim LahurT.
  Metal; no seam observable; one plug visible.
Diameter 94 mm; wall thickness ca. 3 mm. Full
set of constellation figures with about 1018 stars
indicated by inlaid silver points. Ecliptic latitude
circles; equatorial tropic and polar circles. Eclip-
tic and equator graduated by single degrees with
every sixth labeled; ecliptic repeats every 30°,
while equator is numbered continuously from
the vernal equinox. Divisions between 6° inter-
vals indicated by dotted lines. Names of zodiacal
houses engraved along ecliptic; constellations,
equatorial poles, and some stars labeled; naskhi.
Constellations numbered in order. Inscription
near south equatorial pole begins "The work of
{"amal) the least of the servants Diya^ al-DTn
Muhammad . . . ." The word Qa^im has been
defaced and the preceding Muhammad added in
its place.
See Figure 45 for an illustration.
Stand and rings are missing.
Citations
Kliiber. A lengthy study of the globe with two photographs
  and several lithograph reproductions.
Nadvi [1935], 629, no. 5.
27.
   Location: Edinburgh. Royal Scottish Museum, Depart-
ment of Technology; Inventory No. 1890-331 (Richard
Collection).

Date:    1074 H [AD 1663-1664].
   Maker: Diya" al-DTn Muhammad ibn Qa^'im Muhammad
ibn Mulla "Tsa ibn Shaykh Allahdad AsturlabT HumayunT
LahurT.
  Metal; no seam observable; three plugs visible.
Diameter 142 mm. Full set of constellation fig-
ures with about 1018 stars indicated by inlaid
silver points. Ecliptic latitude circles; equatorial
polar circles; equatorial tropic circles lightly en-
graved. Ecliptic and equator graduated by single
degrees with every fifth labeled; ecliptic repeats
every 30° while equator is numbered continu-
ously from the vernal equinox. Divisions between
5° intervals indicated by dotted lines. Names of
zodiacal signs engraved along ecliptic; equatorial
poles, constellations, and many stars labeled; nas-
khi. Constellations are numbered in order. In-
scription around southern equatorial pole begins:
"The work of {"amal) the least of the servants
Piya^ al-DTn Muhammad . . . ."
  The quadruped metal stand on circular base
(height 170 mm) with central support has an
horizon ring graduated by single degrees with
every fifth labeled and bearing the four cardinal
points near outer edge of ring. Horizon ring is
numbered in non-consecutive segments of 90°
beginning at the east-west points and ending at
the north-south points. The ring has notches at
the north-south points for the meridian ring and
also notches at the east-west points, possibly for
a semicircular ring over the top (now missing) to
mark the zenith and prime vertical. Ungraduated
meridian ring, but with holes for adjustment for
six different terrestrial latitudes. Horizon ring
appears to be of same alloy as globe and hence
possibly contemporary with globe.
See Figures 16 and 35 for illustrations.
   Bought in 1890 from F. Du Cane Godman
after being shown in the Richard Collection at
the Paris Exhibition of 1889. In 1931 it was
exhibited at the International Exhibition of Per-
sian Art at the Royal Academy of Arts in Lon-
don. Exhibited at the exhibition "Science and
Technology in Islam" held at the Science Mu-
seum, London, as part of the Festival of the
World of Islam, April-August 1976. Displayed
at the exhibition "Science in India" held at the
   
NUMBER 46

231



Science Museum, London, March-August 1982.
Commonly called the "Barlow Globe."
Citations
Photograph, Illustrated London News, December, 1930, p.
1025.
Exhibition catalog of the "International Exhibition of Per-
sian Art at the Royal Academy of Arts, London, 1931."
In the first edition it was stated that it belonged to Mr.
Barlow (no. 309A), hence the name "Barlow globe.' In
second edition, described on p. 175 (no. 309B); and in
third edition, on p. 193 (no. 309B).
Wiet [1933], 59-60, no. 58, and 139, where it is incorrectly
said to be part of the Harari collection.
Wiet [1935], 17.
Wiet [1936], 99 no. 24.
Nadvi [1935], 629 where it is stated to be an astrolabe.
Maddison 8c Turner, 82.
Anderson, 34, no. 124
28.

been rolled. The diameter of the probe is greater
than that of the holes in the globe at the poles.
Therefore, it must have been inserted, along
with the papers, before the final plug was put in
place on the globe.
  The metal three-legged heavy stand has an
horizon ring graduated by single degrees and
labeled every 5° in non-consecutive segments of
90° converging at the two notches for the merid-
ian ring; may be contemporary with the globe
and is quite similar to that of globe No. 18. A
semicircular arc attached to the underside of the
horizon ring supports the ungraduated meridian
ring which is a later replacement.
See Figures 40, 43, and 47 for illustrations.
Formerly in the M. Henri Michel Collection.
                       Citation
Josten, 20, no. 25; and photograph in plate VII.
                     

                     
   Location: Oxford. Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/25.
Date: 1074 H [AD 1663-1664].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn Mulla "Isa ibn Shaykh Allahdad AsturlabT HumayQnT
LahurT.
  Metal; no seam observable; two plugs ob-
served; has a rather dull black finish. Diameter
175 mm; weight 2097.9 grams, including merid-
ian ring. Full set of constellation figures with
about 1018 stars indicated by inlaid silver points.
Ecliptic latitude circles; equatorial polar circles.
Equator and ecliptic graduated by single degrees
with every fifth labeled; ecliptic repeats every
30° while equator is numbered continously from
the vernal equinox. Division between every 5°
interval indicated by dotted lines. Zodiacal names
engraved along ecliptic; equatorial poles, con-
stellations and many stars labeled; naskhi. Con-
stellations numbered in order. Inscription
around south equatorial polar circle begins; "The
work of {"amal) the least of the servants Diya^ al-
DTn Muhammad . . . ."
  Inside the sphere there are three rolled-up
pieces of paper with writing on them (amulets ?)
which have been sewn together with twine, and
also a rather long heavy metal probe, pencil
shaped, around which the papers appear to have

29.
Location:    London. Private collection.
Date:    1074 [H] [AD 1663-1664].
Maker:    Muhammad Salih TatawT.
  Metal; no seam observable; large figure-eight
plug near north ecliptic pole. Diameter un-
known. Full set of constellation figures with
about 1018 stars indicated by inlaid silver points.
Ecliptic latitude circles; equatorial tropic circles.
Ecliptic graduated by single degrees with every
fifth labeled in both Arabic and in Sanskrit (De-
vanagarT characters); both sets of numerals are
written alongside each other and repeat every
30°. The equator is graduated by single degrees
with every fifth labeled in Arabic {abjad numer-
als) on one side and every sixth labeled in San-
skrit (DevenagarT characters) on the opposite
side; numbering is continuous from the vernal
equinox. Names of zodiacal houses written along
ecliptic in fairly small script, not extended to the
width of each house, in both Sanskrit and Arabic,
using the word burj in Arabic for zodiacal house.
Ecliptic latitude circles, tropics, ecliptic poles, the
mayl kulli, constellations, and major stars labeled
in Arabic, and the constellation names and sig-
nature inscription are also inscribed in Sanskrit,
  
232

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



The script is very precise and clear; naskhi and
DevanagarT, The constellations are carefully and
attractively executed in Mughal style; precision
and uniformity good. Inscription beneath South-
ern Fish reads: "The work of {"amal) Muhammad
Salih TatawT [in the] year 1074," There is also a
Sanskrit inscription. See globe No. 25 for an-
other globe by this maker, on which he spells his
name with a slight variation.
  Metal quadruped stand with spindle shaped
legs is on a base formed of two cross bars. Hori-
zon ring is graduated by single degrees with
every sixth labeled in DevanagarT characters in
non-consecutive segments of 90° converging at
the two notches for the meridian ring. Attached
to the underside of the horizon ring is a semicir-
cular bar, with a finial at the bottom, which
supports a modern ungraduated meridian ring.
The horizon ring and stand appear to be contem-
porary with globe.
See Figure 18 for an illustration.
Formerly in collection of Alain Brieux of Paris.
  Entry based on photographs in the files of A.
Brieux and F.R. Maddison.
30.
Location:    Paris. Private collection.
Date:    1090 H [AD 1679-1680].
  Maker:    Diya^ al-DTn Muhammad ibn Mulla Qa^im Mu-
hammad ibn Hafiz "Isa ibn Shaykh Allah-dad HumayiinT.
  This is not really a globe, but an open, cut-out
sphere consisting of two raised or hammered
hemispheres in which spaces between the con-
stellation outlines and great circles have been cut
out or worked a jour. The hemispheres are easily
separated and meet along the circle of the equa-
tor. Metal, covered with gilt paint and varnish;
has been repaired in places. The gilt finish, cov-
ered with varnish, is also inside the sphere where
it has not worn off, as it has on many areas of
the outside surface. Diameter 164 mm. Full set
of constellation figures, with details engraved on
the figures and backgrounds cut out. About 1018
stars indicated by holes perforating the sphere,
some of which seem to be filled with glass or
mica. Ecliptic latitude circles each about 6 mm

wide, and equatorial and ecliptic polar circles.
Holes at both ecliptic and equatorial poles. Eclip-
tic and equator, each about 15 mm wide, are
graduated by single degrees with every fifth in-
dicated by a dotted line and labeled; equator
numbered continuously from vernal equinox;
ecliptic repeats every 30°, Zodiacal names en-
graved along ecliptic; each ecliptic latitude circle,
the poles, equatorial polar circles, constellations,
and major stars labeled in a mixture of Arabic
and Persian; naskhi. The Persian inscription writ-
ten in five parts around the two southern polar
circles reads as follows:
In the reign of His Majesty MuhyT al-DTn Muhammad Au-
rangzTb Bahadar "AlamgTr, this sphere was invented {ikh-
tird"l) by order of {bi-farmdyish) this excellent nawab, an
exalted and publicly esteemed sovereign and learned
builder. This invented {ikhtird"l) terrestrial-celestial sphere
is the work of {"amal) the weakest of the servants {ada"fal-
'^ibdd) Diya^ al-DTn Muhammad ibn Mulla Qa^im Muham-
mad ibn Hafiz "Isa ibn Shaykh Allah-dad HumayunT in the
year 1090 H.
  This inscription implies that this sphere, of
special design for the Mughal emperor "AlamgTr
I, was both a terrestrial and celestial sphere—
hence we may suppose that at one time a small
terrestrial globe was placed in it. By having the
stars as well as the spaces between the constella-
tions and circles cut out, a sufficient amount of
light would be introduced into the sphere to
allow inspection of the enclosed terrestrial globe.
Thus this celestial sphere seems to have been
part of a demonstrational armillary sphere and
as such is of a unique design among extant Islam-
icate instruments. The signature of the maker
shows slight differences from his signatures on
other globes, notably the different orthography
used in writing Allah-dad, the introduction of
the name Hafiz, and the omission of LahQrT.
Stand and rings missing.
See Figure 17 for an illustration.
  Formerly in the collection of Leonard Linton
of Point Lookout, New York, before being sold
in Paris in 1980.
Citation
Collection of Leonard Linton. Sale: Nouveau Drouot, 9 rue
Drouot, 9-10 October 1980. Commissaires-Priseurs: E.

NUMBER 46

233



Libert and A. Castor. Expert pres les Douanes: A. Brieux
(sale catalog). Item No. 185, p. 131, and photograph p.
130.
31.
  Location: Leningrad. Lomonosov Museum (Muzej M.V.
Lomonosova).
  Date: Begun the eleventh day of Shawwal, 1112 H [21
March AD 1701] with equivalent date given in the Era of the
Flood as well as the Coptic, Alexandrian, Yazdijird and
Malikshah eras, and completed in Safar, 1113 H [8 July to 5
August AD 1701].
Maker:    Ridwan.
  Metal; hemispheres. Diameter 183 mm. Full
set of constellation figures; about 1018 stars in-
dicated by inlaid silver points. Considerable
traces of what was once complete gilding are
visible. Ecliptic latitude circles; equinoctial col-
ure. Ecliptic and equator graduated by single
degrees marked by dots; equator has every fifth
degree indicated by a line and labeled with stan-
dard (not abjad) numerals, numbering continu-
ously from the vernal equinox; ecliptic has every
sixth degree indicated by a line with numbering
in standard numerals repeating every 30°. Zo-
diacal names engraved along ecliptic; equatorial
and ecliptic poles, constellations and many stars
labeled; naskhi. Constellations are numbered
with standard numerals in order; some individual
stars are numbered with abjad numerals. Under
Eridanus there is a lengthy Arabic inscription,
ending with a quotation from the Qur^an
111,167, which reads:
The construction {"amal) of this globe was carried out {ku-
mila) with the engraving on it of the fixed stars—whose
number is 1022—after the adjustment of their longitudes
according to the principles of Ulugh Beg, author of the
recent observations at Samarqand, by the order of {bi-rasm)
Mawlana Hasan EfendT, one-time RuznamijT of Cairo, al-
KhalwatT al-DamurdashT, by {"aid yad) the weakest of the
servants and most in need of God's mercy on the day of
resurrection, humble in his deficiencies and shortcomings,
Ridwan, the unworthy. That was at the time of the first day
of the 4803th year of the Era of the Flood {tufanlyah)—that
is, the day the great light [sun] passed into the House of
Aries—being Monday the eleventh day of the month of
Shawwal in the year 1112 H, and that is the fourteenth of
Barahat [seventh month] of the 1417th Coptic {qubtiyah)

year, and the tenth of Adar of the year 2012 of the Greek
Alexandrian era, and the twentieth of Tirmah of the year
1070 of the Yazdijird era, and the first day of FarwardTn
[first month] of the year 623 of the Malikshah era. The
completion of the remainder of the globe, with the help of
God, may He be exalted!, was in the month of Safar al-Khayr
of the year 1113 H. God is sufficient for us; an excellent
Guardian is He.
  The globe rests in the probably original three-
legged, S-curve, metal stand 324 mm in height
with a central pedestal stand supporting the me-
ridian ring. The meridian ring is graduated by
single degrees with every fifth indicated by a
longer line and labeled in abjad numerals in non-
consecutive segments of 90° beginning at the
poles and terminating at the points opposite the
celestial equator. The horizon ring has four con-
centric circles. The inner circle is graduated by
single degrees with every fifth indicated by two
longer lines and every tenth by an even longer
line. The single line indicating the 10° intervals
extends to the outermost edge of the horizon
ring, through all four concentric circles, while
the double line indicating the 5 ° division passes
through the two innermost circles. The second
concentric circle has the names of the four car-
dinal points and labels each 5 ° interval with abjad
letters in non-consecutive segments of 90° begin-
ning at the east-west points and ending at the
north-south points where there are the notches
for the meridian ring. The third circle bears the
names of the 12 zodiacal signs allowing 30° for
each sign, Reading counterclockwise from east
there are Aries, Taurus, Gemini, (N) Cancer,
Leo, Virgo, (W) Libra, Scorpio, Sagittarius, (S)
Capricorn, Aquarius, Pisces. The outermost cir-
cle bears the names of 104 localities allowing two
or three names per 10° interval. Reading clock-
wise from the east point they are: Ceylon (Sir-
nadTb), Santarah, the Silver Island, SindQn (the
two Sinds), the Indian Island, the Black Sea, al-
Bahrayn, (island of) Socotra, Sa"adah, Qalhat,
Sharjah, Bern Hilal, Hadrmawt, MaqdQshah, al-
Zinj, Sofalah (lit, Suqalah), Marmarah, Sana"a,
Bihamah, "Arafat, al-Ta^ifah, [S], Aden, ZabTd
(in S, Arabia), Bilad al-Za^ila^ Bern Sha'TDah, the
city of Qus, "Ashwah [?], the Yemen, Mara, Saba,
Zafar, Traq, Jabal al-Qamar (Mountain of the
  
234

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Moon, Mount Kilimanjaro), Bilad al-SQdan (the
Sudan), Mahrah, Khasah, al-Mihrjan, Mazhar al-
NTl, TikrQr, Bilad al-Barbar (land of the Berbers),
KQkQ, "Adiyan [?], Danqalah, Zaghaw, Jiddah,
Habashah (Abyssinia), al-Qusayr, [W], Marakash,
Fas (Fes), Tilimsan (Telemsan), Tuwan, TQnis,
Qayruwan (Kairouan in Tunisia), Tarablus al-
GharTb (Tripoli in North Africa), Barqah,
Aujlah, AsyQt, FayyQm, RashTd (Rosette), Dami-
yat (Damiette), Misr (Cairo-Fustat), Mahallah, Is-
tanapQl (Istanbul), Adrinah (Adrianopel), "Aqa-
bah, Muwaylih, TQr (Sinai), Ba"albek, Sayda, Da-
mashq al-Sham (Damascus), Hama, Hams, Wa-
jah, Halab (Aleppo), Malatiyah, Badr, Shatt (Ti-
gris), Rabigh, Masajid "A^ishat al-MadTnah al-
MunQrah (the mosques of Medina?), [N], al-
Mawsil, Babal (Babel), Shiwan, Hit, Baghdad,
KQfa, al-Qadisiyah, QazwTn, Bulghar, Khurasan,
"Abadan (S. of Basra), Basra, NisabQr, KazirQn,
Balkh, Bukhara, Samarqand, SabQr, Isfahan,
Sadd Yajuj (wall or mountain of Yajuj), al-Hind
(India), Kabul, HurmQz, and Hurmuz.
  Formerly in the collection of Clot Bey; given
to library in 1859.
Entry based on published accounts.
Citations
Dorn [1865], 31-44; a thorough description of globe and
inscriptions and star names; no reproduction of constel-
lation outlines.
Stevenson,1:32.
Mayer [1956], 81-82.
Muris Sc Saarmann, 38.
Tchenakal, 293.
Czenakal, 57 and photograph figure 3.
32.
Location:    Oxford.  Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/26.
Dated:    [ 1 ] 121 [H] [AD 1709-1710].
Maker:    Unsigned.
  Metal; no seam observable nor felt with probe
inside; no plugs observed; no visual inspection of
inside was possible; some sand and pebbles inside,
possibly remaining from casting; dark finish. Di-
ameter 150 mm. Full set of constellation figures
with more that 1018 stars indicated by inlaid
silver points surrounded by engraved circles.

Ecliptic latitude circles. Equator and ecliptic
graduated by intervals equivalent to 5° with each
labeled; ecliptic repeats every 30° and equator is
numbered continuously from vernal equinox.
Zodiacal names are not engraved along ecliptic.
Constellations and some stars labeled; naskhi.
Stars are arranged in a decorative manner and
are so poorly positioned on globe that it is im-
possible to determine epoch for which it was
intended. The lines representing the ecliptic and
equator are wobbly and the constellation figures
poorly drawn and placed. A technically non-
functional and inaccurate globe, probably Indo-
Persian in origin.
  A decorative quadruped metal stand on a ring
base supports a horizon ring graduated by two
degrees with every sixth labeled in non-consecu-
tive segments of 90° converging at the notches
for the meridian ring. The legs of the stand have
at the tops and bottoms an elongated spiral de-
sign while the middle sections are engraved with
human figures and floral motifs; the ring base
has a braided design. To the underside of the
horizon ring beneath the zero points is a semicir-
cular band for supporting the meridian ring and
globe. This band bears a two-part inscription in
Persian reading, "By order of {bar hasb; or bar
khabir, for the knowledgeable) the learned, hon-
orable, gracious master, the prince Muzaffar
"Ubayd al-RahTm, the completion occurred in
the year [ 1 ] 121." This is by no means a definitive
translation. Many of the words are difficult to
make out with certainty and can be read in
several ways.
  The meridian ring is graduated on both sides
by 2° intervals with every sixth degree labeled
in non-consecutive segments of 90° beginning at
the points of attachment to the globe; its outside
edge is inlaid with a row of silver points. The
stand and rings are contemporary with the globe.
  Compare this globe with globes Nos, 45 and
46, which display similarities of design and exe-
cution.
Formerly in collection of M, Henri Michel.
Citations
Michel [1938], photograph in plate VI and on cover.
Josten, 20, no. 26.

NUMBER 46

235



Location:    Present location unknown.
Date:    AD 1842, 1258 H, and 1899 Vikrama Era.
Maker:    Lalah Balhumal LahurT.
  Metal; no seam observable; one large plug near
south poles and several smaller plugs observable.
Diameter 330 mm. Full set of constellation fig-
ures with about 1018 stars indicated by inlaid
silver points. Ecliptic latitude circles; six meridian
circles intersecting equator at 24°, 54°, 84°,
114°, 144°, 174°, 204°, 234°, 264°, 294°,
324°, 254° (there is no meridian circle repre-
senting the equinoctial colure); equatorial trop-
ics; two lesser circles at 12° and 20° from equa-
tor on both sides of equator; equatorial polar
circles. Ecliptic and equator graduated by single
degrees with every sixth degree indicated by
longer line, and labeled; ecliptic repeats every
30°; equator labeled each 6° segment with one
numeral (1, 2, 3, 4, . . . 60) numbering continu-
ously from vernal equinox. Zodiacal names en-
graved along ecliptic; constellations labeled and
numbered; major stars labeled and all stars num-
bered within a constellation; naskhi, tending to-
ward ta"liq. There are traces evident of a very
light marking of the great circles which were
then incised more deeply over the first markings;
particularly evident at the crossings. Graduations
and circles precisely executed. Near south equa-
torial polar circle is a lengthy inscription in Per-
sian with small floral decorations interspersed
among the eight lines:
The celestial globe encompasses the visible nature of the
sphere of the ecliptic and 10 great circles, among which are
the horizon circle [ring], meridian circle [ring], zenith ring
{dd^irat sumut) and altitude ring {dd^irat irtifd")—each of
these four surrounding the globe itself—and the celestial
equator, ecliptic, ecliptic latitude circles {dawd^ir '^urud),
solstitial colure {marah bi-aqtdb arba"ah), lacking [the circle
of] inclination*, and declination circles {dawd^ir muyul, me-
ridians); these six circles move with the body of the globe,
drawn and designed over 48 constellations—i.e., 21 north,
15 .south, and 12 constellations which are along the main
part of the zodiac which exists in the middle—and 1022
observable stars which are indicated by silver nails, resulting
in a summary of the principles relating to time, place and
direction, which are called the Three Principles. By order
of the exalted benefactor and wealthy (gunjiir) Singh Sahib,
benefactor of [his] time, unique in [his] lifetime, the end of

[his] epoch, Nihal Singh Sahib Bahadur AhluwalTyah,f who
in all sciences, especially the science of mathematics—which
is to say astronomy and geometry—is incomparable and
unique in the world, through the efforts of Lalah Balhiimal
LahiirT, astronomer, geometer, and astrologer, who is one
of those attending the honorable Singh $ahib mentioned
before, and who was paid a bonus of 10 ashrafi'^ [in addition
to] 480 ashrafi on completion of the design, an object of
special favours, it was made in 1899 in the Vikrama Era§
corresponding to 1258 H, which is equal to AD 1842. It was
placed at the capital Kapiirthalah.
A three-legged metal stand supports the horizon
ring having attached to the underside a semicir-
cular arc which holds the nadir of the meridian
ring. The S-curve legs on feet are attached to the
horizon ring by winged bolts. Over the top of
the globe, at right angles to the meridian and
horizon rings, is a semicircular arc attached to
the horizon ring, corresponding to the prime
vertical and indicating the zenith; the outer edges
and two sides are decorated with an engraved
floral motif, as are also the three legs. The out-
side edge of the horizon ring is engraved with an
acanthus leaf pattern. The horizon ring is grad-
uated by single degrees with every sixth marked
by a longer line and labeled in non-consecutive
segments of 90° beginning at the decorated up-
per semicircular arc (zenith ring) and ending at
the notches holding the meridian ring. The me-
ridian ring is attached to the equatorial poles of
the globe by two screws. One half of the meridian
ring is nearly flush against the surface of the
globe, while the other half (from pole to pole) is
cut away so as to permit a small slide (6° in width)
attached to it to move along the 180° arc, no
doubt for recording and comparing coordinates.
The slide itself is graduated on its edge by single
degrees. The half of the meridian ring with the
slide is graduated by single degrees and the other
half by single degrees with every sixth indicated
by a longer line. The outside edge of the merid-
ian ring is graduated into 6° intervals with each
labeled 6, 12, . . . in non-consecutive segments of
90° beginning at the poles and converging at the
points alongside the celestial equator. The stand
and rings are probably contemporary with the
globe. The design of the stand, rings, and globe
itself are remarkably similar to the one known

236

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



globe entirely in Sanskrit (No. 54) which is un-
signed and undated. The globe is also strikingly
similar to the anonymous globe No. 90 (see No.
127 in the Addendum to this catalog, for another
globe made in the same year by Lalah Balhumal).
See Figure 24 for an illustration.
  Formerly in the collection of the Maharajah of
KapQrthalah; offered for sale in Geneva in 1976.
  Entry based on photographs in the files of the
Whipple Museum of the History of Science, the
Museum of the History of Science at the Univer-
sity of Oxford, and the files of A. Brieux and
F.R, Maddison,
Citations
Galerie Genevoise horologerie ancienne, 9 rue de la Corra-
terie, Vente aux encheves. III, Expert: O. Patrizzi, 3
October 1976 (sale catalog). Item no. 229, where it is
incorrectly stated to be made in 1048 H [AD 1628]; pho-
tograph, plate 36.
   * It reads simply "^adlm al-mayl (lacking inclination); the
meaning is obscure. Perhaps the maker intends to say that
the globe is lacking the equinoctial colure, which is called
dd"irat al-mayl, thus reading it ''adlm dd^irat al-mayl.
   ■]" AhluwalTyah is clearly intended, although the engraver
has mistakenly written al-HuwalTyah.
   J The text speaks ofgulddr ashrafi; ashraflme gold coins,
and gulddr meaning "decorated with fiowers" must refer to
a coin of a particular fioriate design.
   § An era by tradition founded in 58 BC by King Vikra-
maditya of Ujjaim to commemorate a victory, but thought
by many to have been founded by Azes I. The era is still
current in India today.
Anonymous and Undated
34.
Location: Paris. Bibliotheque Nationale, Departement des
C.artes et Plans; Inventory No. Ge. A. 325.
Estimated Date:    AD 1080 (not inscribed).
Maker:    No signature.
  Metal; hemispheres joined at equator. Diame-
ter 190 mm. Full set of constellation figures with
1004 stars indicated by engraved circles having
a dot in the center. Ecliptic latitude circles; small
circles mark ecliptic poles. Ecliptic and equator
graduated by single degrees with every fifth la-
beled; ecliptic repeats every 30°, equator is la-
beled continuously from vernal equinox. Zodia-

cal names engraved along ecliptic; equatorial
poles, constellations, and major stars labeled;
MaghribT KQfic.
  The style of the nude, rubbery figures so
closely resembles the globe No. 1 by IbrahTm ibn
Sa"Td al-SahIT made in Spain in AD 1080, that it
might well be attributed to the same maker. The
difference in longitude of the stars corresponds,
according to Destombes, to the year AD 1067,
slightly earlier than globe No. 1.
  It has a curved quadruped stand meeting to
form a pedestal support for the two semicircular
arcs holding the horizon ring. Horizon ring grad-
uated by single degrees with every fifth labeled
in Kufic script; numbering of graduations is in
non-consecutive segments of 90° converging at
notches for meridian ring. No meridian ring.
  Acquired in 1936; formerly in collection of D.
Schiepati of Milan.
Citations
Schiepati. Asserts it was made in 463 H/AD 1070.
Sedillot [1841], 115-141. Lists all the constellations and
numbers of stars in each; asserts it is of thirteenth-century
Egyptian origin.
Jomard, no. 13 and no. 14; includes two plates of litho-
graphs.
Vallee, 48, no. 304.
Stevenson, 1:31.
Fiorini, 37.
Destombes [1957], 318, no. I.
Destombes [1958] 305-306.
Muris 8c Saarman, 36.
35.
Location: Paris. Private collection of Marcel Destombes.
Estimated Date: AD 1309-1315 (not inscribed).
Maker: Unsigned.
  Metal; very clear seam along ecliptic with lip
of lower hemisphere over which upper hemi-
sphere fits. Dark-colored surface, slightly dented.
Diameter 209 mm. Full set of constellation fig-
ures with 930 stars indicated by silver inlaid
points, each star numbered within the constella-
tion. Square holes at both ecliptic and equatorial
poles. Ecliptic latitude circles. Ecliptic and equa-
tor graduated by single degrees, with every fifth
indicated by a longer line and labeled; the equa-
tor beginning at the vernal equinox has three
  
NUMBER 46

237



segments of 100° and one of 60° while ecliptic
repeats every 30°. Zodiacal names engraved
along ecliptic; celestial poles, constellations and
145 stars labeled; Kufic.
  The iconography of the figures suggests Ilk-
hanid design, and the style is analogous to that
of globe No. 5. After studying the differences in
longitude of 800 stars, Destombes concludes that
this globe is quite close in date to No. 5, suggest-
ing a date of about five years later, (AD 1309-
1315), and a place of origin of Maragha, Sul-
tanye, or Tabriz. This globe, as well as the other
globes produced before 1500, will be analyzed
in detail in a forthcoming publication by Marcel
Destombes.
Modern pedestal stand. No rings.
  Acquired in 1953. Formerly in Chadenat Col-
lection, no. 6804.
Citations
Destombes [1957], 320. no. 7.
Destombes [1958], 307, note 2.
Muris Sc Saarmann, 37-38.
36.
  Location: London (Greenwich). National Maritime Mu-
seum, Department of Navigation; Inventory No. G.141.
36..302.
Date: Not inscribed.
Maker: Unsigned.
  Painted wood, gilt ground with black, white
and red oil paint; lacquered. Diameter 165 mm.
Full set of constellation figures outlined in black;
animals are filled in with red dots, while human
figures wear striped clothing. About 1018 stars
indicated by indentations filled in with white
paint. Ecliptic latitude circles; ecliptic polar cir-
cles; equatorial polar and tropic circles in red.
Ecliptic and equator are indicated by two close
parallel red circles and are graduated by single
degrees, each indicated by a dot, with every third
degree indicated by a line. Between the equator
and each tropic there are two additional red
parallel, labeled circles: the tropic {maddr) of
Gemini and Leo and the tropic of Taurus and
Virgo about 4 ° and 12°, respectively, from the
Tropic of Cancer, while at 4° from the Tropic
of Capricorn is a tropic of Aquarius and Sagittar-

ius and about 12° away is the tropic of Scorpio
and Pisces, Zodiacal names written along ecliptic;
major stars are labeled in black; naskhi. Stars
seem well placed and their positions relative to
the equinoxes conform to those of the early
seventeenth century. The representation of the
constellation figures differs considerably from
those on other known globes. Iconography as
well as calligraphy suggests the eastern borders
of Persia, possibly Bukhara in Afghanistan or
across the Oxus in Samarqand as a point of
origin.
  The metal pedestal stand supporting two sem-
icircular arcs to which the horizon ring is at-
tached is recent. The horizon ring is graduated
by two degrees with every tenth labeled in non-
consecutive segments of 90°, converging at the
two notches which hold the meridian ring. Me-
ridian ring ungraduated.
See Figure 19 for an illustration.
  Exhibited at the exhibition "Science and Tech-
nology in Islam" held at the Science Museum,
London, as part of the Festival of the World of
Islam, April-August 1976.
Citation
Maddison Sc Turner, 84.
37.
Location:    Bombay.  Cama Oriental  Institute,   136B Sa-
macher Margl., Bombay, Fort.
Date:    Not given.
Maker:    Unsigned, but bears patron's name.
   Metal; seamless with large plug near south
poles which interferes slightly with the inscrip-
tion giving patron. Diameter 203 mm. Full set
of constellation figures; stars indicated by inlaid
silver points. Ecliptic latitude circles; equatorial
polar circles. Equator and ecliptic graduated at
2° intervals. Zodiacal houses engraved along
ecliptic; constellations, major stars, and equato-
rial poles labeled; naskhi tending toward ta"liq.
Dress of the human constellation figures does
not appear to be Mughal but rather reflects other
Indian influences. The name of the patron for
whom the globe was executed is given in an
inscription which reads: "The one [who is] in
need of God, AbQ al-Qasim, son of our master
   
238

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



and teacher, the paragon of his age and unique
in his time "^Abd al-Rahman ibn Hasan, requested
it be made and figured for him. May God bestow
much mercy on them both."
  Thus one Abu al-Qasim ibn "Abd al-Rahman
ibn Hasan requested that the globe be executed.
Nothing is known of either the son nor the
learned father.
  The globe is supported by a stand, details
unknown.
  No further information available; no photo-
graph published. Information on inscription and
general appearance furnished by Dr, Ismail K,
Poonawala of the Department of Near Eastern
Languages, UCLA, who was kind enough to
inspect the globe for me.
Citations
Rehatsek, 329. Gives the inscription.
Mayer [1956], 33. Gives the name Abu al-Qasim ibn "Abd
al-Rahman ibn Hasan as the maker rather than the patron.
38.
  Location: Washington, D.C. Smithsonian Institution, Na-
tional Museum of American History, Division of Physical
Sciences; Inventory No. NMAH 330,781.
  Metal; seamless with many plugs observed; cast
by cire perdue. Diameter 217 mm; weight 2976.7
grams. Full set of constellation figures; 1019 stars
indicated in inlaid silver points. Ecliptic latitude
circles. Appears incomplete; northern equatorial
tropic circle not correctly placed and only par-
tially engraved. Ecliptic and equator graduated
by '/2° intervals with every fifth full degree la-
beled; ecliptic labeled continuously from vernal
equinox; equator numbered in two 180° seg-
ments. Zodiacal names engraved along ecliptic;
names of the 28 lunar mansions also engraved
along ecliptic on opposite side from zodiacal
names and at marked intervals of approximately
13° Constellations and many stars labeled; nas-
khi.
Stand and rings missing.
See Figures 49-68 and 70-88 for illustrations.
  Acquired by museum in 1974; formerly in
collection of Ernst Nagel of San Francisco.

Citations
Previously unpublished; this globe is the focus of the present
monograph.
39.
Location:    Present location unknown.
  Metal; no seam observable; several small plugs
visible. Diameter 115 mm. Full set of constella-
tion figures with about 1018 stars indicated by
inlaid silver points. Ecliptic latitude circles. Eclip-
tic and equator graduated by 5° intervals (with
no single degree graduation); ecliptic repeats
numbering every 30°, while equator is num-
bered in four segments of 90° each. Zodiacal
names engraved along ecliptic. Constellations
and major stars labeled; naskhi This is definitely
a product of the seventeenth-century Lahore
workshop. Star positions relative to equinoxes
are same as those of Lahore workshop. The globe
closely resembles globe No. 23 made by Diya^al-
DTn Muhammad in 1068 H/AD 1657-1658, al-
though there are certain characteristics, such as
the depiction of Argo and the dress of Auriga,
which are reminiscent of the work of the father
Qa^im Muhammad as seen on globes No. 11 and
13.
  Ungraduated meridian ring. Later base not
illustrated in published photograph.
  Entry based on sales catalog when offered for
sale in London by Spink 8c Son in 1980.
Citation
Islamic Art from India (sale catalog). Spink Sc Son. 24 April-
10 May 1980. Item 5, with photograph.
40.
  Location: (Chicago, Illinois. Adler Planetarium and Astro-
nomical Museum, Mensing Collection; Inventory No. Ml4.
  Metal; no seam observable; several plugs visi-
ble including one large oblong one; rattles;
clearly polished on lathe. Diameter 169 mm;
weight 1835.7 grams. Full set of constellation
figures with about 1018 stars indicated by inlaid
silver points. Ecliptic latitude circles. Ecliptic and
equator graduated by single degrees with every
  
NUMBER 46

239



fifth labeled; ecliptic repeats every 30°; equator
is numbered continuously from the vernal equi-
nox. Zodiacal names engraved along ecliptic;
equatorial and ecliptic poles, constellations and
many stars labeled; naskhi. Has several small dark
engraved circular patches ringed by dark dots
(one in Serpentarius and three in Leo) whose
purpose is not known. Clearly two different en-
gravers worked on globe, for Bootes, Hercules,
and the Northern Crown are rather crudely
drawn and lightly incised with some parts re-
drawn, while the other figures are well executed
and deeply engraved. Star positions relative to
the equinoxes conform to those of the seven-
teenth-century Lahore workshop. Iconography
Mughal.
  Quadruped metal stand on eight-lobed ring
base resting on four spherical legs, with two cross
bars holding a central support for a semicircular
arc beneath the globe. This semicircular arc
bears the names of 42 localities with their cor-
responding longitudes and latitudes; there are
23 on one side and 19 on the other, each side
beginning with the words asmd" bildd, tul, and
"urd (names of cities, longitude, and latitude).
The cities are: Mawsil (Mosul), al-Zubayr, Mar-
and, Nakhjawan (Nakhichevan in Azerbaijan),
Maragha, Tabriz, KQfah, Baghdad. Basra, "Aba-
dan, FirQzabad (the earlier Gor in Persia, not the
one in India), ShTraz, Yezd, Shahrazur, Naha-
wand (Nehawend) Abhar, Hamadan, Narakh,
QazwTn (Kazvin), Isfahan, Qum, Rayy, Astarabad
(Astrabad); Medina (Medinah RasQl), Mecca
(Meccah Mu"azzamah), Hajar in Bahrayn, Jeru-
salem (Bayt al-muqaddas), Mahdia, Qayruwan
(Kairowan), Tarablus Maghrib (Tripoli in North
Africa), Andalus (Southern Spain), Alexandria,
Misr (Cairo-Fustat), Aden, Yemen, ZabTd, Asqa-
lan (Ascalon), Tiberias, Tarablus Sham (Tripoli
in the Lebanon), Damascus, Baalbek, Halab
(Aleppo). The horizon ring is graduated by single
degrees with every fifth labeled and bears the
names of the four cardinal points and the four
intermediate directions near outer edge of ring;
numbering is in non-consecutive segments of 90°
beginning at the north-south points, where there
are also two notches for the meridian ring, and

converging at the east-west points. The outside
edge of horizon ring is divided into about 16
groupings of three segments each, one large and
two smaller, separated by engraved double lines;
in one grouping there is engraved asmd" al-bul-
ddn (the names of cities) in the large section, tUl
(longitude), and "urd (latitude) Tn the two smaller
sections; the other spaces are left blank, indicat-
ing that the ring and stand were never com-
pleted. On the underside of the horizon ring are
also the words asmd" bildd, tul, and "urd (names
of cities, longitude, and latitude).
  The meridian ring is graduated on one quad-
rant by single degrees with every fifth labeled
beginning at a point marked North and proceed-
ing clockwise. From the point labeled North,
there are a series of holes anti-clockwise along
the ring labeled 10°, 18°, 24°, 27°, 32°, 40°,
48°, 54°, 66° 30', 72°, and 90°, the latter hole
having a special dark inset and being labeled
"up" {fawq). Continuing anticlockwise from "up"
the next quadrant is ungraduated and contains
no holes. Continuing with the remaining quad-
rant, the zero point is labeled South, and there
are holes labeled 10°, 18°, 24°, 27°, 32°, 40°,
48°, 54°, 66° 30', and 72°. Thus if the ring is
positioned on the horizon ring with the North
point on the north point of the horizon ring (and
the South point on the south point) and the point
marked "up' at the top to mark the zenith, the
axis of the globe can then be set at 10 specified
northern geographical altitudes; there is no ac-
commodation for southern latitudes. A pin passes
through the meridian ring and the equatorial
(celestial) poles.
Citations
Engelmann, 12.
Guide to the Adler Planetarium, 25 (photograph).
41.
  Location:    Chicago, Illinois. Adler Planetarium and Astro-
nomical Museum; Inventory No. A 140.
   Metal; no seam observable; five plugs visible
and one figure-eight hole where there was once
a plug. Diameter 165 mm; weight 1942 grams;
wall thickness 2 mm.  Full set of constellation
   
240

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



figures with about 1018 stars indicated by inlaid
silver points. Ecliptic latitude circles. Ecliptic and
equator graduated by single degrees with every
fifth labeled; ecliptic repeats every 30°; equator
numbered continuously from vernal equinox.
Zodiacal names engraved along ecliptic; constel-
lations and major stars labeled; naskhi, engraving
filled in white. Star positions relative to the equi-
noxes conform to those of the seventeenth-cen-
tury Lahore workshop. While star placement,
engraving of circles and graduations are precise,
the constellation figures are rather crudely
drawn. Surface of globe badly scratched and
beaten. Probably Indo-Persian product of sev-
enteenth-century Lahore.
  Metal quadruped stand on ring, eight-lobed
base with cross bar, but with no undersupport
for globe. Stand and horizon ring clearly made
to go with a different globe, for the ring is too
small for this globe and does not allow globe to
rest with ring at its mid-point. Horizon ring
graduated by single degrees with every fifth la-
beled in non-consecutive segments of 90° con-
verging at the two notches for the meridian ring
which are labeled north and south; bears the
names of the four cardinal points near outside
edge of horizon ring. A 90° graduated arc (same
pattern as ring) is attached and hinged to the top
of the horizon ring to serve as a marker for
zenith and prime vertical. Ungraduated (recent)
meridian ring.
Citation
May be the globe referred to by Yonge, 83, as being at the
Adler Planetarium; otherwise unpublished.
42.
   Location: London (Greenwich). National Maritime Mu-
seum. Department of Navigation; Inventory No. G
175.36.378.
  Metal; no seam observable; two plugs visible;
rattles slightly; and two white pebbles inside,
possibly remnants of casting process; weight is
biased. Diameter 108 mm; weight 104.8 grams.
Full set of constellation figures with about 1018
stars indicated by inlaid silver points. Ecliptic
latitude circles. Ecliptic and equator graduated

by single degrees with every sixth labeled; ecliptic
repeats every 30°; equator numbered continu-
ously from vernal equinox. Zodiacal names en-
graved along ecliptic; constellations and many
stars labeled; naskhi. The 12 zodiacal figures and
the solsitial colure are engraved with dotted lines.
Star positions relative to the equinoxes conform
to those of the seventeenth-century Lahore
workshop. Iconography Mughal and similar to
several of the globes by Diya^ al-DTn Muhammad,
Metal quadruped stand on ring base with sem-
icircular support for globe attached to underside
of horizon ring. Horizon ring graduated by sin-
gle degrees with every sixth labeled in non-con-
secutive segments of 90° beginning at the two
notches for the meridian ring; bears the names
of the four cardinal points near outside edge of
ring. At the center of the semicircular undersup-
port in the notch to accommodate the meridian
ring is a hole possibly threaded for a screw to
hold firm the meridian ring. Meridian ring miss-
ing. Horizon ring and stand appear to be contem-
porary with globe.
43.
Location:    Present location unknown.
  Metal; no seam observable; one round plug
visible. Diameter 180 mm; overall height with
stand and meridian ring 285 mm. Full set of
constellation figures and about 1018 stars indi-
cated by inlaid silver points having a dot in the
middle of each; some variation according to mag-
nitude. Ecliptic latitude circles; equinoctial col-
ure; equatorial tropic and polar circles; ecliptic
polar circles. Ecliptic and equator graduated by
single degrees with every sixth labeled; ecliptic
repeats every 30°; equator labeled in four seg-
ments of 90° each. Zodiacal names engraved
along ecliptic; tropics, ecliptic polar circles, con-
stellations, and major stars labeled; naskhi. Star
positions relative to the equinoxes conform with
those of the seventeenth-century Lahore work-
shop. Iconography Mughal, Carefully executed
engraving.
  Modern stand consisting of four ebony spindle
legs on round feet resting on two ebony cross-
  
NUMBER 46

241



bars with a brass plaque in the center. The legs
support an ungraduated metal horizon ring in-
laid with a ring of ebony. The metal semicircular
support for the globe which is attached to the
underside of the horizon ring terminates in an
ebony finial. Ungraduated metal meridian ring.
The design of the modern stand is of Mughal
inspiration.
  Entry based on photograph and description in
sales catalog.
  Sold in Paris in 1980, Formerly in collection
of Leonard Linton of Point Lookout, New York,
Citations
Collection of Leonard Linton (sale catalog). Sale: Nouveau
Drouot, 9 rue Drouot, 9-10 October 1980. Commissaires-
Priseurs: E. Libert and A. Castor. Expert pres les Douanes:
A. Brieux. Item no. 224, p. 176; photograph p. 177.
44.
  Location:    Chicago. Adler Planetarium and Astronomical
Museum; Inventory No. A 115.
  Metal; no seam observable; three plugs visible
including two large round plugs in line on op-
posite sides of the sphere; rattles slightly. Diam-
eter 115 mm; weight 921.4 grams. Full set of
constellation figures and about 1018 stars indi-
cated by very small inlaid silver points. Ecliptic
latitude circles; it appears that a circle with the
south equatorial pole as center was begun but
not finished; the circle would have been between
the south tropic circle and the south polar circle
had it been completed. For this reason work was
probably stopped on globe and no holes were
drilled at either the equatorial or ecliptic poles,
and the signature of maker was not put on it (cf.
No, 38), Ecliptic and equator graduated by single
degrees with every fifth labeled; ecliptic repeats
every 30°; equator is numbered continuously
from the vernal equinox. Zodiacal names en-
graved along ecliptic; constellations and major
stars labeled; naskhi. The 12 zodiacal constella-
tion figures are engraved with dotted lines. No
hole drilled at either the equatorial or ecliptic
poles; however, two (Precent) large holes drilled
opposite each other about 25° east of the equi-
noxes; purpose unknown. Star positions relative

to the equinoxes conform to those of the seven-
teenth-century Lahore workshop. Iconography
Mughal, Fairly carefully executed engraving.
Stand and rings missing.
  On loan for exhibit to the Hayden Planetar-
ium, New York, from 1964 through March
1974,
45.
   Location: London (Greenwich). National Maritime Mu-
seum. Department of Navigation; Inventory No.
G.7.36.383.
  Metal; no seam observable and no seam felt
inside; one large (Precent) plug in southern hem-
isphere; visual inspection of interior not possible.
Diameter 115 mm; weight with meridian ring
550 grams. Full set of constellation figures with
at least 1018 stars indicated in inlaid silver points.
Ecliptic latitude circles. Ecliptic and equator
graduated by 5° intervals (not single degrees)
with every unit equivalent to 5° labeled; ecliptic
repeats every 30° while equator is numbered
continuously from the vernal equinox. Zodiacal
names engraved along ecliptic; constellations and
major stars labeled; naskhi. Zodiacal constellation
figures engraved with dotted lines. Has two holes
drilled opposite each other about 25° east of the
equinoxes whose purpose is unknown (similar in
this respect to No. 44). The ecliptic and equator
are poorly drawn and somewhat wobbly. Star
positions are very poor, with extra stars fre-
quently being added in a somewhat decorative
manner, filling in many areas both inside and
outside the outlines which should otherwise be
without stars. Position of the stars with regard to
equinoxes is difficult to determine since stars so
poorly placed; both Spica and Regulus (both
labeled) are misplaced and far from the ecliptic.
A technically non-functional and inaccurate
globe. Iconographical design basically similar to
preceding globe (No. 44) but crudely executed.
Could possibly have been made by same maker
as produced No. 32 and No. 46. Possibly an
eighteenth to nineteenth-century Indo-Persian
product.
  A decorative quadruped metal stand on a ring
base supports an horizon ring graduated by two
  
242

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



degrees with every sixth labeled in non-consecu-
tive segments of 90° converging at the notches
for the meridian ring. The legs of the stand are
engraved with human figures and other decora-
tions and have three knobs in the middle of each.
A modern semicircular band attached to the
horizon ring supports the globe and meridian
ring. The meridian ring is graduated by two
degrees with every sixth labeled in non-consecu-
tive segments of 90° beginning at the two points
at which the ring is attached to the globe. The
ring base bears an undeciphered inscription, pos-
sibly nonsensical, engraved in bas relief on a
black background, enclosed in a band. Horizon
ring, stand, and meridian ring may be contem-
porary with globe, although their graduations do
not match those of the globe.
46.
  Location: Bloomfield Hills, Michigan. Cranbrook Insti-
tute of Science; Inventory No. T-88.
  Metal; according to the Curator there is a seam
in the plane halfway between that of the equator
and that of the ecliptic, but this is not visible in
the photographs. Diameter 150mm. Constella-
tion outlines with stars indicated by engraved
circles. Ecliptic latitude circles. Ecliptic and equa-
tor graduated by 5° intervals (not single degrees)
each numbered, equator labeled continuously
from vernal equinox; ecliptic repeats every 30°.
Zodiacal names not engraved along ecliptic. Con-
stellations and some stars labeled; naskhi. Con-
stellation outlines crudely executed; stars very
inaccurately placed and seem arbitrarily posi-
tioned so as to fill spaces symmetrically in and
outside the constellation outlines. The iconog-
raphy has some interesting features such as a bird
in place of Delphinus and Equus. Position of the
stars with regard to the equinoxes is impossible
to determine since the stars are so poorly placed.
A technically non-functional and inaccurate
globe produced for artistic rather than scientific
purposes.
  Metal quadruped stand; the spindle legs each
have a center knob and are covered with en-
graved floral patterns. The horizon ring is grad-

uated by two degrees with every sixth labeled in
non-consecutive segments of 90° converging at
the notches for the meridian ring. The meridian
ring is graduated by two degrees with every sixth
labeled in non-consecutive segments of 90° be-
ginning at the points of attachment to the globe.
Graduations uneven and poorly executed. There
appears to be no undersupport for the globe,
and the meridian ring and globe do not fit
properly into the horizon ring.
See Figure 23 for an illustration.
  There is considerable similarity between this
globe and stand and the globe and stand in No.
45, as well as some similarity with No. 32. Possi-
bly an eighteenth or nineteenth century Indo-
Persian product.
  Bought in 1953 in Cairo and given to the
museum.
  Entry based on information and photographs
supplied by the Cranbrook Institute of Science
and a photograph in the files of the Museum of
the History of Science, Oxford.
Citation
Yonge, 88.
47.
Location:    Present location unknown
  Metal; no seam observable; one large round
plug visible. Diameter 147 mm. Full set of con-
stellation figures with somewhat rubbery human
figures, with about 1018 stars indicated by inlaid
silver points. Ecliptic latitude circles. Ecliptic and
equator indicated by simple lines with no gradu-
ations. Constellations are labeled, but no word
for constellation is used; naskhi. Zodiacal names
are not written along ecliptic. Positions of stars
with relation to the equinoxes conform roughly
to the late seventeenth century. Although there
are no graduations on the globe, the stars and
constellation outlines are basically well placed
and the engraving of the circles is uniform. Icon-
ography is Mughal.
Rings missing.
  Exhibited in Paris at the Cinquieme Congres de
I'Union Astronomique in 1935, and in 1936 at the
  
NUMBER 46

243



exposition of Instruments et Outils d'Autrefois at
the Musee des Arts Decoratifs, where it was listed
as no. 13 in the catalog. Formerly in the collec-
tion of C. Chadenat (no. 53) until 1956, when it
was sold at a public sale in Paris.
  Description based on the one published pho-
tograph and two photographs in the files of F.R.
Maddison.
                       Citation
Chadenat, sale item no. 43; photo, plate IX.
48.
  Location: London (Greenwich). National Maritime Mu-
seum. Department of Navigation; Inventory No. G.4.36.
360.
  Metal; seamless; can see inside that there is no
seam and that there are protruding stumps from
the casting process. There is a hole 26 mm in
diameter where a plug once was. Diameter 127
mm; weight with steel axis rod 735,3 grams; wall
thickness 0.43 mm. Full set of constellations
somewhat crudely drawn and lacking in detail,
with about 1018 stars indicated by inlaid silver
points varying in size approximately according
to magnitude. Ecliptic latitude circles indicated
by very faint circles, the solstitial colure being a
dotted line. Ecliptic indicated by a simple un-
graduated dotted line and the equator by an
extremely faint simple solid line. Zodiacal names
not engraved along ecliptic; constellations and
some stars labeled; naskhi tending toward ta"liq.
Very similar in design and execution to globe
No, 47.
  Ungraduated horizon ring on quadruped
stand with central square knob on legs on ring
base, with inverted semicircular support for me-
ridian ring. Meridian ring ungraduated.
49.
  Location: Glasgow. Private collection of Arthur Frank,
housed at Charles Frank Ltd., 145 Queen Street.
  Metal; not known if in hemispheres. Globe is
said to be "engraved with the ecliptic and equa-
tor, with several of the constellations,"
Exhibited at the Glasgow Art Gallery and Mu-

seum, April-June 1973,
  No further information available. Entry based
on catalog description.
                       Citation
Nuttall, 44; where attributed to ca. AD 1600.
50.
Location:    Present location unknown.
  Metal. Diameter 97 mm. Constellation figures
are engraved. Ecliptic latitude circles; polar cir-
cles. Quadruped stand on circular base; legs en-
graved in scroll pattern, connected in center by
ring of twisted fluting. Meridian and horizon
rings graduated by 3°. Height of support 160
mm.
  No further information; entry based on sale
catalog.
  Formerly in the collection of Ch. Chadenat
until 1956, when it was sold at a public sale in
Paris.
Citation
Chadenat, sale item no. 51; date estimated as ca. AD 1790-
1795.
51.
Location:    Present location unknown.
  Metal. Diameter 140 mm. Constellation fig-
ures engraved; some star names. Stand composed
of horizon ring supported by four small baluster
columns engraved with frondescent pattern,
standing on a circular base.
  In the collection of C. Chadenat unil 1956,
when it was sold at a public sale in Paris. Entry
based on sale catalog; no further information
available.
                        Citation
Chadenat, sale item no. 53; attributed to AD 1795.
52.
Location:    Present location unknown.
  Metal; not known if in hemispheres. "Surface
engraved with the ecliptic and equator and with
  
244

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



several of the constellations." Metal stand of
three curved legs and horizon ring graduated by
single degrees; height 270 mm.
  This globe was sold by Sotheby 8c Co., London,
in 1968; entry from sale catalog; no further
information available.
Citation
Sale catalog of Sotheby Sc Co., London, dated Monday, 22
July 1968.
53.
Location:    Present location unknown.
  Metal; not known if in hemispheres or not; no
seam or plug observable in photograph. Diame-
ter estimated to be between 450 mm and 600
mm. Full set of constellation figures with about
1018 stars indicated by small inlaid silver points.
Ecliptic latitude circles; six meridian circles inter-
secting equator at what would be (if equator had
been numbered continuously) 24°, 54°, 84°,
114°, 144° . . . 354°—that is, a 6° shift, with
no meridian circle representing the equinoctial
colure. Equatorial tropic circles; two lesser circles
each 12° from equator and two lesser circles
each 20° from equator. Silver points at ecliptic
poles and at crossing of ecliptic and equator.
Equator graduated by single degrees with each
6° segment assigned one number (1, 2, 3, 4, . . .
60) numbering in both abjad and standard nu-
merals continuously from vernal equinox. Each
single degree segment contains three dots ar-
ranged in a triangle. Ecliptic graduated by V2°
with every 5° segment numbered with an abjad
as well as standard numeral, repeating every 30 °.
Zodiacal names deeply incised and blackened
along ecliptic; the 48 constellations are labeled
and numbered; naskhi tending toward ta"liq. No
stars appear to be labeled, although they are
numbered within each constellation.
  Stars appear to have been placed after constel-
lation outlines engraved; autumnal equinox not
well drawn, with uneven number of degrees on
either side. Stars rather inaccurately positioned,
although corresponding for the most part to that
of the middle nineteenth century. Although the
star positions and  general  design  correspond

roughly to the globe made by Lalah Balhumal
LahQrT (No. 33), the iconography of the constel-
lations and the engraving and possibly metallurg-
ical techniques differ substantially from those of
BalhQmal; furthermore the globe does not dis-
play the precision and accuracy of the Lalah
BalhQmal globe.
  A three-legged metal stand supports the hori-
zon ring with a semicircular arc (attached by two
winged bolts on each side) holding the nadir of
the meridian ring. The slender S-curved legs on
rounded feet are attached to the horizon ring by
means of a winged bolt. Over the top of the
globe, at right angles to the meridian ring, is a
semicircular arc attached to the horizon ring; the
two sides of this zenith ring are decorated with
an engraved plant motif, with the outer edge
having a figure of a European (P) soldier carrying
a bayonet and also a floral decoration. On the
triangular projection at the top of the semicir-
cular arc marking the zenith there appears to be
another engraved human figure in Indian dress.
The outside edge of the horizon ring is engraved
with an abstract pattern of arcs or peacock feath-
ers. The horizon ring has on its top surface a
floral design near the edge with an inner circle
of numerals and writing not decipherable from
the available photographs. The meridian ring
has the two holes that should contain the axis of
the globe labeled as north and south poles; three
parallel circles are engraved around the outside
edge of the meridian ring. One face of the ring
is graduated by single degrees with every sixth
labeled in abjad as well as standard numerals,
beginning at the poles in non-consecutive seg-
ments of 90° converging at the points opposite
the celestial equator (if mounted properly). The
axis is missing so that the globe is not actually
attached to the meridian ring. Stand and rings
contemporary with globe.
  Possibly middle nineteenth century or later
Indo-Persian product.
  Formerly in the collection of Hew Kennedy of
Shropshire. Said to have been bought in Kapur-
thalah along with a large astrolabe by Lalah
Balhumal LahQrT,
  Entry based on photographs in the files of F,R.
Maddison and A, Brieux,
  
NUMBER 46

245



54.
  Location: New York City. Columbia University, Butler
Library, David Eugene Smith Collection; Inventory No. 27-
244.
  A Sanskrit globe. Metal; no seam observable;
two large plugs and several small ones visible.
Diameter 205 mm. Full set of constellation fig-
ures with about 1018 stars indicated by inlaid
silver points, now blackened. Engraving has been
whitened. Ecliptic latitude circles; six meridian
circles intersect the celestial equator at points (if
the equator had been numbered continuously)
24°, 54°, 84°, 114°, 144°, 174°, 204°, 234°,
264°, 294°, 324°, and 354° (that is, there is no
meridian circle representing the equinoctial col-
ure, for the set of six circles has been shifted 6°
westward); equatorial tropic circles; two lesser
circles each 20° from equator; two lesser circles
each 12° from equator; equatorial polar circles.
Ecliptic and equator graduated by single degrees
with every sixth indicated by longer line and
labeled; ecliptic repeats every 30°; equator has
the 6° segments of each 30° interval between
meridians labeled 1, 2, 3, 4, 5. Sanskrit zodiacal
names engraved along ecliptic; constellations la-
beled and numbered; stars numbered. Writing
and numerals in Sanskrit, DevanagarT characters.
Stars are carefully positioned, corresponding to
middle of nineteenth century, and graduations
and circles precisely executed.
  The three-legged metal stand supports an ho-
rizon ring with attached semicircular arc which
supports the nadir of the meridian ring. The
legs, in S-curve leaf design, resting on small dec-
orated feet, are attached to the horizon ring by
means of winged bolts. The outside edge of the
horizon ring is engraved in an abstract pattern
based on leaf and flower motifs. The horizon
ring is graduated by single degrees with every
sixth indicated by a longer line and labeled in
non-consecutive segments of 90° beginning at
the two notches which would hold the zenith ring
(the semicircular arc corresponding to the prime
vertical; now missing) and converging at the
notches holding the meridian ring. The meridian
ring is attached to the equatorial poles of the
globe by screws and is constructed so that one

half (from pole to pole) is nearly flush while the
other is cut away from the surface of the globe
so as to permit a small slide of 6° width to slide
along it. The slide is graduated by single degrees.
The outside edge of the meridian ring is divided
into 6° intervals, each labeled 6, 12 ... in non-
consecutive segments of 90° beginning at the
poles and ending at the points opposite the celes-
tial equator. The edges near the globe are grad-
uated by single degrees. Stand and rings probably
contemporary with globe (zenith ring missing).
See Figure 25 for an illustration.
  Indo-Persian product of the middle of the
nineteenth century. The design of both the globe
and stand and rings, including the iconography
and star positions, are strikingly similar to that
of globe No. 33 by Lalah BalhQmal LahQrT, and
it is likely that the globe was produced by the
same maker or by someone trained in his work-
shop.
Citations
Yonge, 83; where it is dated AD 1640.
Stevenson, 1: fig. a (opposite p. 17), reproduces a photo-
graph of it incorrectly labeled "Anonymous Arabic Globe,
1635."
55.
  Location: New Delhi, India. Red Fort Archaeological
Museum.
  Metal; no seam or plugs observable in available
photograph. Diameter 150 mm. Full (P) set of
constellation figures, but no stars. Ecliptic lati-
tude circles. Ecliptic and equator graduated by
single degrees with every fifth indicated by
longer line and labeled; ecliptic repeats every
30° while equator is numbered continuously
from vernal equinox. Zodiacal names engraved
along ecliptic; naskhi. The graduations and cir-
cles appear to be precisely and uniformly exe-
cuted. With no stars it is difficult to estimate the
epoch, but the general positions of the constel-
lation outlines corresponds to the late seven-
teenth century. This may be an unfinished globe.
  Quadruped stand 121 mm in height, with six-
sided tapered legs on rounded feet resting on an
eight-lobed  ring  base  with  center  cross-bars.
  
246

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Semicircular support for globe is attached to
underside of horizon ring. Meridian ring appears
to be ungraduated; not known if horizon ring is
graduated.
  Probably Indo-Persian in origin. Entry based
on short published description and a photograph
in the archives of G.R. Kaye preserved at the
Museum of the History of Science, Oxford.
Citation
Stone, 160.

ticularly since even the constellation forms are
irregularly positioned. It is not unreasonable that
this is a late eighteenth- or early nineteenth-
century, possibly Persian, product. A technically
non-functional and inaccurate globe represent-
ing a separate tradition in celestial globe design.
Stand and rings missing.
See Figure 29 for an illustration.
  Acquired in 1965 from Sotheby &: Co., Lon-
don.
Citation
Weil, sale item no. 183; illustrated on the cover. This catalog
is undated, but was issued between 1955 and 1965.

56.

  Location: London (Greenwich). National Maritime Mu-
seum, Department of Navigation; Inventory No. G180. A65-
37.
  Lacquered papier mache on hollow wooden
core. Diameter 152 mm. Black ground with
white circles except for red equator; figures
painted in white, browns and red; red labeling.
There are approximately 48 constellation figures
painted on this globe, but they are not actually
constellation outlines, for no stars are depicted
in them. In fact, there are no stars on this globe.
Ecliptic latitude circles; equatorial tropics; equa-
torial and ecliptic polar circles; equinoctial col-
ure. Ecliptic indicated by a simple, ungraduated
white circle, and equator by a simple red circle.
Zodiacal names are not written along ecliptic;
only the constellation figures are labeled (no
word for constellation); naskhi. Large holes at
the ecliptic poles and only very small holes at the
celestial equatorial poles.
  The iconography is radically different from
that usually found on celestial globes. For ex-
ample, Gemini and Taurus are facing in the
opposite directions from usual, Perseus has the
sword and head in reversed hands and has his
head toward the East. There are actually two
Cassiopeias {dhdt al-kursi or "lady with the
chair"); one roughly above Perseus where it
ought to be, and the other along the equator
north of Aquarius. Since there are no stars on
the globe it is impossible to approximate the
intended epoch or the time of construction, par-

57.
Location:    Paris. Private collection.
  Metal; seam along equator. Diameter 200 mm.
Stars indicated by inlaid points, some of which
protrude from the surface of the globe. Full (?)
set of constellation figures. Ecliptic latitude cir-
cles; equatorial tropic circles; circles about 12°
and 20° from equator in northern hemisphere
only. Equator and ecliptic graduated by two de-
grees with every tenth indicated by a longer line
and labeled; ecliptic repeats every 30°, equator
numbered continuously from vernal equinox.
Zodiacal names engraved along ecliptic; constel-
lations, equator, and mayl kulli labeled; naskhi.
The constellation figures do not seem, from the
one published photograph, to cover the required
areas and represent a different iconographical
tradition from the usual constellation diagrams.
There are some stars, but only a few, within the
outlines of the constellation figures. There ap-
pear to be many fewer stars than are common
on Class A globes. The graduations are even and
precise and the circles well executed, but the star
positions are so irregularly spaced that the epoch
to which they were intended to conform cannot
be determined, at least from the one available
photograph.
  Metal pedestal stand on a circular decorative
base with a semicircular arc attached to under-
side of horizon ring to support nadir of globe.
Both rings graduated by single degrees with ev-
  
NUMBER 46

247



ery fifth indicated by a longer line.
  Entry based on the one published photograph;
no further information available.
                        Citation
Guye 8c Michel, 210, plate 199.
58.
   Location: La Chaux-de-Fonds, Switzerland. Musee d'hor-
ologerie.
  Metal; no seam observed; no plugs observed.
Diameter not recorded. Globe has no stars. Does
not have the usual constellation figures, but
merely ornamental figures which appear to fill
up the areas of the globe other than the area
where the zodiacal names are engraved near the
ecliptic in very large, double-outlined writing.
Ecliptic latitude circles; ecliptic tropic circles;
equatorial polar circles; equinoctial colure. Eclip-
tic and equator graduated by single degrees with
every fifth labeled. Circles are rather poorly
drawn on the globe and graduations are uneven.
A non-functional and inaccurate globe.
  Metal quadruped stand on a ring base with
ornamental engraving on the legs. Horizon ring
appears to be graduated by two degrees with
every sixth labeled; meridian ring is ungraduated
with ornamental engraving.
  Entry based on photograph taken in May of
1972 by Uta C. Merzbach of the Smithsonian
Institution.
CLASS B—CELESTIAL GLOBES WITHOUT
CONSTELLATIONS
Signed or Dated
59.
Location:    Tehran. Muzeh-i Iran-i Bastan.
Date:    535 H [AD 1140-1141].
Maker:    Badr ibn "Abdallah Mawla BadT" al-Zaman.
  Metal; in hemispheres with seam along ecliptic;
probably cast hemispheres. Diameter 140 mm;
height on stand 160 mm. No constellation fig-
ures; about 50 stars or star groups indicated by

inlaid silver points or groups of silver points.
Ecliptic latitude circles. Ecliptic and equator
graduated by single degrees with every fifth la-
beled; ecliptic repeats every 30°; equator labeled
in three segments of 100° each and one segment
of 60° starting at vernal equinox. Zodiacal names
engraved along ecliptic; stars labeled; lunar man-
sions written along ecliptic; KQfic. KQfic inscrip-
tion north of the Houses of Gemini and Cancer
reads: "The work of {san"at) Badr ibn "Abdallah
Mawla BadT" al-Zaman, year 535." A much later
inscription in naskhi script has been added in the
southern hemisphere on the other side of the
equator from the previous one; it is dated 1304
H [AD 1886-1887] and records that the globe
belonged to {min mumtalakdt) "Imad al-Dawlah
ibn "Imad al-Dawlah ibn Muhammad "AIT
ibn ... h "AIT Shah.
  A meridian ring (not contemporary with
globe), plain and ungraduated, is attached to the
ecliptic poles. The meridian ring sits in a metal
stand consisting of a horizon ring supported by
four legs resting on a circular ring base. The
horizon ring is graduated by single degrees with
every fifth labeled in KQfic abjad numerals in
non-consecutive segments of 90° beginning at
the notches holding the meridian ring. A non-
functional hole has been drilled into the globe
between the ecliptic and equator at the House of
Cancer, possibly to see if anything was inside.
  Entry based on information and photographs
in the files of F.R. Maddison.
  Displayed at the exhibition "Science and Tech-
nology in Islam" held at the Science Museum,
London, as part of the Festival of the World of
Islam, April-August, 1976.
Citation
Maddison 8c Turner, 76.
60.
Location:    Oxford. Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/181.
Date:    718 [H] [AD 1318-1319].
Maker:    "Abd al-Rahman ibn Burhan al-MawsilT.
  Metal; no seam observable; 2 plugs visible; no
seam detected by probing the rough interior
  
248

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



surface or with limited visual inspection of inte-
rior. Diameter 153 mm; weight 1488.3 grams.
No constellation figures; about 100 stars indi-
cated by inlaid silver points. Ecliptic latitude cir-
cles; equatorial and ecliptic polar circles. Ecliptic
and equator graduated by single degrees with
every fifth labeled; both equator and ecliptic
repeat every 30°. Zodiacal names engraved
along ecliptic in narrow double-outlined writing;
names of the 28 lunar mansions also engraved
near ecliptic; equatorial poles, equatorial polar
circles, and most stars and star groups labeled;
naskhi. The inscription near the north ecliptic
polar circle reads: "The work of {"amal) "Abd al-
Rahman ibn Burhan al-MawsilT in the months of
the year seven hundred and eighteen."
  Near the south poles are two inscriptions lack-
ing many diacritical points and written rather
poorly; they may possibly be by a different hand.
The sense of the first seems to be: "Undertaken
at the instigation of the poor, the weak {al-faqir
al-haqir) Ghiyath known as al-Mansur in the
months of the year .... (lacuna) and seven
hundred." (The expression for "at the instigation
of" is the mixed Arabic and Persian matwd bar
nazar; for a similar use of nazar in Arabic see
Dozy, 2:694.) The lacuna occurs where an insert
(P of more recent origin) has been placed in the
hole at the ecliptic pole and interferes with the
inscription. The second inscription reads: "By
order of {bi-rasm) the cabinet of treasuries {khi-
zdnah) of the Sultan al-Matt al-"Adil Ulugh Beg.'
{Al-Matt [the proud] could, with some modifica-
tion, be read as al-malik [the ruler].) The dates
in the first two inscriptions are written out in
words rather than in numerals; this in known to
occur on only one other globe, where the inscrip-
tion is in Persian (No, 33).
  The star groups representing the lunar man-
sions have been placed near the ecliptic in geo-
metrical patterns which show a reliance by the
maker upon the cosmology of al-QazwTnT or the
occult writings of al-Bunl, both of the thirteenth
century, or similar writings in which geometrical
patterns are associated with the lunar mansions
(see Savage-Smith 8c Smith, 40-41, Table 2),
Holes are drilled at both the ecliptic and equa-

torial poles. The lines representing the ecliptic
and equator are somewhat wobbly and the grad-
uations rather uneven. The star positions near
the ecliptic are so poorly placed that it is difficult
to tell to what epoch they conform. If the date
on the globe is considered correct, one must
assume the last inscription to have been added
later, since Ulugh Beg lived in the fifteenth cen-
tury.
  A metal stand consists of three curving legs
meeting at the bottom of a pedestal stand sup-
porting two crossed semicircular arcs to which
the horizon ring is attached. The horizon ring is
engraved only with a line at every 30° interval;
ungraduated meridian ring. Around the top of
the pedestal support is a curious inscription
which is transcribed in Chapter 8.
  This is a technically non-functional and inac-
curate globe which might possibly be of a more
recent date of execution than that indicated in
the inscriptions.
  See Figures 7 and 8 for illustrations of the
globe.
  Exhibited in Paris in 1935 at the exposition
Instruments et Outils d'Autrefois; no. 11 of catalog.
In the Chadenat collection until May 1957, when
it was sold at public sale in Paris.
Citations
Chadenat, item no. 42, plate XII (photograph).
Maddison [1957], 40, no. 181, photograph, on cover.
Mayer [1956], 31.
Destombes [1957], 320 no. 8.
61.
Location:    Paris. Private collection of Marcel Destombes.
Date:    813 H [AD 1410-1411].
Maker:    Muhammad ibn Ja"far al-AsturlabT.
  Metal; hemispheres joined at ecliptic. Diame-
ter 67.5 mm. No constellation figures; 35 stars
indicated by inlaid silver points. Ecliptic latitude
circles. Ecliptic and equator graduated by single
degrees with every fifth labeled; ecliptic repeats
every 30°; equator numbered continuously from
vernal equinox. Zodiacal names engraved along
ecliptic; equatorial poles and stars labeled; Kufic.
Holes have been drilled at both the ecliptic and
  
NUMBER 46

249



equatorial poles. The graduations are irregular
and the precision not perfect. In the southern
hemisphere beneath the houses of Sagittarius
and Capricorn is the inscription, written without
diacritical points, reading: "The work of {san"at)
Muhammad ibn Ja"far al-AsturlabT in the year
813 H." The date could also be interpreted as
818, 853, or 858 H, but the earliest date seems
most likely in terms of other instruments made
by him and by his father. The maker is the son
of the maker of globes Nos, 7 and 8,
Stand and rings missing.
  Formerly in the collection of M.A. Lee, this
globe was sold in 1962 at Sotheby & Co., Lon-
don. A detailed study of this globe is to be
published by M. Destombes, who interprets the
date (written in Kufic abjad letters without dia-
critical points) as 803 H [AD 1400-1401].
Citation
Sale catalog of Sotheby & Co., London, dated 26 February
1962, p. 8. Plate IV has two photographs.
62.
   Location: London. British Museum, Department of Ori-
ental Antiquities; Inventory No. 26.3.23.
   Date: 834 H, 805 Yazdijird, and 1733 Alexandrian era
[AD 1430-1431].
  Maker: Muhammad ibn Ja"far ibn "Umar al-AsturlabT
known as Hilal.
  Metal; hemispheres joined at ecliptic; dark fin-
ish. Diameter 105 mm. No constellation figures;
about 60 stars indicated by inlaid silver points.
Ecliptic latitude circles. Equator and ecliptic
graduated by single degrees with every fifth la-
beled; ecliptic repeats every 30°; equator num-
bered continuously from vernal equinox. Zodia-
cal names engraved along ecliptic; stars and star
groups labeled; Kufic. Near the north equatorial
pole there is the inscription: "The work of
{san"at) Muhammad ibn Ja'^far ibn "Umar al-
AsturlabT al-mulaqqab (known as) Hilal." The last
name can also be read as Jalal since it is written
without diacritical points. Near the south pole
another inscription reads: "These stars were
placed {rusimat) after increasing their longitude
7°5' according to the Constellations of Abu al-

Husayn ibn "Abd al-Rahman al-SQfi in the year
834 H, 805 Yazdijird, or 1733 of the Alexandrian
Era." The year 1733 could be read as 1738 since
there are no diacritical points. Holes have been
drilled at both equatorial and ecliptic poles.
Graduations are irregular and the precision not
the best.
  A slender metal pedestal stand is on a quad-
ruped base formed by two intersecting semicir-
cular arcs. The upper part of the pedestal sup-
ports two crossed semicircular arcs that hold the
horizon ring, which is graduated by single de-
grees with every fifth labeled in non-consecutive
segments of 90°, converging at the notches for
the meridian ring. The ungraduated meridian
ring is semicircular, covering only the top half of
the globe; holes are placed in it at 5° intervals
for different terrestrial latitude settings. One end
of a chain is attached to one end of the meridian
ring where it is joined to the horizon ring; the
other end of the chain holds a pin which passes
through the desired latitude hole in the meridian
ring and the celestial north pole of the globe.
Stand and rings appear contemporary with
globe.
See Figure 10 for an illustration.
Citations
Mayer [1956], 68, with photograph of signature on plate
Xlla.
Destombes [1957], 321, no. 10.
Destombes [1960], 206-207, no. 3; reads the Yazdijird date
as 800 and prefers the reading of Jalal to Hilal; also states
that the correct Alexandrian date would be 1743.
63.
Location:    Chicago. Adler Planetarium and Astronomical
Museum; Inventory No. A 114.
Date:    1012 [H] [AD 1603-1604].
Maker:    Unsigned, but bears patron's name.
   Metal; hemispheres with seam at ecliptic. Di-
ameter 150 mm; weight 1389.1 grams. Only 12
figures representing the zodiacal houses are en-
graved; they are not actually constellation out-
lines but only decorative medallions. There are
38 stars indicated by inlaid silver points sur-
rounded by a  circle.   Ecliptic  latitude  circles;
   
250

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



equatorial tropic and polar circles; ecliptic tropic
circles; equinoctial colure and one circle 30° to
the west passing through the equatorial poles (a
meridian). There are also circles about 12° inside
each of the tropic circles, parallel to both the
ecliptic and equatorial tropic circles. Equator
represented by a simple line with no graduations;
ecliptic represented by the line of the seam. The
figures representing the Houses are not the usual
constellation figures (e.g.. Libra is a cross-legged
human figure holding the scales like a yoke over
the shoulders) and there are no stars indicated in
them; each figure is enclosed in a circle and is of
a decorative style found on some other forms of
Islamicate metalwork. Each medallion figure is
labeled with the name of the zodiacal house
outside the circle; equinoxes, solstices, ecliptic
latitude circles, meridian, both colures, both sets
of poles, ecliptic, equator, mayl kulli, tropic and
polar circles are labeled; naskhi in floriated dou-
ble-outlined writing on recessed stippled ground.
A unique feature of the globe is a large crescent
moon engraved on it between the north equato-
rial pole and the House of Aries. Within a car-
touche located between the Houses of Aries and
Pisces and the south ecliptic pole is an inscription
reading: "By command of {hasb al-amr) the sultan
of sultans, the honorable and powerful Bandeh
Shah Walayat "Abbas, completed 1012." The
patron was Shah "Abbas I, who ruled AD 1587-
1629 (for the formula used for his title see Ra-
bino di Borgomale, 32-33).
  Globes No. 82 and 83 are very similar in
design.
Stand and rings are missing.
See Figure 20 for an illustration.
Citations
May be the globe of diameter 150 mm referred to by Yonge,
83. Otherwise unpublished.
64.
Location:    Cairo. Museum of Islamic Art. Inventory No.
15266.
Date:    Not given.
Maker:    Muhammad Zaman.
Metal; hemispheres joined at ecliptic. Diame-

ter 74 mm. No constellation figures; about 40
stars, which had probably been indicated by in-
laid silver points that are now missing. Ecliptic
latitude circles; equatorial tropic and polar cir-
cles; equinoctial colure. Equator and ecliptic
graduated by 2° intervals with every tenth de-
gree labeled; ecliptic repeats every 30°; equator
is numbered continuously from vernal equinox.
Zodiacal names engraved along ecliptic; equato-
rial north pole labeled and 36 stars labeled; nas-
khi. Near south poles it is signed: "The work of
{san"at) Muhammad Zaman." Star positions rel-
ative to the equinoxes conform to the middle of
the seventeenth century. Maker also made globes
with constellations, Nos. 16 and 17.
Stand and rings missing.
  Exhibited in Cairo in 1935 at an exposition of
Persian art.
Formerly in Harari collection, no. 402.
Citations
Wiet [1939], 16, no. 35.
Mayer [1956], 79.
65.
   Location: Baltimore, Maryland. Walters Art Gallery; In-
ventory No. 54.712.
   Date: 1056 H, 1004 Yazdijird, and 1986 Alexandrian
era [AD 1646-1647].
Maker:    No signature.
  Metal; hemispheres with seam passing through
the ecliptic poles but not coinciding with any of
the ecliptic latitude circles; very dark finish. Di-
ameter 100 mm. No constellation figures; about
150 stars indicated by inlaid silver points. Ecliptic
latitude circles. Ecliptic and equator graduated
by single degrees with every fifth labeled; ecliptic
repeats every 30°; equator numbered continu-
ously from vernal equinox. Zodiacal names en-
graved along ecliptic; names of 28 lunar man-
sions engraved along ecliptic; ecliptic and equa-
torial poles and most stars labeled; naskhi tending
toward ta"liq. Beneath the Southern Fish there is
the Arabic inscription: "These stars were placed
{rusimat) after increasing their longitude 4 [de-
grees] in accordance with the observatory of
Samarqand in the year  1056 H,   1004 of the
  
NUMBER 46

251



Yazdijird era, and 1986 of the Alexandrian era,"
It is probable that 1956 was intended instead of
1986, which would involve a very slight altera-
tion of the reading. The entire globe is badly
corroded and the writing of the inscription
rather crowded and not very clear. Holes have
been drilled at both the ecliptic and equatorial
poles. The precision of the placement of stars
relative to the equinoxes is not consistent, but
the majority correspond to the middle of the
seventeenth century.
Stand and rings missing.
  The globe was acquired from Kelekian by the
Walters Art Gallery in 1928. It was exhibited in
1952 and 1955-1956 at the Walters Art Gallery
(no. 20 in the catalog, "The World Encom-
passed," 7 October-23 November 1952, where
it was listed as thirteenth century or later) and in
1955-1956 in Newark at the Newark Museum,
Citation
Yonge, 83 says it is undated but attributes it to the seven-
teenth century; a very short entry. Otherwise unpub-
lished.
66.
   Location: Patna (Bihar), India. Private Collection of
Maulvi Yusuf Sahib in the Phulvari district (as of 1935).
Date:    1058 H. [AD 1648-1649].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn Mulla "Tsa ibn Mulla Allahdad AsturlabT HumayunT
LahiirT.
  Metal; no seam observable; one very large
circular plug under signature. Diameter un-
known; weight 567 grams. No constellation fig-
ures; about 20 stars indicated by rather large
inlaid silver points. Ecliptic latitude circles; equa-
torial tropic and polar circles. Equator and eclip-
tic graduated by two degrees with every sixth
labeled. Zodiacal names engraved along ecliptic;
most of the stars labeled; naskhi. Signature near
end of Eridanus reads: "The work of {"amal)
Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn Mulla Tsa ibn Mulla Allahdad AsturlabT Hu-
mayunT LahQrT in the year 1058 H." Although
the stars are large in proportion to the globe,
being about 4° in diameter, they appear to be
positioned like those on other globes of the sev-

enteenth-century Lahore workshop.
  Metal quadruped stand on circular base with
central pedestal support for the semicircular arc
attached to underside of horizon ring. Outside
edge of meridian ring graduated by 6° intervals,
each of which is labeled; meridian ring contains
holes for setting globe for different terrestrial
latitudes.
  Entry based on published statements and one
photograph.
  The family in Patna has owned the globe since
AD 1822-1823 and it was still in their possession
in 1935.
Citations
Nadvi [1935], 628 no. 1.
Kluber, 5 (photograph).
67.
Location:    Paris. Private collection of Marcel Destombes.
Estimated Date:    AD 1650 (not inscribed).
Maker:    Lutf Allah ibn "Abd al-Qadir al-Muhibb al-Astur-
labT.
  Metal; no seam observable; large circular plug
15 mm in diameter around north ecliptic pole;
lead segregation evident. Diameter 67 mm. No
constellation figures; 62 stars indicated by inlaid
silver points. Ecliptic latitude circles; equatorial
tropic and polar circles. Ecliptic and equator
graduated by two degrees with every tenth la-
beled; ecliptic repeats every 30°, while equator
is numbered continuously from vernal equinox.
Zodiacal names engraved along ecliptic; ecliptic
poles, equatorial poles, and stars labeled; rather
crude naskhi. Inscription near north poles reads:
"The work of {san"at) the weakest of servants
Lutf Allah ibn ^Abd al-Qadir al-Muhibb al-Astur-
labi, may God forgive his sins." Star positions
conform to middle seventeenth century.
  Quadruped metal stand with horizon ring. Me-
ridian ring pierced so as to allow for the globe to
be set at different terrestrial latitudes at intervals
of 4°. Height of globe on stand 137 mm.
  In the collection of M.A. Lee before being sold
in 1962 at Sotheby 8c Co., London.
                     Citation
Sale catalog of Sotheby Sc Co., London, dated 26 February
1962, p. 8 no. 15.
                   
252

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



68.
   Location: Cambridge, England. Whipple Museum of the
History of Science; Inventory No. 1255.
Date:    21 Shawwal 1065 [24 August AD 1655].
   Maker: Hamid ibn Muhanmiad MuqTm ibn "Isa ibn
Allahdad AsturlabT LahiirT HumayiinT.
  Metal; no seam observable; one large plug of
diameter 40 mm and several small dark plugs
visible. Diameter 99.3 mm. No constellation fig-
ures; about 67 stars indicated by silver dots inlaid
with indentations in the middle of each; one such
silver dot is inlaid at each ecliptic pole. Ecliptic
latitude circles. Ecliptic and equator graduated
by single degrees with every fifth labeled; ecliptic
repeats every 30° while equator is labeled in two
segments of 180° each beginning at the equi-
noxes. Zodiacal names engraved along ecliptic;
all stars and equatorial poles labeled; naskhi,
rather poorly engraved with resulting mistakes.
Maker's name and date inscribed near the south-
ern equatorial pole: "^n/ [an error for san"at, the
work of] the unworthy servant {qalil al-"ibdd)
Hamid ibn Muhammad MuqTm ibn "Isa ibn
Allahdad AsturlabT LahQrT HumayQnT, written
[tahrir, meaning the act of writing or composing,
possible here in sense of engraving] on the date
{fi al-tdrikh) the twenty-first of the month of
Shawwal of the year 1065." Star positions im-
mediately along ecliptic indicate an epoch of
about the second half of the seventeenth century,
but those stars further from ecliptic are rather
poorly positioned. The star at 5° House of Pisces
(beneath the signature) was incorrectly labeled
dhanab al-qitus (tail of Cetus) and re-engraved
sdq sdkib al-md" (shinbone of Aquarius). The lat-
ter star, b Aquarii, is frequently found on astro-
labes and on Class B globes even though it is not
usually labeled on Class A globes.
See figure 14 for an illustration.
Stand and rings missing.
69.
   Location: C'airo. Museum of Islamic Art; Inventory No.
3800.
Date:    1070 H [AD 1659-1660].
   Maker: Diya^ al-DTn Muhammad ibn Qa^im Muhammad
ibn "Isii ibn Allahdad AsturlabT HumayunT LahiirT.
  
Metal; no seam observable; one square plug
visible. Diameter 60 mm. No constellation fig-
ures; about 60 stars indicated by a dot enclosed
in a small circle. Ecliptic latitude circles. Equator
and ecliptic graduated by two degrees with every
sixth labeled. Zodiacal names engraved along
ecliptic; stars labeled; naskhi Signature around
south equatorial pole begins: "The work of
{"amal) Diya^ al-DTn Muhammad . . . . " Holes at
both ecliptic and equatorial poles.
  Metal horizon ring on metal quadruped stand,
92 mm high, on ring base. Meridian ring has
labeled holes for setting at different latitudes.
Probably contemporary with globe.
  No further information recorded at time of
examination.
70.
Location:    Karachi. National Museum of Pakistan.
Date:    Fourth of Ramadan  1083 [H] [25 December AD
1672].
Maker:    Unknown.
  Metal; no seam observable in photograph; one
possible large circular plug. Diameter unknown.
No constellation figures; about 50 stars indicated
by small inlaid silver points. Ecliptic latitude cir-
cles; equatorial tropic and polar circles and circles
12° and 20° parallel to equator on both sides;
equinoctial colure. Ecliptic (Pand equator) grad-
uated by 2 ° (P) intervals with every sixth degree
labeled; ecliptic repeats every 30°. Zodiacal
names engraved along ecliptic; most stars la-
beled; naskhi tending toward ta"liq. Near the
southern polar circle beneath the Houses of Sag-
ittarius and Capricorn is a three-line inscription,
the first line of which is nearly illegible in the
photograph, only the word Muhammad in the
middle being legible. The other two lines read:
"on the date {fitdrikh) of the fourth of the month
of Ramadan [in the] year 1083."
  Three-legged metal stand with the S-curve legs
being shaped in a leaf design and resting on
rounded decorated feet. A semicircular arc at-
tached to the underside of the horizon ring sup-
ports the globe; over the top of the globe is a
  
NUMBER 46

253



heavy, semicircular zenith ring, held in position
by two wing-bolts at the sides of the horizon ring.
The horizon ring is graduated by 2 ° (P) intervals
with every sixth degree labeled in non-consecu-
tive segments of 90° beginning at the notches
holding the zenith ring and converging at the
notches for the meridian ring; the outside edge
of the top of the horizon ring is engraved in a
decorative pattern. The meridian ring is also
graduated by 6° intervals on the outside face.
One half of the ring may be thinner than the
other half, though it is difficult to tell from the
photograph. The stand and rings appear to be
contemporary with the globe.
  The globe is probably Indo-Persian in origin.
The stand and rings bear a striking similarity to
those of No. 54 and No. 33, although the date
of this globe is considerably earlier than those
globes. On the other hand, the calligraphy of this
maker resembles the poorly inscribed signature
of No. 68 made by Hamid ibn Muhammad Mu-
qTm of the Lahore workshop in 1065 H/AD 1655.
  Entry based on a photograph taken by Simon
Digby, formerly of the Ashmolean Museum; pho-
tograph now on deposit with the Museum of the
History of Science, Oxford.
71.
Location:    New Delhi. Red Fort Archaeological Museum.
Date:    1087 [H] [AD 1676-1677].
  Maker:    Diya^ al-DTn Muhammad ibn Mulla Qa^im Mu-
hammad ibn Hafiz "Isa ibn Shaykh Allah-dad HumayunL
  Metal; no seam observable in available photo-
graphs; the three "inlaid patches" described by
Kaye are probably plugs. Diameter 65 mm. No
constellation figures; 92 stars indicated by a dot
enclosed in a small circle. Ecliptic latitude circles;
equatorial tropic and polar circles. Equator and
ecliptic graduated by 2° intervals with every
sixth degree labeled; ecliptic repeats every 30°
while the equator repeats every 90°. Zodiacal
names engraved along ecliptic; equatorial poles
and 91 stars are labeled; naskhi. Signature
around south equatorial pole beings: "the work
of {"amal) the weakest of the servants {ahqar al-
"ibdd) Diya"" al-DTn . . . . " The name of the maker

resembles that on globe no. 30 rather than his
other products, in that the name, Hafiz is added,
Allah-dad is written with different orthography,
and LahQrT is omitted.
  Heavy metal quadruped stand with straight
six-sided legs on an eight-lobed ring base with
center cross-bars holding a central pedestal sup-
port for the meridian ring. Horizon ring is grad-
uated by 2° intervals with every sixth degree
labeled in non-consecutive segments of 90° con-
verging at the notches for the meridian ring
which are also labeled north-south; names of the
four cardinal points engraved near outer edge of
horizon ring. There are also indentations on the
top of the horizon ring at the zero points, appar-
ently to accommodate a semicircular arc serving
as a zenith ring, now missing. Ungraduated me-
ridian ring is probably a recent replacement.
  Entry based on published account and photo-
graphs in the files of A. Brieux and in the ar-
chives of G.R. Kaye, now at the Museum of the
History of Science, Oxford.
Citations
Kaye [1921], 16-19, with photographs in figure 9 of plate
  II. Kaye makes several mistakes in transliteration.
Stone, 160.
72.
Location:    Cairo. Museum of Islamic Art; Inventory No.
15364.
Date:    1215 [H]/[AD 1800-1801].
Maker:    Muhammad Ashraf TuqadT Zadah.
  Painted wooden globe, or possibly painted
ivory; brown ground with black lines and label-
ing; lacquered. Diameter 78 mm. No constella-
tion figures; about 50 stars marked with a dot of
black paint. Meridian circles cross the equator at
30° intervals; there are no ecliptic latitude cir-
cles. Equatorial and ecliptic polar and tropic cir-
cles. The circles parallel to the equator at 12°
and 20° (as well as 24° which mark the tropics)
north and south are prominently indicated.
There is also a complete set in very fine ink lines
of parallels at 2° intervals north and south of the
equator, and a similar set at 2° intervals parallel
to the ecliptic. The ecliptic and equator them-
  
254

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



selves are indicated by simple lines with small
dots marking the 2° intervals with every 10°
interval indicated by a short dash at right angles
to the great circle. The zodiacal names are writ-
ten along the ecliptic, the stars are labeled; small
naskhi tending toward ta"liq. Near the south ce-
lestial pole is the signature "The design of {ras-
mat) Muhammad Ashraf TQqadT Zadah [in the]
year 1215."
  The rings and stand are clearly contemporary
with this globe, for the stand has the signature
"the work of {"amal) Muhammad Ashraf 1215."
The metal quadruped stand (72 mm high) is on
a ring base bisected by a bar to which is attached
a central support for the meridian ring. The
horizon ring is graduated by 2° intervals with
every tenth degree labeled in non-consecutive
segments of 90° converging at the notches for
the meridian ring; horizon ring has names of the
four cardinal points engraved near outside edge;
notches labeled north-south. The metal meridian
ring is also graduated by 2° intervals with every
tenth degree labeled in non-consecutive seg-
ments of 90° beginning at the points of attach-
ment to the globe and converging near the equa-
tor. The pin running through the globe as an
axis is connected to a small metal plate which is
attached to the meridian ring at the north pole
with a small screw.
  Both globe and rings exhibit remarkable pre-
cision of design and execution. Perhaps the con-
trast between "amal on the stand and rasmah on
the globe indicated that Muhammad Ashraf ac-
tually constructed as well as graduated the metal
rings and stand, but designed and executed the
surface details of the globe onto a sphere that
had been fashioned by another craftsman. Pos-
sibly Muhammad Ashraf also wished to make
clear by the term rasmah that he himself designed
the rather unusual style of the globe and that the
design was not copied from another globe. Com-
pare this intricate design with that of the anony-
motis globe No. 76.
Formerly in Harari Collection, No. 159,
Citation
Mayer [19561, 64.

73.
Location:    Cairo. Museum of Islamic Art; Inventory No.
3774.
Dated:    23 Jumada I, 1221 H [ 9 August AD 1806].
Maker:    Muhammad "AIT al-HusaynT.
  Painted and lacquered papier mache over hol-
low wood core. Diameter 234 mm. Globe painted
with dark rich brown ground, red and black
circles, black labeling, gilt stars, and red, black
and gilt figures. Only 12 figures representing the
zodiacal houses are depicted on the globe; they
are not actual constellation outlines, but only
small decorative figures, with no stars indicated
in them. There are approximately 50 stars indi-
cated by very large gilt dots, most of which are
labeled. Ecliptic latitude circles; equatorial tropic
and polar circles indicated by black circles. Equa-
tor and ecliptic each indicated by two red and
one black parallel circles, graduated by single
degrees with every fifth labeled; ecliptic repeats
every 30°; equator numbered continuously from
vernal equinox. Zodiacal names written along
ecliptic; ecliptic latitude circles, poles, polar cir-
cles, and mayl kulli, and most stars labeled; naskhi.
The mayl kulli is labeled mayl kulli janubi {sham-
dli) 23 darajah wa 30 daqiqah wa 17 thdniyah
(southern [northern] greatest distance of ecliptic
from equator is 23 degrees 30 minutes and 17
seconds). Holes at both ecliptic and celestial
poles. Graduations are uneven and circles wobble
slightly. The surface of the globe is not uniformly
spherical. Calligraphy quite good. Technically
non-functional and inaccurate, though star posi-
tions conform roughly to those of the beginning
of the nineteenth century.
  Between the south celestial pole and the zodia-
cal houses of Sagittarius, Capricorn and Aquarius
is a long Arabic inscription which includes quo-
tations from the Qur^an whose identification I
have noted in the translation:
Glory be to God, there is no god but He, Creator of the
heavens and earth. Who placed in the sky the zodiac, and
we adorned them for the beholders [Qur^an XV, 16] and the
sun and the moon and the stars were made to pursue their
course only by His authority, to Him belong creation and
command. Praise be to God, Lord of the Worlds, that in the

NUMBER 46

255



difference of the night and the day there are signs [of His
Sovereignty] for those of vision [Qur^an 111,190]. Praise be
to God the Best Creator. And the sun runs to a fixed resting-
place that is the ordaining of the All-mighty and the All-
knowing [Qur^iin XXXVI,38] and the moon—we have de-
termined it by mansions until it returns like an aged palm-
tree bough [Qui^an XXXVI,39], that they might know the
number of the years and the reckoning [Qur^an X,5] that in
that are tokens [of His sovereignty] for those of understand-
ing. Finished by the hand of {tummat "ala yad) the hopeful
[of God's mercy] and contrite servant [of God] Muhammad
"AIT al-HusaynT, supervisor of the teaching facilities, and on
his master [? employer] a thousand salutations and praise.
On Saturday, the twenty-third of Jumada I, 1221 H.
  The zodiacal figures are representative of a
different iconographic tradition from that found
on celestial globes on which the figures serve as
outlines of the asterisms. The figures are remin-
scent of the Persian iconographic tradition de-
picted in the medallions of globes No. 63 and
No. 83 and on other forms of Islamic metalwork,
but also bear some resemblences to the zodiacal
figures in the constellations pictured on globe
No. 56, also of lacquered papier mache.
Stand and rings missing.
74.
Location    London. Victoria and Albert Museum, Depart-
ment of Metalwork; Inventory No. M.24-1882.
Date:    1241 H and 1195 Yazdijird era [AD 1825-1826].
Maker:    Muhammad KarTm.
  Metal; no seam observable; several plugs visi-
ble including large square plug around south
equatorial pole; rattles slightly. Diameter 126
mm. No constellation figures; about 80 stars
indicated by inlaid silver points which have tar-
nished and several of which are missing. Ecliptic
latitude circles; equatorial tropic and polar cir-
cles. Ecliptic and equator graduated by single
degrees with every sixth labeled; ecliptic repeats
every 30°; equator is numbered continuously
from vernal equinox. The solstitial colure is grad-
uated by single degrees with every sixth indicated
by a longer line and labeled every 6° in non-
consecutive segments of 90° beginning at the
ecliptic and converging at the ecliptic poles. Zo-
diacal names engraved along ecliptic; names of
the 28 lunar mansions engraved along ecliptic at

intervals of about 12'/2° indicated by short lines;
equatorial poles and most of the stars are labeled;
naskhi tending toward ta"liq. Within the south
equatorial polar circle there is the signature:
"The work of {"amal) Muhammad KarTm [in
the] year 1195 Yazdijird Solar {shamsi) era."
Outside the polar circle there is the date, 1241
H. Star positions relative to the equinoxes con-
form to those of the late seventeenth or early
eighteenth century. Executed with considerable
precision.
Stand and rings missing.
  Was acquired in India and given to the Victo-
ria and Albert Museum in 1882.
Citation
Mayer [1956], 69.
75.
Location:    Istanbul. Collection of Oduglii.
Date:    3 RabT" 1 1299 [23 January AD 1882].
Maker:    Husayn Hiisnii KanqarIT.
  Metal; seam along equator; dented on one side.
Diameter unknown. No constellation figures;
about 100 stars indicated by relatively large en-
graved asterisks or circles with radiating lines.
Ecliptic latitude circles; full set of meridians;
equatorial tropic and polar circles. Equator and
ecliptic graduated by single degrees with every
fifth indicated by longer line. Along the ecliptic
every 10° interval is labeled in standard numer-
als, repeating every 30°; every 10° interval of
the equator is labeled in standard numerals, 10°,
20°, 30° ... 180°, possibly proceeding contin-
uously around the equator, although it may be
labeled in two 180° segments, for it cannot be
determined from the one available photograph.
The names of the zodiacal houses and the nu-
merals are written with the top of the writing
toward the north poles—that is, in the reverse
direction from that on all other known Islamicate
globes. Among the star names are some constel-
lation names; the majority of the stars are la-
beled; naskhi. The inscription in Ottoman Turk-
ish between the north polar circle and the House
of Cancer  reads:   "A  student  at  the   Normal
  
256

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



School, Husayn Husnu Kanqarli made {ikhtira")
this celestial globe {kurah samdwi) [in the] year
1229, the third of RabT"."
Stand and rings missing.
  This is the only known example of an Ottoman
Turkish celestial globe. While the circles are
uniformily executed and the graduations of mod-
erate precision, the star positions are imprecise
and vary greatly in the epoch to which they
conform.
  Entry based on photograph and transcription
of the inscription in the files of A. Brieux and
F.R. Maddison.
Anonymous and Undated
76.
Location    Damascus. Musee Nationale; Inventory No. A.
4485.
   Painted paper on a wooden core; yellow, red,
and black. Diameter 224 mm. No constellation
figures; about 60 stars or star groups indicated
by small circles with radiating lines, or by groups
of such circles. Ecliptic latitude circles drawn in
red; set of six meridians drawn in black; equato-
rial tropic and polar circles in black; ecliptic poles
have small circles about 6° in diameter around
them. Equator indicated by simple line; ecliptic
graduated by single degrees with every fifth in-
dicated by a short line extending on either side
with dots at either end. Zodiacal names written
in red along ecliptic; lunar mansions are for the
most part written along ecliptic at regular inter-
vals; directions and seasons written along equi-
noctial colure; most stars labeled; na^^^f tending
toward ta"liq. The equinoctial colure is gradu-
ated by single degrees with every fifth labeled in
non-consecutive segments of 90° beginning at
the equinoxes and ending at the poles. There is
a network of half-great circles originating at the
intersection of the celestial equator and the sol-
stitial colure at the winter solstice and extending
to the equinoctial colure at 5° intervals. The
hemisphere thus defined is also covered by a
series of parallels (parallel to equinoctial colure)
at   1°   intervals, drawn in  red, with every 5°

parallel drawn in black. The purpose of the net-
work, perhaps unique on extant globes, is un-
clear.
  The wooden horizon ring on a stand of four
round and tapered legs on circular base with
central support for meridian ring is contempo-
rary with the globe. A geometrical floral pattern
is on the base of the stand. The horizon ring is
graduated near the inner edge by single degrees
with every fifth labeled in non-consecutive 90°
segments beginning at the points where the me-
ridian ring is inserted. Every 30° is assigned a
zodiacal house in counter-clockwise direction,
the insertion of the meridian ring occurring be-
tween Gemini and Cancer (summer solstice) and
Sagittarius and Capricorn (winter solstice). A cir-
cle of dots marks the separation between the
zodiacal houses. Near the outer edge of the ho-
rizon ring is a table of place names arranged in
pairs, one written in black and one in red. Six
pairs of place names are aligned with each zodia-
cal house. From the available photographs, it was
impossible to make out any but a few of the 144
localities on the horizon ring. Meridian ring is
ungraduated (? a recent replacement).
  Since the equator is indicated by only a simple
line and the meridian ring is ungraduated, it is
possible that this unusual network of half-great
circles and parallels was intended as an aid to
measuring the right ascension and declination of
a star; the graduations of the equinoctial colure
would supply the necessary scale for measuring
the declination while the graduations along the
equator, given by the intersections of the paral-
lels along the southern half-circle of the equator,
would serve as a scale for the right ascension
measurement. In addition the parallels would
delineate the intervals between the equinoxes.
The celestial longitudes of the stars indicate an
epoch of the end of the eighteenth century; the
celestial latitudes of the stars, however, are in
several instances quite poor. For example, Re-
gulus {qalb al-asad) is on the summer tropic
rather than the ecliptic, and the nineteenth lunar
mansion, al-shawlah, is indicated near the equa-
tor rather than the ecliptic. There are some
similarities in craftsmanship, calligraphy, and the
precise execution of the circles and parallels be-
  
NUMBER 46

257



tween this globe and that made by Muhammad
Ashraf TuqadT Zadah in 1215 H/AD 1800-1801
(No. 72).
See Figure 28 for an illustration.
  Displayed at the exhibition "Science and Tech-
nology in Islam" held at the Science Museum,
London, as part of the Festival of the World of
Islam, April-August 1976.
  Entry based on photographs and information
supplied by F.R, Maddison of the Museum of the
History of Science, Oxford, who suggests that
the unusual network of semi-great circles and
parallels was intended for use in measuring the
coordinates of the ecliptic, by placing the axis of
the globe in a horizontal position and tilting it at
an angle equivalent to the obliquity of the eclip-
tic.
Citation
Maddison Sc Turner, 83.
77.
  Location: London (Greenwich). National Maritime Mu-
seum, Department of Navigation; Inventory No.
G.3.36.379.
  Metal hemispheres joined at equator. Diame-
ter 70 mm; weight 370.8 grams including merid-
ian ring (4.4 X 1,5 mm). No constellation figures;
about 20 stars indicated by inlaid silver points.
Ecliptic latitude circles; equatorial tropic and po-
lar circles; ecliptic tropic circles. Equator and
ecliptic graduated by 2° intervals with every
tenth labeled; ecliptic repeats every 30°; equator
numbered continuously from vernal equinox.
Zodiacal names engraved along ecliptic; poles,
tropic and polar circles, the mayl kulli, and stars
labeled; naskhi. Positions of stars relative to the
equinoxes conform to middle to late seventeenth
century. Circles are precisely engraved, but grad-
uations slightly uneven.
  Metal pedestal stand on circular base with two
crossed semicircular arcs supporting the globe
and attached to the underside of the horizon
ring. Horizon ring is graduated by two degrees
with every tenth indicated by a longer line. Al-
though the rings are graduated slightly more
carefully than the globe, they appear to be con-
temporary with the globe.

78.
  Location: London (Greenwich). National Maritime Mu-
seum, Department of Navigation; Inventory No.
G.6.36.382.
  Metal; hemispheres joined at equator with sol-
dering evident; cracking visible; probably two
cast hemispheres. Diameter 50 mm; weight
141.2 grams including axis rod of diameter 1.98
mm. No constellation figures; about 20 stars
indicated by inlaid silver points. Ecliptic latitude
circles; equatorial tropic and polar circles; equi-
noctial colure. Equator and ecliptic divided by
6° intervals (no single degrees) with every unit
of 6° labeled. Numerals are written sideways
rather than in the same direction as the zodiacal
names. Ecliptic repeats every 30°; equator is
numbered in opposite directions from the vernal
equinox in two segments of 180°, one running
eastward and one westward. Zodiacal names en-
graved along ecliptic; poles, tropic and polar
circles and the mayl kulli, and stars labeled; nas-
khi. Positions of stars in relation to the equinoxes
conform to the late seventeenth or early eigh-
teenth century.
  Three-legged stand on ring base with inverted
semicircular arc attached to ring base to support
the globe. Both the horizon ring and meridian
rings are ungraduated. Non-functional.
79.
  Location: London. The Victoria and Albert Museum;
Inventory No. 1 149-1883. Not assigned to any department.
  Metal; hemispheres joined at equator. Diame-
ter 82 mm. No constellation figures; about 20
stars indicated by inlaid silver points. Ecliptic
latitude circles; equatorial tropic and polar cir-
cles. Zodiacal names engraved along ecliptic;
ecliptic latitude circles, solstitial colure, and some
stars labeled; naskhi. Ecliptic and equator grad-
uated by single degrees with every sixth labeled;
ecliptic repeats every 30°, equator numbered
continuously from the vernal equinox. Positions
of stars in relation to equinoxes conform to mid-
dle of seventeenth century.
  Metal quadruped stand, one leg of which is
missing, on eight-lobed ring base with center
  
258

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



crossbars in the center of which is a pedestal
support with a semicircular arc to support the
globe attached to underside of horizon ring at
the notches for the zenith ring; zenith ring is
missing. Horizon ring graduated by 5° intervals
labeled in non-consecutive segments of 90° con-
verging at the notches for the meridian ring.
Meridian ring is graduated by 5° intervals la-
beled in non-consecutive segments of 90° begin-
ning at the points of attachment to the globe.
Although the stand and rings appear to be about
the same age as the globe, they may not have
originally been made for this globe, since they
are graduated with a different scale and the
precision is not good.
80.
   Location: Oxford. Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/184.
  Metal; has a rough large seam at equator with
soldering visible. Diameter 79 mm; weight 355
grams. No constellation figures; about 20 stars
indicated by inlaid silver points. Ecliptic latitude
circles; equatorial tropic and polar circles; equi-
noctial colure. Ecliptic and equator graduated by
2° intervals with every tenth degree labeled;
ecliptic repeats every 30°; equator is numbered
continuously from the vernal equinox. Zodiacal
names engraved along ecliptic; poles, tropic and
polar circles and stars labeled; naskhi. Very few
stars are near ecliptic, but their positions gener-
ally conform to the late seventeenth century. A
pin passes through the equatorial poles with a
ring by which it can be suspended.
Stand and rings missing.
                       Citation
Maddison [1957], 41, plate XXVIM (photograph).
81.
  Location: Oxford. Museimi of the History of Science,
Billmeir Collection; Inventory No. 57-84/182.
  Metal; there is a seam along ecliptic which can
be felt from inside, although no seam is observ-
able on outside; several cracks; probably two cast
halves. Diameter 131 mm; weight 599 grams. No

constellation figures; 37 stars indicated by en-
graved dots within circles. Ecliptic latitude cir-
cles; equatorial tropic and polar circles, and a full
set of circles parallel to the ecliptic at 5° intervals
on both sides of the ecliptic. Has the unique
feature of having arcs drawn and labeled through
a star (labeled "ayyuq, a Aurigae) indicating the
different coordinate systems. The semi-great-cir-
cle representing the declination circle {dd"irat
mayl) is marked by a dotted line, while the arc
on   which   the   celestial   latitude   is   measured
{dd"irat "ard) is an engraved solid line. In addi-
tion, there is engraved and labeled on the surface
of the globe a circle corresponding to the horizon
ring {dd"irat ufq), another for the meridian ring
{dd"irat nisf al-nahdr), along with the arc repre-
senting the prime vertical {dd"irat awwal sumut),
the circle passing through the zenith and inter-
secting the horizon circle at the east-west points.
The great circle passing through the zenith and
the ecliptic poles is also indicated. Equator and
ecliptic graduated by single degrees with every
fifth labeled; ecliptic repeats every 30°; equator
numbered continuously from  vernal  equinox;
stars are given numbers corresponding to their
magnitudes.  In addition to the labels already
mentioned,  the  zodiacal  names are  engraved
along the ecliptic, and the poles, tropic and polar
circles, solstitial colure, solstices, and the mayl
kulli are labeled; naskhi. The engraving was not
always carefully executed, with several circles
redrawn; solstitial colure was incorrectly drawn
and labeled as passing through the ecliptic at 15°
House of Cancer, which the maker then noted
was an error before labeling the correct circle.
The accuracy of the star positions and graduation
is good, however. By having the horizon rings
indicated directly upon the surface of the globe,
the coordinate systems indicated are valid for
only one particular geographical latitude, in this
case 32° north, which corresponds to the latitude
of Yazd (or roughly that of Lahore). Star posi-
tions relative to the equinoxes conform roughly
to those of the early seventeenth century. Be-
cause of the method of construction (two cast
hemispheres), it is likely that this sphere of un-
usual design is a Safavid product of a seven-
teenth-century workshop in or near Yazd.

NUMBER 46

259



See Figure 22 for an illustration.
  Stand and rings missing, if indeed they were
intended to be used with a globe meant to dem-
onstrate the coordinate systems directly upon its
surface.
  Formerly in Chadenat Collection until May of
1956, when it was sold at a public sale in Paris.
Citations
Chadenat, item no. 44.
Maddison [1957], 40, No. 182, plate XXVIII (photograph).
82.
Location: London (Greenwich). National Maritime Mu-
seum, Department of Navigation; Inventory No. G. 142. NA
9022-40C.
  Metal; hemispheres with seam along ecliptic.
Diameter 180 mm; weight 926.6 grams. Twelve
zodiacal figures are engraved in circular medal-
lions; these are not actually constellation outlines
but only decorative devices. About 20 stars in-
dicated by inlaid silver points. Ecliptic latitude
circles; equatorial tropic and polar circles. Equa-
tor and ecliptic indicated by simple line and seam
of globe; no graduation. Zodiacal medallions,
poles, circles, and some stars labeled; naskhi in
floriated double-outlined writing on recessed
stippled ground. Stars are poorly positioned and
without graduations it is impossible to tell to
which epoch it was intended to conform.
  A technically non-functional globe, but exe-
cuted with first-rate calligraphy and well-drawn
medallion figures by a very skilled metalworker.
The figures representing the zodiacal houses are
not the usual constellation figures, and there are
no stars indicated in them. They are representa-
tive of a different iconographical tradition to be
seen also on globes Nos. 63 and 83. This globe
is strikingly similar in nearly every respect to that
made for Shah "Abbas I in 1012 H/AD 1603-
1604 (No. 63).
  Probably a Safavid Persian product of the early
seventeenth century.
  Has a six-legged metal stand with horizon and
meridian rings very inaccurately graduated by
approximate 1 ° intervals with every sixth indi-
cated by a longer line. Not contemporary with
globe.

83.
   Location: Cambridge, England. Whipple Museum of the
History of Science; Inventory No. 1410.
  Metal; seam along ecliptic. Diameter 121 mm.
No constellation figures but 12 zodiacal figures
engraved in circular medallions. About 42 stars
(unlabeled) indicated by inlaid silver points some-
times having circles engraved around them.
Ecliptic latitude circles; incomplete set of merid-
ians consisting of equinoctial colure and one 30°
to the west; equatorial and ecliptic tropic circles;
circles about 12° inside each of the tropic circles
parallel to both the ecliptic and equatorial tropic
circles; equatorial polar circles. Ecliptic and equa-
tor ungraduated. Zodiacal houses labeled outside
medallions; ecliptic, equator, equinoxes, sols-
tices, the mayl kulli, ecliptic latitude circles, me-
ridians, tropic circles, polar circles, equatorial
and ecliptic poles labeled; naskhi tending toward
ta"liq. Along the ecliptic are 12 circular medal-
lions with the zodiacal figures which represent a
different iconographic tradition from true con-
stellation diagrams. It is very similar in this re-
spect to globes Nos. 82 and 63. The unlabeled
stars are spaced regularly about the globe, re-
sulting in little if any astonomical accuracy.
  The globe is probably of Safavid Persian origin
of the early seventeenth century.
  The globe is placed in a meridian ring and
horizon ring with stand which are not contem-
porary with the globe and of a different alloy.
The horizon ring, supported by three tapered
legs on a ring base, has approximate 5° segments
numbered in an irregular and scrambled se-
quence, with irregular smaller divisions. The me-
ridian ring, also marked with irregular divisions
and disordered numerals, is mounted at the
equatorial poles of the globe and permanently
attached to the horizon ring with the axis of the
globe parallel to the horizon.
84.
Location: London (Greenwich). National Maritime Museum,
Department of Navigation; Inventory No. G.26.36.384.
  Metal; no seam observable; one plug visible;
rattles when shaken for it has a white pebble in
  
260

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



it which is larger than the bored holes. Diameter
60 mm; weight 236.5 grams including meridian
ring (1.75 X 3.8 mm). No constellation figures;
about 20 stars indicated by inlaid silver points.
Ecliptic latitude circles; equatorial tropic and po-
lar circles; equinoctial colure. Ecliptic and equa-
tor graduated by 6° intervals (not single degrees)
with each 6° segment labeled; ecliptic repeats
every 30°; equator repeats every 90°. Zodiacal
names engraved along ecliptic; names of the 28
lunar mansions engraved on opposite side of
ecliptic at intervals of about 12° indicated by a
short line; stars labeled; naskhi. Precisely made
globe. Star positions in relation to the equinoxes
conform generally to those of the seventeenth-
century Lahore workshop.
  Metal quadruped stand, with central square
knob on each leg, on a ring base. A semicircular
arc attached to underside of horizon ring sup-
ports globe; in the center of the arc there is an
indentation with a hole bored through it that
must have been used with a pin to secure the
meridian ring. The horizon ring is graduated by
2° intervals with every sixth labeled in non-
consecutive segments of 90° converging at the
notches for the meridian ring. Meridian ring
ungraduated; later replacement.
85.
Location: (Greenwich) London. National Maritime Museum,
Department of Navigation, Inventory No. G.5.36.381.
  Metal; no seam observable nor felt inside;
three plugs visible including two about 10 mm
in diameter on opposite sides of globe along the
equator; dull finish. Diameter 80 mm; weight
408.1 grams. No constellation figures; about 20
stars indicated by inlaid silver points. Ecliptic
latitude circles; equatorial tropic and polar cir-
cles; equinoctial colure. Ecliptic and equator
graduated by single degrees with every sixth
labeled; ecliptic repeats every 30°; equator num-
bered in two 180° segments. Names of zodiacal
houses engraved along ecliptic; names of the 28
lunar mansions engraved along ecliptic at inter-
vals of about 12° indicated by short lines; stars
labeled; naskhi. A well executed and precise
globe. Star positions relative to the equinoxes

generally conform to those of the seventeenth-
century Lahore workshop.
  Quadruped metal stand, with squarish knobs
having decorative notches on top, center and
bottom of each leg, on circular base. There is no
undersupport for the globe. Globe does not set
far enough down in the ring for it to function
properly. Hence stand useless. Both horizon ring
and meridian ring are ungraduated and not con-
temporary with globe.
86.
Location:    Oxford. Museum of the History of Science, Lewis
Evans Collection; Inventory No. 2913.
  Metal; no seam observable outside nor felt
inside; two plugs visible; slight dent; dull dark
finish. Diameter 100 mm; weight 376 grams. No
constellation figures; 57 stars indicated by small
engraved Xs. Ecliptic latitude circles; equatorial
tropic and polar circles. Ecliptic and equator
graduated by 2° intervals with every sixth degree
labeled; ecliptic repeats every 30°; equator num-
bered continuously from vernal equinox. Zodia-
cal names engraved along ecliptic; stars labeled;
naskhi, tending toward ta"liq. The engraved la-
bels and the Xs indicating the stars have been
whitened to stand out against the dark surface of
the globe. Several holes have been drilled in the
globe, possibly at a later date. There are holes at
both the ecliptic and equator and at about 25°
on either side along the solstitial colure (compare
globes No. 7 and No. 8). Quite precise and
uniform graduation. The star positions relative
to the ecliptic generally conform with those of
the seventeenth-century Lahore workshop.
  Metal quadruped stand with S-curve legs
shaped in a leaf design and resting on rounded
feet. A semicircular arc attached to the underside
of the horizon ring supports the globe. Horizon
ring graduated by two degrees with every sixth
numbered in non-consecutive segments of 90°
converging at the notches for the meridian ring.
Meridian ring is graduated by 2° intervals with
every sixth labeled in non-consecutive segments
of 90°. One half of the ring (from zero-point to
zero-point) is thicker than the other semicircle
and is nearly flush against the surface of the
  
NUMBER 46

261



globe. Under the thinner half of the ring there
rotates a thin 90° arc rather carelessly graduated
by single degrees which is attached at the mid-
point, or 90° label, of the thinner half-circle. At
the zero-points on either side of the ring are two
protruding lips that should rest on the horizon
ring to hold the meridian ring immobile. The
stand is made so that the thicker part (labeled on
the lower right-hand quadrant "southern lati-
tude") must be at the bottom to fit in the stand
properly. The pin holding the graduated rotat-
ing arc would therefore always be at the top or
zenith and could serve to measure the altitude,
and the axis of the globe could be adjusted to
different terrestrial latitudes by inserting it
through the appropriate set of holes drilled
through the meridian ring. There are holes la-
beled for northern geographical latitudes of 21 °,
24°, 32°, 48°, 66° 30', and for southern lati-
tudes of 21 ° (approximate, for unlabeled), 24°,
29°, and 72°. Labeling has been whitened to
stand out against the dark finish of the stand and
rings, as on the globe. A long, decorative pin
serves as axis for the globe; this pin does not pass
freely through the holes labeled 72° and 24°
south and 21 ° north.
  Stand and rings contemporary with globe. The
stand and horizon ring are strikingly similar to
those of globe No. 70 (made in 1083 H/AD 1672
and also to the later globes Nos. 33 and 55. The
design of the meridian ring is very similar to that
of No. 18 made by Diya^ al-DTn Muhammad of
the Lahore workshop in 1055 H/AD 1645-1646,
  Probably an Indo-Persian product, possibly of
the seventeenth-century Lahore workshop.
  Globe was bought from Sayyid Bahadur Shah
in Lahore by H, Beveridge and later sold to
Lewis Evans. Displayed at the exhibition Science
in India held at the Science Museum, London,
March-August 1982.
Citation
Anderson, 34, no. 125.
87.
Location:    Present location unknown.
  Metal; no seam observable; two large circular
plugs of diameters 27 and 28 mm opposite each

other, one below the House of Aquarius and
Pisces and the other above the Houses of Leo
and Virgo; walls appear to be very thick. Diam-
eter 75.2 mm; weight 610 grams. No constella-
tion figures; about 60 stars indicated by engraved
dots within small circles. Ecliptic latitude circles;
equatorial polar circles; equatorial tropic circles
and circles 12° either side of equator. Ecliptic
and equator graduated by 2° intervals with every
sixth labeled; ecliptic repeats every 30°, while
equator is numbered in two segments of 180°
each. Zodiacal names engraved along ecliptic;
celestial poles and nearly all stars labeled; naskhi.
Holes at both ecliptic and equatorial poles. The
graduations are somewhat uneven and the dis-
tances between circles are not uniform. Star po-
sitions conform in general to those of middle to
late seventeenth century, although the latitude
of some stars is irregular. A technically inaccur-
ate and imprecise globe, probably of seven-
teenth-century Indo-Persian origin.
  Quadruped stand on ring base, with center
knob of three rings with two rings at top and
bottom on each leg. Semicircular arc attached to
underside of horizon ring supports the globe.
Horizon ring graduated by irregular divisions of
more or less 1 ° with every fifth labeled in non-
consecutive segments of 90° beginning at the
notches for the meridian ring. Meridian ring
missing. Stand not contemporary with globe.
  Entry based on photographs and information
obtained from F.R. Maddison of the Museum of
the History of Science, Oxford.
  Formerly in a private collection in France.
Offered for sale at Christie's auction house in
1982 (Christie's reference no. CQ 314).
88.
  Location: Oxford. Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/183.
   Metal; no seam observable outside nor detect-
able inside; no plugs visible on outside, but
stumps for casting visible inside; has appearance
of a soft, heavily leaded alloy with possible lead
segregation evident; rattles. Diameter 76 mm;
weight 344 grams. No constellation figures; 22
stars indicated by inlaid silver points. Ecliptic
latitude circles; equatorial tropic and polar cir-
   
262

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



cles; equinoctial colure. Ecliptic and equator
graduated by 2° intervals with every sixth degree
labeled; ecliptic repeats every 30°; equator also
repeats every 30°. Zodiacal names engraved
along ecliptic; stars labeled; naskhi. Graduations
and circles precisely engraved. Positions of stars
relative to the equinoxes are not consistent, for
those near autumnal equinox seem to conform
to those of early seventeenth century while those
around the vernal equinox appear to conform to
the late seventeenth century. A rod passes
through the equatorial poles with end bent at
each end to form small hooks by which it might
be suspended.
Stand and rings missing.
  Formerly in Chadenat Collection until May,
1956, when it was sold in Paris.
Citations
Chadenat, item no. 45.
Maddison [1957], 41, no. 183, plate XXVIII (photograph).
89.
  Location: London. The Science Museum, Department of
Astronomy and Geophysics; Inventory No. 1914-597. Lent
by the Royal Astronomical Society.
   Metal; no seam observable; three large plugs
and seven small black squares about 2 mm across
are visible; rattles; dark finish which appears very
blotchy and discolored in irregular patches. Di-
ameter 140 mm. No constellation figures; about
150 stars indicated by large inlaid silver points,
each within an engraved circle. Ecliptic latitude
circles; equatorial tropic circles in dotted engrav-
ing; ecliptic and equatorial polar circles in dotted
engraving; equinoctial colure with dotted en-
graving. Ecliptic and equator graduated by single
degrees with every sixth indicated by dotted lines
and labeled; ecliptic repeats every 30°; equator
is numbered continuously from vernal equinox.
Zodiacal names engraved along ecliptic; poles,
polar circles, both colures, equinoxes and stars
labeled; naskhi. Graduations and circles precisely
inscribed. Positions of stars in relation to the
equinoxes conform to those of late seventeenth
century. Modern threaded insert has been put
into equatorial poles.
Meridian  ring is labeled and graduated by

single degrees with every sixth labeled in non-
consecutive segments of 90°. One semicircle
from zero-point to zero-point is nearly flush
against the surface of globe. The other semicircle
is recessed from the globe surface so as to allow
for a rotating 90° arc to be attached at the
midpoint, or 90° label; this arc is now missing,
although the pin to hold it is still intact. This
semicircle of the meridian ring also extends out
past the edge of the other half-circle so as to rest
securely on the horizon ring and hold the ring
immobile. In this position the rotating arc would
serve to measure altitude, and the globe could
be adjusted to a given terrestrial latitude by
setting the axis of the globe through one of the
holes in the meridian ring. In the ring, in addi-
tion to holes at the zero-points (equivalent to the
equator), there are holes for northern latitudes
labeled of 24°, 27°, 32°, 51 °, and 66°30', and
for southern latitudes of 18 °, 29°, 35 °, and 72 °.
The numbers of the lower half of the circle,
which should always be below the horizon ring if
properly positioned, are written with the bottom
of the numerals toward the outside of the ring,
while the bottom of the numerals on the upper
half are turned to the inside of the ring. In this
way, both semicircles can be easily read in this
stationary position. The point where the pin for
the now missing arc is inserted is labeled nisf al-
nahdr (midday). Meridian ring is contemporary
with globe and is very similar in design to the
rings of globes Nos. 18 and 86. The quadruped
wooden stand, with wooden ring which does not
actually serve as a true horizon ring, is recent
and non-functional.
Citations
Rothman [1840].
Rothman [1842]; lists some of the stars labeled on the globe
and attributes it to the eighteenth rather than the seven-
teenth century.
Calvert, 2 (photograph).
90.
   Location:    Rockford, Illinois. The Time Museum; Inven-
tory No. 1175.
  Metal; no seam observable inside or outside; 8
small round plugs visible; a very large circular
  
NUMBER 46

263



hole at the north equatorial pole where probably
a large plug was once placed. Diameter 184 mm;
wall thickness 2.5 mm. No constellation figures;
32 stars indicated by inlaid silver points and
numbered and labeled. Ecliptic latitude circles;
six meridian circles intersect the equator at 24°,
54°,84°, 114°, 144°, 174°, 204°, 234°,264°,
294 °, 324 °, and 354 ° (there is no meridian circle
representing the equinoctial colure); equatorial
tropics; two parallel circles at 12° and 20° from
the equator on both sides. Ecliptic and equator
graduated by single degrees with every sixth
indicated by a longer line and labeled; ecliptic
repeats every 30°; equator numbered continu-
ously from vernal equinox. Zodiacal names en-
graved along ecliptic; south equatorial pole and
stars labeled; naskhi tending toward ta"liq. Stars
are numbered with standard numerals. All circles
and graduations were lightly traced on the globe
first and then engraved over more deeply. Small
silver points are inlaid at the equinoxes and eclip-
tic poles and the points of intersection of the
meridians with the equator. Very accurate and
uniform graduations and circles. Positions of the
stars relative to the equinoxes conform to those
of the middle of the nineteenth century.
  The unusual design (with a set of six meridians
shifted 6° westward), the engraving technique of
the circles and graduations, the calligraphy, the
great precision, and the star positions are so
similar to those of globe No. 33 made by Lalah
Balhumal LahQrT in AD 1842 that with consider-
ably certainty this globe can be attributed to
Lalah or to someone trained in his workshop.
For an illustration see Figure 26.
Stand and rings now missing.
Citation
Turner (in press).
CLASS C—GLOBES WITH NEITHER STARS NOR
CONSTELLATIONS
Signed or Dated
91.
  Location:    Chicago. Adler Planetarium and Astronomical
Museum; Inventory No. A 116.

Date:    1238 [H] [AD 1822-1823].
Maker:    Muhammad Sam^ TarT.
  Metal; no seam observable; no plugs visible.
Diameter 102 mm; weight 411.1 grams. No con-
stellation figures and no stars. Ecliptic latitude
circles; equatorial tropic and polar circles; equi-
noctial colure (with the ecliptic latitude circle
intersecting the equinoctial colure omitted).
Ecliptic and equator indicated by simple lines
with no graduations. Zodiacal names engraved
along ecliptic; tropic and polar circles, solstices,
solstitial colure, the mayl kulli, ecliptic, equatorial
poles, equinoxes, and solstices labeled; naskhi,
written with floriated double-outlined writing on
dark recessed ground with a band around each
inscription, except for the cartouche near the
vernal equinox in which the date and name of
maker are given. The latter is lightly engraved
with a stippled background; possibly added later.
The inscription in the cartouche could be inter-
preted in several ways, perhaps reading: "the
w/ork of {"amal) Muhammad Sam^TTarTin 1238."
The top part of the cartouche seems to read
bandah mubarrd (a freed slave), though this is by
no means clear. Globe is probably of Persian
origin.
Stand and rings missing.
92.
Location:    London:  The  Victoria and Albert  Museum,
Indian Section; Inventory No. 06,475 (I.S.).
Date:    1241 H [AD 1825-1826].
Maker:    Unsigned, but bears patron's name.
  Metal; hemispheres joined at equator. No con-
stellation figures and no stars. Diameter 100 mm.
Ecliptic latitude circles; equatorial tropic and po-
lar circles; equinoctial colure. Ecliptic and equa-
tor indicated by simple lines with no graduations.
Zodiacal names engraved along ecliptic; poles,
tropic and polar circles, the mayl kulli, solstices,
and equinoxes labeled; naskhi tending toward
ta"liq. Inscription, which appears to be in same
hand as rest of globe, reads: "Its owner is {mdli-
kuhu) Nawab Muhammad "AIT Khan Sahib [in
the] year 1241 H." It was probably made for
Nawab Muhammad "AIT Khan (also called Madali
Khan), the eighth Khan of Khokand, who ruled
  
264

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



in Khokand in Transoxania from AD 1822-1842.
It has been suggested by the Victoria and Albert
Museum that the globe was made in Oudh in
India, but this does not seem as likely.
  Meridian ring graduated irregularly by single
degrees with every sixth indicated by a longer
line; mounted to equatorial poles of globe. Stand
and horizon ring missing.
Anonymous and Undated
93.
  Location: Oxford. Museum of the History of Science;
Inventory No. 69-186.
  Painted papier mache and plaster over some
kind of fiber core; tan ground with black ink
circles and red ink lettering; lacquered; in dete-
riorated condition. Diameter 178 mm, weight
907.2 grams including brass meridian ring and
axis rod. Ecliptic latitude circles; meridian circles
through the equatorial poles every 60° begin-
ning with equinoctial colure; equatorial polar
and tropic circles; parallel circles 1 ° inside each
tropic and 4° inside each tropic, the latter two
labeled "The Circle of the Start of Gemini and
Virgo/Aquarius and Sagittarius." Equator and
ecliptic graduated by single degrees with every
tenth labeled; ecliptic repeats every 30°; equator
is numbered continuously from vernal equinox.
Zodiacal names written along ecliptic, each name
being enclosed in a double circle; poles, tropic
and polar circles, the mayl kulli, equinoxes, and
solstices labeled; naskhi. A precisely executed
globe. Since it has no stars, the globe's date
cannot be estimated according to precession, as
is the case with all globes in this category; prob-
ably seventeenth or eighteenth century.
  Ungraduated metal meridian ring, not con-
temporary with globe. Stand and horizon ring
missing.
See Figure 27 for an illustration.
94.
  Location:    Oxford. Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/30.
  
Papier mache on wooden core, painted and
lacquered; tan, brown, and gilt paints with black
ink labels; rattles. Diameter 70 mm; weight 57
grams. No constellation figures and no stars.
  Ecliptic latitude circles; equatorial tropic and
polar circles; equinoctial colure. Equator and
ecliptic indicated by simple lines with no gradu-
ations; circles in gilt paint. Zodiacal names writ-
ten along ecliptic; the mayl kulli, solstices, solsti-
tial colure, equinoxes, equator, equinoctial col-
ure, ecliptic latitude circles, poles, and tropic and
polar circles labeled; naskhi, tending toward
ta"liq. Some Persian words mixed with the Ara-
bic; where the mayl kulli are labeled the obliquity
of the ecliptic is given as 23°30' {mayl kulli kah
janubi {shamdli) 23 wa nim).
  Wooden meridian ring with every interval
equal to 5° labeled in non-consecutive segments
of 90° beginning at the two points where it is
attached to the equatorial poles of the globe. Pin
with a hook for suspension of globe passes
through the meridian ring and celestial poles of
globe; horizon ring missing.
Formerly in the collection of M. Henri Michel.
Citation
Josten, 21.
95.
Location:    Present location unknown.
   Metal; no plugs visible; appears to have a seam
along equator; badly scratched surface. Diameter
unknown. No constellation figures and no stars.
Ecliptic latitude circles; equatorial tropic and
ecliptic tropic circles (except both sets of tropics
are drawn about 5° too far away); also a circle
about 20° on either side of both ecliptic and
equator; equinoctial colure. Equator ungrad-
uated and represented by seam (P), while ecliptic
graduated rather crudely into single degrees with
every tenth labeled, repeating every 30°. At
various places on the globe the lunar mansions
are named, but no stars are indicated. Zodiacal
names are written in bold double-outlined writ-
ing in circular medallions along ecliptic; lightly
incised; naskhi. Graduations uneven and circles
poorly spaced.
   
NUMBER 46

265



  Metal meridian ring has only one half of ring
graduated; this semicircle is graduated by single
degrees with every fifth labeled both in abjad
numerals and standard numerals in non-consec-
utive segments of 90° converging at the mid-
point. The graduated half of the ring appears to
be wider than the ungraduated half, and to rest
on the surface of the horizon ring. The globe is
at present mounted at the zero-points; there may
be holes in the ring for various terrestrial lati-
tudes, although this is not visible in the available
photograph. There is a metal horizon ring sup-
ported by four short legs, with knob centers,
about the height of the radius of the globe, which
in turn rest on another ring of the same size
which is supported by four rounded feet; it is not
known if the horizon ring is graduated. There
appears to be no central undersupport for the
meridian ring and globe; however, if the ring
were of the type which is stationary, with the
graduated semicircle resting immobile on the
horizon ring, the meridian ring would support
the weight of the globe and an undersupport
would be unecessary. It may be that this ring and
horizon ring are not contemporary with the
globe, since the globe seems to have been very
badly scratched possibly by a very close fitting
meridian ring, while the present meridian ring is
a considerable distance from the surface of the
globe.
  Entry based on the photograph in the sales
catalog when it was offered for sale in Paris in
1977.
Citation
C. Boisgirard—A. de Heekeren, Instruments Scientifiques
Anciens, Le jeudi 28 Avril 1977, Salle No. 8: Expert:
Alain Brieux (sale catalog. Item no. 669).
96.
Location: Present location unknown.
  Metal; no plugs visible; appears to have a seam
at equator. Diameter unknown. No constellation
figures and no stars. Ecliptic latitude circles;
equinoctial colure; both equatorial and ecliptic
tropic circles; both equatorial and ecliptic polar
circles, with two outer circles about 5 ° and 10°

distant concentric with each. Ecliptic and equator
graduated unevenly by single degrees with every
fifth labeled; ecliptic repeats every 30° while
equator is numbered continuously from vernal
equinox. Names of zodiacal houses engraved in
double-outlined writing in small circular medal-
lions; the mayl kulli and poles labeled; naskhi.
   Ungraduated meridian ring. The horizon ring
and stand are clearly not contemporary with
globe for they do not fit the globe properly. The
horizon ring is graduated by single degrees with
every fifth labeled (pattern not discernible); a
semicircular undersupport for the globe and ring
is attached to the underside of the horizon ring
and terminates in a large finial. The horizon ring
is supported by four very tall slender legs having
central square knobs and X-shaped decorations;
the height of the legs is about one and a half
times that of the diameter of the globe; these
legs in turn rest upon a ring the same size as the
horizon ring. This lower ring has decorative
palmettes on the upper surface and is supported
by four rounded feet. The horizon ring and
stand may be Indo-Persian while the globe may
be a Persian product.
  Entry based on a photograph in the files of
F.R. Maddison, sent to him by Saeed Motamed
of Frankfurt am Main sometime in the 1960s.
97.
Location:    Paris. Private collection.
Maker:    Unsigned; but bears patron's name.
  Metal; no seam observable; no plugs visible.
Diameter 69 mm; height on stand 141 mm. No
constellation figures and no stars. Ecliptic lati-
tude circles; equatorial tropic and polar circles;
equinoctial colure. Ecliptic and equator gradu-
ated by 2° intervals with every tenth degree
labeled; ecliptic repeats every 30° while equator
is numbered continuously from the vernal equi-
nox. Zodiacal names written in large double-
outlined writing along ecliptic; mayl kulli labeled
and probably some of the circles; naskhi. Gradu-
ations somewhat uneven and circles wobble
slightly.
Meridian ring is graduated rather unevenly by

266

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



single degrees with every fifth indicated by a
longer line, but with each larger interval labeled
as if it were 6°: that is, it is labeled 6, 12, 18,
etc., even though only 5° are actually indicated
in each interval, in non-consecutive segments of
90° beginning at the points of attachment to the
globe. The horizon ring is graduated but the
pattern cannot be discerned from the photo-
graph; to the underside of the horizon ring is
attached a semicircular arc for supporting the
globe which terminates in a four-sided finial. The
horizon ring rests on four four-sided legs which
in turn rest on a ring the same size as the horizon
ring; at the tops, bottoms and centers of each of
the long straight legs are square knobs. The
bottom ring of the stand is engraved with an
inscription in four parts, which names the patron,
reading:
Suitable for the salon of His Excellency, the most learned
savant of the ages, mujtahad* of the era and the time, the
Imam of the Muslims, drawing together the derivative insti-
tutes of the law and the fundamentals, combining sciences
based on reasoning and transmitted knowledge,t the model
of distinguished men, the glory of scholars, MTrza Muham-
mad TaqT Khatarat, may his efforts endure.
Ring and stand contemporary with globe.
  Entry based on photograph and description
from a sale catalogue when it was sold in Paris in
1976, and from a photograph in the files of A.
Brieux and F.R. Maddison.
Citation
Histoire  des  sciences.   Livres-Instruments  (sales  catalog).
Alain Brieux. Paris. June 1976. (Item 8551, p. 74).
* ShT"ite authority in religious jurisprudence.
"I" Jurisprudence, theology, cosmology, reports {hadlth) of
the Prophet.
98.
Location:    Paris. Private collection of Marcel Destombes.
  Metal; seam along equator; no plugs visible,
quite light in weight; perhaps raised hemispheres;
badly damaged and dented. Diameter 67.6 mm.
No constellation figures and no stars. Ecliptic
latitude stars; equatorial tropic and polar circles;

equinoctial colure. Ecliptic and equator gradu-
ated by 2° intervals with every tenth labeled.
Ecliptic repeats every 30°; equator is numbered
continuously from vernal equinox. Zodiacal
names engraved along ecliptic in broad double-
outlined writing; poles labeled; naskhi. Circles
well drawn and graduations uniform and precise.
Stand and rings missing.
99.
Location:    Vienna. Private collection of R. Schmidt (as of
1963).
  Metal; no seam observable; no plugs visible.
Diameter unknown. No constellation figures and
no stars. Ecliptic latitude circles; equatorial tropic
and polar circles; ecliptic tropic circles; equinoc-
tial colure (poorly drawn so that it does not pass
properly through the equinoxes). Ecliptic and
equator graduated by 2° intervals with every
tenth degree labeled; ecliptic repeats every 30°
while equator is numbered continuously from
the vernal equinox. Zodiacal names along the
ecliptic, the polar circles and the mayl kulli are
labeled in large double-outlined writing; equa-
torial tropic circles and equinoctial colure labeled
in regular script; naskhi. Graduations are uneven
and the circles wobble, and the intersections of
ecliptic and equator are poorly executed.
  The meridian ring is graduated rather un-
evenly in 2° segments with every sixth degree
indicated by a longer line and labeled in non-
consecutive segments of 90° beginning at the
points of attachment to the globe. The horizon
ring is graduated unevenly by single degrees with
every fifth indicated by a longer line, but with
each larger interval labeled as 6° (just like the
meridian ring on globe No. 97) in non-consecu-
tive segments of 90° beginning at the two
notches for the meridian ring. To the underside
of the horizon ring is attached a semicircular arc
supporting the globe which terminates in a four-
sided finial. The horizon ring rests on four four-
sided legs, which in turn rest on a ring the same
size as the horizon ring. At the tops, bottoms,
and centers of each of the long straight legs are
  
NUMBER 46

267



square knobs. Stand and rings appear to be con-
temporary with globe.
  The design and execution of the globe, stand,
and rings is so nearly identical to globe No. 97
that it is very likely by the same marker.
  Entry based on six photographs in the files of
the Museum of the History of Science, Oxford,
100.
  Location: Oxford. Museum of the History of Science;
Inventory No. 2901.
  Metal; no seam observable; no plugs visible,
but seam can be felt inside along ecliptic; dull
dark brown mottled finish which has been ap-
plied to globe and which in one spot has been
rubbed off. Diameter 89 mm; weight 186 grams.
No constellation figures and no stars. Ecliptic
latitude circles; equatorial tropic and polar cir-
cles; ecliptic tropic and polar circles; equinoctial
colure. Ecliptic and equator graduated by 2°
segments with every tenth degree labeled; eclip-
tic repeats every 30°; equator is numbered con-
tinuously from vernal equinox. Zodiacal names
along ecliptic and the mayl kulli are labeled in
large double-outlined writing; the polar circles,
tropic circles, and equinoctial colure are labeled
in regular script; naskhi. Where the mayl kulli is
labeled, the obliquity of the ecliptic is also given
(23° 30') in abjad numerals. There are holes at
the ecliptic poles and a very small hole at the
north equatorial pole. Appears to have been
mounted on an axis at the southern ecliptic pole,
for the surface has been rubbed by a circular
plate about 45 mm in diameter centered at the
south ecliptic pole. Graduations rather uneven
and circles wobble slightly.
Stand and rings missing.
101.
  Location: Chicago. Adler Planetarium and Astronomical
Museum; Inventory No. A 39.
  Metal; no seam observable; no plugs visible;
rattles. Diameter 69 mm; weight 127.6 including
brass meridian ring. No constellation figures and
no stars. Ecliptic latitude circles; equatorial tropic

and polar circles; ecliptic tropic and polar circles;
equinoctial colure. Ecliptic and equator gradu-
ated by 2° intervals with every tenth degree
labeled; ecliptic repeats every 30°; equator num-
bered continuously from vernal equinox. Zodia-
cal names engraved along ecliptic; poles, equa-
torial tropic and polar circles, and the two mayl
kulli labeled; naskhi. Some of the engraving is in
broad double-outlined writing and some is filled
in in white. Graduations are uneven and circles
somewhat wobbly.
  Metal quadruped stand on ring base with top,
center and bottom square knobs on the straight
legs. The semicircular support under the globe
attached to horizon ring has a screw at bottom
that can tighten to hold meridian ring in place.
Horizon ring graduated by 2° intervals with
every sixth degree labeled in non-consecutive
segments of 90° beginning at notches for merid-
ian ring. Meridian ring graduated by 2° intervals
with every sixth labeled in non-consecutive 90°
segments beginning at the poles where it is at-
tached to the globe by a pin. Rings and stand
appear contemporary with globe.
                       Citation
Yonge, 80; brief entry with no inventory number.
102.
  Location:    Chicago. Adler Planetarium and Astronomical
Museum; Inventory No. A 38.
  Metal; no seam observable; no plugs visible;
rattles. Diameter 70 mm; weight 134.7 grams
including metal meridian ring. No constellation
figures and no stars. Ecliptic latitude circles;
equatorial tropic and polar circles; ecliptic tropic
and polar circles; equinoctial colure. Ecliptic and
equator graduated by 2° intervals with every
tenth degree labeled; ecliptic repeats every 30°;
equator numbered continuously from the vernal
equinox. Zodiacal names engraved along ecliptic;
poles, equatorial tropic and polar circles and the
two mayl kulli are labeled; naskhi with some large
double-outlined writing. Graduations and circles
uneven.
Metal quadruped stand on ring base with hex-

268

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



agonal knobs at the top, center, and bottom of
each tall straight leg. The semicircular support
under globe attached to horizon ring has a very
large wing-headed screw at the bottom (mid-
point), which can tighten to hold meridian ring
in place; the semicircular support terminates in
a four-sided finial at midpoint. Horizon ring
graduated by 2 ° intervals with every sixth degree
labeled in non-consecutive segments of 90° be-
ginning at notches for meridian ring. Meridian
ring graduated by 2° intervals every sixth degree
labeled in non-consecutive segments of 90° be-
ginning at the points where it is mounted to the
globe by a pin. Stand has been repaired. Probably
contemporary with globe.
  Is similar to preceding globe (No.  101), but
engraving not quite as well executed.
Citation
Yonge, 80; very brief entry.
103.
  Location:    Oxford.  Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/186.
  Metal; no seam observable outside but seam
can be felt from inside to be along equator; no
plugs visible; probably of two raised hemispheres;
rattles. Diameter 88 mm; weight 172 grams. No
constellation figures and no stars. Ecliptic lati-
tude circles; equatorial tropic and polar circles;
ecliptic tropic and polar circles; equinoctial col-
ure. Equator and ecliptic graduated by two de-
grees with every tenth labeled; ecliptic repeats
every 30°; equator is numbered continuously
from vernal equinox. Zodiacal names engraved
along ecliptic in double-outlined writing; naskhi.
Graduations are uniform, but some circles are
slightly uneven; generally a precise example of
this style of globe.
  Metal quadruped stand on ring base with top,
center, and bottom square knobs on the straight
four-sided legs. Semicircular support for globe is
attached to the underside of the horizon ring.
Horizon ring graduated by 2° intervals with
every sixth degree numbered in non-consecutive
segments of 90° beginning at the notches for the
meridian ring; meridian ring also graduated by

2° intervals with every sixth degree labeled in
non-consecutive segments of 90° beginning at
points where attached to celestial poles of globe.
Probably contemporary with globe.
  Formerly in collection of C. Chadenat until
1956, when it was sold at public sale in Paris,
Citations
Chadenat, item no. 48.
Maddison [1957], 41, no. 186, plate XXVIII (photograph).
104.
   Location: London (Greenwich). National Maritime Mu-
seum, Department of Navigation; Inventory No.
G.176.36.377.
  Metal; no seam observable; no plugs visible;
noticeably light weight. Diameter 62 mm; weight
149.8 grams including metal meridian ring. No
constellation figures and no stars. Ecliptic lati-
tude circles; equatorial tropic and polar circles;
ecliptic tropic and polar circles; equinoctial col-
ure. Equator and ecliptic are graduated by 2°
intervals with every tenth degree labeled; ecliptic
repeats every 30°; equator is numbered contin-
uously from vernal equinox. Zodiacal names en-
graved along ecliptic in double-outlined writing;
poles, equinoctial colure, solstices, the two mayl
kulli, tropic and polar circles labeled; naskhi.
Circles wobble slightly, but graduations are even.
  Metal quadruped stand on ring base with top,
center, and bottom square knobs with triangular
decorations on each four-sided straight leg. A
semicircular support for globe is attached to un-
derside of horizon ring and terminates at mid-
point in a four-sided finial. Horizon ring gradu-
ated by 2° intervals with every sixth degree
labeled and enclosed in a scalloped design in non-
consecutive segments of 90° beginning at the
notches for the meridian ring. Meridian ring
undergraduated; probably a recent replacement.
105.
  Location:    Oxford. Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/185.
  Metal; no seam is visible, but seam can be felt
from inside to be along equator; no plugs visible;
  
NUMBER 46

269



dented and rippled surface. Diameter 68 mm;
weight 80 grams. No constellation figures and
no stars. Ecliptic latitude circles; equatorial tropic
and polar circles; equinoctial colure. Equator and
ecliptic indicated by simple lines with no gradu-
ations. Zodiacal names written along ecliptic; nas-
khi. Circles wobble slightly.
  Metal quadruped stand on ring base, with top,
center, and bottom square knobs on each straight
leg. Horizon ring graduated by 2 ° segments with
every sixth degree labeled in non-consecutive
segments of 90° beginning at the notches for the
meridian ring. A semicircular arc attached to
underside of the horizon ring supports the globe
and terminates at midpoint in a four-sided finial.
Horizon ring and stand may be contemporary
with globe and closely resemble those of globes
No. 97 and No. 99. Meridian ring appears to be
of later construction; graduated by 2° segments
with every sixth degree numbered in non-con-
secutive segments of 90° beginning at the points
where it is attached to the celestial poles of the
globe.
  Formerly in the collection of C. Chadenat until
1956, when sold at public sale in Paris.
Citations
Chadenat, item no. 47.
Maddison [1957], 41, no. 185, plate XXVIII (photograph).
106.
Location:    Present location unknown.
  Metal; seam clearly visible in photograph. Di-
ameter 69 mm. No constellation figures and no
stars.
  Has a meridian ring and an horizon ring which
rests on four thin legs, with a square knob at top,
center, and bottom of each leg; the four legs in
turn rest on a ring base. The semicircular under-
support for the globe is attached to the underside
of the horizon ring and terminates in a finial.
  No further information available. Entry based
on the short description in a sale catalog when it
was offered for sale in Zurich in 1975.
Citation
La collection Greppin: Instruments scientifiques, Galerie Koller,
Zurich, 1975 (sale catalog). Item no. 3137, p. 37; preced-

ing plate I is a color illustration of a large number of the
instruments grouped together, among which is this globe.
107.
  Location: London (Greenwich). National Maritime Mu-
seum, Department of Navigation; Inventory No. G.2. NA.
9038-36.
  Metal; seam along equator; dented. Diameter
95 mm; weight 220.3 grams including metal
meridian ring (5.0 X 3.1 mm). No constellation
figures and no stars. Ecliptic latitude circles;
equatorial tropic and polar circles. Ecliptic and
equator graduated by single degrees with every
fifth indicated by a longer line. Zodiacal names
engraved along ecliptic; poles, tropic and polar
circles, equator and ecliptic labeled; naskhi. Cir-
cles wobble slightly.
  Ungraduated horizon ring on quadruped
stand 116 mm high; the central support for the
globe has broken off. Ungraduated meridian
ring.
Formerly in collection of Sir James Caird,
108.
  Location: Philadelphia. Private collection of Mr. and Mrs.
David H.H. Felix; Inventory No. B.
  Metal; hemispheres joined at equator. Diame-
ter 72 mm. No constellation figures and no stars.
Ecliptic latitude circles; equinoctial colure; equa-
torial polar and tropic circles and parallel circles
about 12° and 20° either side of equator and
also two parallel circles about midway between
tropics and polar circles (the latter may be the
same as those circles on globe No. 112 that are
43° from the equator and tangent to the ecliptic
polar circle at the solstitial colure). Zodiacal
names engraved along ecliptic; equator, ecliptic,
parallel circles, the two mayl kulli, and the poles
labeled; naskhi, tending toward ta"liq. Ecliptic
and equator graduated by single degrees with
every sixth labeled. Probably a seventeenth- or
eighteenth-century Persian product.
  Meridian ring, possibly contemporary with
globe, graduated by uneven single degrees with
every sixth labeled in non-consecutive segments
of 90° beginning at the points of attachment to
  
270

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



the globe and converging at points opposite the
celestial equator. The abjad numerals are not
written with the bottoms of the letters toward
the inside of the ring, which is customary, but
rather with the letters turned sideways with the
bottoms pointing toward the two points labeled
90°. Horizon ring missing.
  Exhibited at the Philadelphia Museum of Art
in 1977.
  Entry based on photographs and information
supplied by owners.
109.
  Location: Philadelphia. Private collection of Mr. and Mrs.
David H.H. Felix; Inventory No. A.
  Metal; hemispheres joined at equator with sol-
dering visible; dented. Diameter 55 mm: height
with stand 174 mm. No constellation figures and
no stars. Ecliptic latitude circles; equinoctial col-
ure; equatorial tropic circles and parallels about
12° and 20° from equator; equatorial polar cir-
cles. Zodiacal names engraved along ecliptic;
some circles labeled; naskhi Ecliptic and equator
graduated by single degrees with every sixth
labeled.
  Meridian ring missing. Horizon ring of differ-
ent alloy and calligraphy. Ring supported by four
very tall legs with four oblique notches cut in
each leg, which rest on a ring in turn supported
by four short slender legs; there appears to be
no undersupport for the globe and meridian
ring, even though there are two notches in the
ring to allow a meridian ring to pass through.
Horizon ring has the four cardinal points labeled
and is graduated unevenly by single degrees with
every fifth labeled in non-consecutive segments
of 90° beginning at the notches (labeled north-
south) and terminating at the east-west points.
At present the pin passing through the axis of
the globe rests horizontally on the horizon ring.
  Exhibited at the Philadelphia Museum of Art
in 1977.
  Entry based on photographs and information
supplied by owners.

110.
  Location: Bloomfield Hills, Michigan. Cranbrook Institute
of Science; Inventory No. T-29.
  Metal; seam along equator. Diameter 100 mm.
No constellation figures and no stars. Ecliptic
latitude circles. Zodiacal names engraved along
ecliptic; the two mayl kulli \abe\ed; naskhi Several
tears and irregular holes less than 5 mm across
are along the equator, which have been repaired
with plugs of a different alloy soldered into place.
Ecliptic and equator ungraduated. Holes 8 mm
in diameter at equatorial poles.
  Four short S-curve heart-shaped legs support
a ring base which in turn supports a quadruped
stand for the ungraduated horizon ring. The
upper legs are decorated in a twisted rope design.
To the underside of the horizon ring is attached
a semicircular cradle for holding the globe and
meridian ring. Meridian ring is graduated into
36 unlabeled intervals of 10° each; each interval
has a hole through which the pin passing through
the axis of the globe can be set. In this manner
the globe may be rotated by 10° increments to
indicate equivalent changes in terrestrial lati-
tude.
  Bought by the Institute in 1944 from Count
Walewski.
  Entry based on information and photographs
supplied by the Institute.
Citation
Yonge, 86; very brief entry.
Ill,
Location: Present location unknown.
  Metal; of silvered alloy; it is a noticeably heavy
globe; no seam observable; one plug visible. Di-
ameter 89 mm; weight 1450 grams. No constel-
lation figures and no stars. Ecliptic latitude cir-
cles; equatorial tropic and polar circles. Equator
and ecliptic indicated by simple lines with no
graduation. Zodiacal names engraved along
ecliptic; celestial poles, equinoxes, and solstices
labeled; naskhi, tending toward ta"liq; script filled
with yellow coloring and rather poorly formed.
  
NUMBER 46

271



  Metal, ungraduated meridian ring attached by
screws to globe, Ungraduated metal horizon ring
supported by three straight legs attached to a
lower ring, which in turn is supported by three
S-curve legs. Rings are held in place by metal
screws with flat heads.
  Probably an Indo-Persian product of the nine-
teenth century.
  Acquired from a dealer in Bombay. Offered
for sale in April of 1982 at a public auction in
Paris.
Citation
Collection de I'ancienne Maison Gasselin, Instruments de chi-
rurgie humain et veterinaire . . . instruments scientifiques an-
ciens (sale catalog). Nouveau Drouot, Salle No. 8, 28 Avril
1982. Commissaires-Priseurs: J. Lenormand and P.
Dayen. Commissaires-Priseurs, SCP: E. Libert and A.
Castor. Expert pres les Douanes: Alain Brieux. (Item no.
90).
112.
Location:    Present location unknown.
  Metal; may have a seam along equator; no
plugs visible in the only available photograph.
Diameter 110 mm. No constellation figures and
no stars. Ecliptic latitude circles; equatorial tropic
circles; two parallel circles at 12° and 20° from
the equator in both hemispheres; ecliptic and
equatorial polar circles; in each hemisphere there
is also a circle parallel to equator at about 43°
from the equator that is tangent to the ecliptic
polar circle at the solstitial colure. Ecliptic con-
sists of a narrow band with very crude gradua-
tions of approximately one degree with every
tenth, roughly, marked with a short dash outside
the band and labeled, repeating every 30°; equa-
tor consists of a similar narrow band numbered
every sixth degree continuously from vernal
equinox. Zodiacal names engraved along ecliptic;
the equinoxes, solstices, the mayl kulli, and every
parallel circle are labeled; naskhi Graduations
crude and uneven; circles wobble slightly.
  Metal meridian ring has crude graduations of
approximately single degrees with every sixth
labeled in non-consecutive segments of 90° be-

ginning at the points of attachment to the globe.
The horizon ring appears to be graduated and
labeled, but pattern cannot be discerned from
the photographs. The horizon ring is supported
by a semicircular arc attached to the underside
which also supports the globe. A tapered and
turned pedestal stand holds the semicircular arc
at its midpoint; the pedestal stand rests on a
circular plate with scalloped upturned edges.
  Entry based on photograph and description in
the sale catalog when it was offered for sale in
Paris in 1977, and also on an earlier photograph
published in 1970.
Citations
Instruments scientifiques anciens (sale catalog). 28 Avril 1977,
Salle no. 8. Commissaires-Priseurs: C. Boisgirard, A. De
Heekeren. Expert pres les Douanes: A. Brieux. (Item no.
668).
What appears to be the same globe was illustrated earlier by
Guye Sc Michel, 210, plate 200. It is said to be in a private
collection in Paris.
113.
Location:    Paris. Private collection of Alain Brieux.
  Metal; no seam observable; one plug visible.
Diameter 112 mm; height on stand 190 mm. No
constellation figures and no stars. Ecliptic lati-
tude circles; equatorial tropic and polar circles;
ecliptic polar circles. Ecliptic and equator grad-
uated by single degrees with every fifth labeled;
ecliptic repeats every 30°; equator is numbered
continuously from vernal equinox. The solstitial
colure is also graduated by single degrees with
every fifth indicated by a longer line; space is left
for numbering the larger intervals, but it has not
been filled in. Zodiacal names engraved along
ecliptic; equatorial poles labeled; naskhi Gradu-
ations are uneven, and the circles are not all
evenly drawn and appear to have been incised
lightly twice before the final deep engraving.
The alloy seems to have been very soft. Holes at
both equatorial and ecliptic poles.
  The outside face of the meridian ring is grad-
uated somewhat unevenly by single degrees with
every fifth labeled in non-consecutive segments
  
272

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



of 90° converging at the points of attachment to
the globe. Horizon ring is also graduated, al-
though pattern cannot be determined from avail-
able photographs. To the underside of horizon
ring is attached a semicircular arc to support the
globe. The horizon ring is supported by four
slender octagonal legs resting on an eight-lobed
ring based with central cross-bars. Stand and
rings appear to be contemporary with the globe.
See Figure 34 for an illustration.
  Entry based on photograph and description in
the sale catalog when it was offered for sale in
Paris in 1979, and on a photograph supplied by
the owner.
Citation
Instruments scientifiques anciens medicine, chirurgie, curiosites
medicales: Important Collection de Lunettes . . . Instruments
de Mathematiques, d'astronomie, de gnomonique . . . (sale cat-
alog). Drouot Rive Gauche Salle No. 16, 22 Juin 1979.
Commissaires-Priseurs, SCP: E. Libert et A. Castor. Ex-
pert pres les Douanes: A. Brieux (item 155 with photo-
graph).
114.
Location:    Present location unknown.
  Metal; no seam observable; two small brown
plugs about 1 mm in diameter are visible. Di-
ameter 84 mm; weight 500 grams. No constel-
lation figures and no stars. Ecliptic latitude cir-
cles; equatorial tropic circles; ecliptic and equa-
torial polar circles; equinoctial colure. Ecliptic
and equator graduated by single degrees with
every sixth degree labeled and indicated by dot-
ted circles. The equator is numbered in two 180°
segments, while the ecliptic repeats every 30°.
The equinoctial colure, tropic circles and polar
circles are indicated by dotted circles. Zodical
names engraved along ecliptic; polar circles,
tropic circles, the two mayl kulli, the equinoctial
colure are labeled; naskhi The graduations are
uneven, but the circles are quite well engraved;
the sphere itself is very even spherically, and an
example of quite good metalwork. The sphere
has a slight musical rattle, and when the meridian
ring is removed and a probe or the axis pin
inserted inside, the globe at first appears to be a

solid sphere. In fact there is an axial tube inserted
between the two celestial poles through which
the pin passes which is connected with the merid-
ian ring; this tube is then lapped over at each
end at the poles and hammered back over the
surface of the sphere.
  The rings and stand appear to be contempo-
rary with the globe itself and of the same alloy.
The meridian ring is on one side graduated by
single degrees with every sixth labeled in non-
consecutive segments of 90°, and on the other
side by 6° intervals (no single degrees) labeled in
the same manner. In two opposite quadrants
there are holes for the north terrestrial latitudes
(labeled) of 21°, 24°, 32°, and 66° 30', while in
the other two opposite quadrants, labeled "south-
ern latitude" are holes for the latitudes of 18°,
24°, 29°, 34°, 40°, and 72°. The pin passing
through the axis of the globe can be adjusted to
one of the latitudes, keeping the zero-points of
the ring, at which there are also holes, steady on
the horizon ring. The horizon ring is attached to
a quadruped stand on ring base with top, center,
and lower hexagonal knobs on each leg; stand
has later soldered repairs. The horizon ring is
graduated by single degrees with every sixth
labeled in non-consecutive segments of 90° con-
verging at the notches for the meridian ring;
there is a semicircular support for globe attached
to the underside of the horizon ring.
  Probably Indo-Persian, possibly nineteenth
century.
  Formerly in a collection in Great Neck, New
York. Offered for sale on 10 December 1981 by
Sotheby Parke Bernet & Co., London.
Citation
Watches . . . Scientific Instruments . . . Barometers . . . Clocks
. . . which will be sold by auction by Sotheby Parke Bernet &
Co. (sale catalog). Day of Sale 10 December 1981. Item
no. 22, photograph p. 10.
115.
  Location: Oxford. Museum of the History of Science,
Billmeir Collection; Inventory No. 57-84/27.
  Metal; no seam observable nor to be felt inside;
eight plugs visible. It has a greenish-brown mot-
  
NUMBER 46

273



tied finish; one slight dent; rattles. Diameter 103
mm; weight 565 grams with meridian ring. No
constellation figures and no stars. Ecliptic lati-
tude circles. Equatorial tropic and polar circles;
equinoctial colure. Simple lines indicate ecliptic
and equator with no graduations. Zodiacal names
engraved along ecliptic; poles, tropic and polar
circles, and the two mayl kulli labeled; naskhi.
Circles slightly uneven.
  Metal meridian ring graduated by 2 ° intervals
with every sixth indicated by a longer line on one
side of ring, while on other side every 30° is
marked by a longer line. Horizon ring is gradu-
ated by 2° intervals with every sixth indicated by
a longer line. Four slender straight legs on an
eight-lobed ring base support the horizon ring.
Four concave arcs are also attached at one end
to the ring base and meet together at the other
ends and are attached to the midpoint of the
semicircular arc which holds the globe and is
attached to the underside of the horizon ring.
Under what would be the east-west points of the
horizon ring are two pointed devices which are
markedly like the gnomons found on three early
globes (Nos. 3, 4, 6). They are positioned over
the curved supports in just the same manner as
these early globes, but, like globe No. 4, the
concave surface of the leg is not graduated.
Citation
Josten, 20. He attributes it to the sixteenth century or
earlier; such an early date is unlikely; it is probably a
middle to late seventeenth-century Indo-Persian product.
116.
Location: Baghdad. "Iraqi Museum; Inventory No. 9718.
  Metal; no seam observable from photograph;
no plugs visible. Diameter ca 97 mm. No con-
stellation figures and no stars. Ecliptic latitude
circles, ecliptic and equator all indicated by en-
graved double lines filled with hatch marks.
Equatorial tropics engraved as well as, in the
northern hemisphere, parallel circles approxi-
mately 4°, 8°, 12°, and 20° from the equator.
Ecliptic and equator graduated by 10° segments
(not single degrees) labeled; ecliptic repeats every

30° while equator is numbered continuously
from vernal equinox. Zodiacal names engraved
along ecliptic; solstitial colure and the two mayl
kulli labeled; naskhi Circles wobbly and poorly
positioned.
  Metal meridian ring graduated with a fair
amount of regularity by single degrees with every
fifth labeled in non-consecutive segments of 90°
beginning at the two points opposite the celestial
equator and terminating at attachments to the
globe. Horizon ring and stand missing.
  Exhibited at the exhibition "Science and Tech-
nology in Islam" held at the Science Museum,
London, as part of the Festival of the World of
Islam, April—August 1976.
  Entry based on photographs in the files of the
Museum of the History of Science, Oxford, and
information supplied by F,R, Maddison and A.J.
Turner.
Citation
Maddison Sc Turner, 87.
117.
Location:    Present location unknown.
  Metal; probably no seam observable, for cata-
log entry says "turned from a hollow casting."
Diameter 79 mm. Ecliptic latitude circles; equa-
torial, tropic, and polar circles. Equator gradu-
ated by single degrees, represented by dots, with
every fifth indicated by a long line (and possibly
labeled); ecliptic similarly graduated. Zodiacal
names engraved along ecliptic; various labels.
Stand and rings missing.
  Entry based on description in sale catalog when
offered for sale in London in 1957 (no photo-
graph published). Following the sale, it was for a
while in the collection of Landau.
Citation
Sotheby & Co. sale catalog; London, sale of Thursday, 14
March 1957. Item no. 150, p. 28.
118.
Location:    Present location unknown.
Metal; diameter  125  mm.  No constellation

274

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



figures and no stars. Ecliptic latitude circles;
equinoctial colure. Equator and ecliptic indicated
by simple lines with no graduations. Zodiacal
names engraved along ecliptic; naskhi Definite
information not available as to whether or not it
bears a maker's name or date or whether a seam
is visible.
  Metal quadruped stand with graduated hori-
zon and meridian rings.
  Entry based on the one short published de-
scription and photograph.
                       Citation
Schmidt, 23-24; photograph plate X.

121.
Location:    Present location unknown:
  Metal; diameter 132 mm. Ecliptic latitude cir-
cles; polar circles; ecliptic and equator. Numer-
ous Arabic inscriptions. Supported by a modern
baluster shaft on a round base with horizon and
meridian rings. Height of support 320 mm.
  In the collection of C, Chadenat until 1956,
when it was sold at a public sale in Paris,
  No further information available; entry based
on sale catalog.
Citation

Chadenat, item no. 50.

119.
Location:    Present location unknown.
  Metal; diameter 79 mm. No constellation fig-
ures and no stars. Ecliptic latitude circles; polar
circles, equator and ecliptic. Many Arabic inscrip-
tions.
Stand and rings missing.
Formerly in collection of C. Chadenat.
  Entry based on short description in sale catalog
when offered for sale in Paris in 1956. This globe
could possibly be the same as the globe No. 117.
Citation
Chadenat, item no. 46.

122.
Location:    Present location unknown.
  Metal; diameter 82 mm. No constellation fig-
ures and no stars. Ecliptic latitude circles; tropic
and polar circles, ecliptic and equator, all inlaid
with silver. The twelve zodiacal names and the
poles [P] are deeply engraved {graves en creux).
Stand and rings missing. No further information
available.
  In collection of Ch. Chadenat until 1956, when
it was sold at public sale in Paris.
Citation
Chadenat, item no. 49.



120.
Location: London. Private collection.
  "A brass celestial globe, possibly Indo-Persian,
engraved with divisions of the ecliptic and var-
ious great circles, with meridian ring. 86 mm
diam., 18th century."
  No further information available; Entry is
quoted from sale catalog when it was offered for
sale in London in 1972.
Citation
Sotheby Sc Co. (sale catalog). London, sale dated 24 July
1972.

123.
Location:    Present location unknown.
  Metal; diameter 90 mm. Engraved with in-
scriptions. Supported by a quadruped stand on a
base with a central ring and crest supporting the
meridian ring and the horizon ring which are
graduated by three degrees. Height of stand 97
mm. No further information available.
  In collection of C. Chadenat until 1956, when
it was sold at a public sale in Paris.
Citation
Chadenat, item no. 52.

NUMBER 46

275



124.
Location:    Present location unknown.
  Metal. No constellation figures and no stars.
Diameter 72 mm. "Engraved with zodiacal and
planetary [sic] names, numerals and various cir-
cles." Stand has graduated horizon ring with
names of four cardinal points. Height of globe
on stand, 185 mm. "Eastern Islamic. 18th cen-
tury."
  No further information available; entry based
on sale catalog.
  In the collection of M.A. Lee, until 1962,
when it was sold at Sotheby 8c Co., London,
Citation
Sotheby 8c Co., London (sale catalog). Sale dated 26 Febru-
ary 1962. Item no. 14, p. 7. The catalog adds: "Some of
the circles engraved on the globe do not appear to be
meaningful and the inscriptions contain errors.'
INSUFFICIENT   INFORMATION   TO   CLASSIFY
Signed or Dated
125.
Location: Cairo: Was in private collection of Hasan bey
Yazdiasof 1935.
Dated:    1040 H [AD 1630-1631 ].
Maker:    Unsigned.
  Metal. Naskhi script. Date engraved near south
pole in small cursive writing.
  Shown in exposition of Persian art in Cairo in
1935,
No futher information available.
Citation
Wiet [1935], 16, no. 34.
Unsigned and Undated
126.
  Location:    Glasgow. Collection of Arthur Frank, housed
at Charles Frank Ltd., 145 Queen Street.
  Persian celestial globe. No further information
available.
  
Exhibited at the Glasgow Art Gallery and Mu-
seum, April-June 1973.
                       Citation
Nuttall, 44; attributed globe to about AD 1450.
Addendum to the Catalog
CLASS A—CELESTIAL GLOBES WITH
CONSTELLATIONS
127.
Location:    Karachi, Pakistan. National Museum;  Inven-
tory No. N.M. 1957.1049.
Date:    AD 1842, 1258 H and 1899 Vikrama era.
Maker:    Lalah Balhumal LahurT.
  Metal; no seam observable; one large plug
visible near autumnal equinox. Diameter 177.8
mm. Full set of constellation figures with about
1018 stars indicated by inlaid silver studs. Ecliptic
latitude circles; six meridian circles intersecting
equator at 24°, 54°, 84°, 114°, 144°, 174°,
204°, 234°, 264°, 294°, 324°, 354° (there is
no meridian circle representing the equinoctial
colure); equatorial tropics; two lesser circles at
12° and 20° from equator on both sides of
equator; equatorial polar circles. Ecliptic and
equator graduated by single degrees with every
sixth degree indicated by longer line and labeled;
ecliptic repeats every 30°; equator has each 6°
segment labeled with one numeral (1, 2, 3,
4 , . . 60) numbering continuously from vernal
equinox. Zodiacal names engraved along ecliptic;
constellations labeled; major stars labeled and all
stars numbered within a constellation; naskhi,
tending toward ta"liq. Graduations and circles
precisely executed. Near the south pole is an
inscription which states that it was made by Lalah
BalhQmal LahQrT for the education of the son of
one Khushi Ram in 1899 Vikrama era, 1258 H
and AD 1842.
  Modern, non-functional quadruped stand at-
tached to south equatorial pole.
  See No. 33 of the catalog for another globe by
the same maker executed in the same year. Entry
based on the one published photograph and short
  
276

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



description, where maker is referred to as Lala
Balhooljee Sahib of Lahore.
                       Citation
Ashfaque, 222-223, no. 37 and plate VIIB.
  
Metal. Diameter 82 mm. No constellation fig-
ures; unknown number of stars indicated. Great
circles indicated. Zodiacal names engraved along
the divisions of the ecliptic; stars labeled; naskhi
Metal meridian and horizon rings.



128.
Location:    Granada, Spain. Museo de la Alhambra.
  A lacquered papier mache (P) globe with con-
stellations.
129.
Location:    Present location unknown.
  Painted wooden globe with constellations and
stand. Diameter 130 mm; height on stand 255
mm. In 1977 was in the possesion of H,P. Kraus
of New York.

Citation
Ashfaque, 222, no. 36.
131.
Location:    Present location unknown.
  A wooden Sanskrit (DevanagarT script) celestial
globe with stars; has metal stand.
Photographs in the files of Alain Brieux, Paris.



CLASS B—CELESTIAL GLOBES WITHOUT
CONSTELLATIONS
130.
   Location:    Karachi,  Pakistan.  National  Museum;  Inven-
tory No. N.M. 1958.210.

132.
Location:    Present location unknown.
  A Sanskrit (DevenagarT script) armillary sphere
that has 21 stars indicated on pointers.
Photographs in the files of Alain Brieux, Paris,

NUMBER 46

277


FIGURE 88.—View of the southern ecliptic pole of the Smithsonian globe. (Photo: Smithsonian
no. 72-3080)

7. Tables of Basic Characteristics of the Globes
(See pages 213-217 for explanation of Tables. Throughout the Tables globes are referred to by Catalog numbers.)
TABLE 1.—Total number of globes by class

Globe classification
Total
Undated
Dated
Catalog Nos.
Class A




Constellations
58
25
33
1-58
with stars




Constellations only
3
3

55, 56, 58
Class B




Stars but no
32
15
17
59-90
constellations




With zodiacal
4
2
2
63, 73,82,83
figures




Class C




No stars or
34
32
2
91-124
constellations




Unclassified
2


125,126
TABLE 2.—Graduations of ecliptic and equator; globes are referred to by Catalog number

Engraved graduations
Class A
Class B
Class C
'/2° with every
38,53
none
none
fifth line longer



1 ° with every
36
none
none
third line longer



1 ° with every
1, 2, 4, 5, 6,
59,60,61,
96, 107, 113,
fifth line longer
7,8,9, 11,
62,^65,
117

12, 13, 17,
68, 73, 75.


18,20,22.
76.''81


23, 24, 27,



28, 29, 30,



31, 34, 35,



40, 41, 44,



55, 58


1 ° with every
3, 16, 26, 29,
74, 79, 85,
108, 109, 112,
sixth line longer
31, 33, 42.
43, 53,^' 54
89, 90
114
1 ° with every
none
none
93,95,'' 111''
tenth line longer



2° intervals
10, 37
none
none
2° with every
none
66, 69, 70,
none
sixth line longer

71, 86. 87,
88

278

NUMBER 46

279

TABLE 2.—Continued

Engraved graduations
Class A
Class B
Class C
2° with every
57
64, 67, 72,
97, 98, 99,
tenth line longer

77,80
100, 101,
102, 103,
104
5° intervals
32, 39, 45,
46
none
none
6° intervals
none
78, 84
none
10° intervals
none
none
116
Ungraduated
47,48, 56
63, 76,=-82,
91,92,94,95,"


83
105. 110.
Ill, 115,
118
Unknown
14, 15, 19,
none
106, 119, 120,

21,25,49,

121, 122,

50.51, 52

123, 124,
125, 126
Standard numerals rather
31
75
none
  than abjad
Standard numerals
53
none
93 (on the
with abjad


ring)
DevanagarT numerals
Equator numbered continuously
from vernal equinox
29, 54
1,2, 3, 11,
12, 13, 16,
18,20, 22,
none
61, 62,65,
67,73, 74,
75,77, 79,
none
93, 96, 97, 98,
99, 100,
101, 102,

23, 24, 26,
80,81, 84,
103, 104,

27, 28, 29,
86, 89, 90
112, 113,

30. 31, 32,

116

34,40,41,



42,44,45,



46, 55, 57


Equator in two
38
68, 78,85,
114
   180° segments
Equator in four
39,43
  87
71
none
  90° segments
Equator in 3 segments of 100°, one of
60°
4, 5, 6, 7, 8,
35
59
none
Equator has each 6° segment
  assigned a number 1 through 60
Both equator and ecliptic
33, 53 (cf.
   54)
none
none
60, 88
none
none
  repeat every 30°
Ecliptic labeled continuously
11,^38
none
none
from vernal equinox



No labeling
7,8,9
none
none
  of ecliptic
No labeling of either ecliptic
10
76
none
  or equator
Solstitial colure
16
74
113
graduated



Equator only.
' Ecliptic only.
Second set repeats every 30'

280

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

TABLE 3.—Accuracy of globes

Accuracy of engraving
Class A
Class B
Class C
Uniform and accurate
1, 2, 3, 4, 5, 7,
59, 64, 65, 66,
98, 110, 111,
graduations
8, 10, 11, 12.
67, 68,
69,
114,'' 118

13, 16, 17,
70, 71,
72,


18, 20, 22,
74,75,
76,


23, 24, 26,
77, 78,
79,


27, 28, 29,
80, 81,
84,


30. 31, 33,
85, 86,
88,


34, 35, 36,
89,90



38, 39, 40,




41, 42,43,




44, 47, 54,




55,57



Ungraduated, uniform
48
none

91,92, 94
circles




Technically inaccurate and
6, 9,32, 45,
60, 61," 62,"
95. 96. 97, 99,
non-functional
46, 53," 58
63,73,
83,87
82,
100, 101, 102,
103, 107, 108,
109, 112, 113,
115," 116
Graduated evenly, circles
none
none

104
  poor
Ungraduated, circles
56
none

105
  poor
Unknown
14, 15, 19, 21,
25, 37,49,
50, 51, 52
none

106, 117, 119,
120, 121, 122,
123, 124, 125,
126
" Moderately inaccurate.
" Graduations uneven, but circles good.

NUMBER 46

281

TABLE 4.—Presence of circles other than ecliptic and celestial equator

Type of circle
Class A
Class B
Class C
Equatorial polar circles
2, 3,23, 26,
60. 63. 64. 66,
91,92, 93, 94,

27, 28, 30,
67. 70, 71,
95, 96, 97, 98,

33, 36, 43,
72, 73, 74,
99, 100, 101,

54, 56, 58
75, 76, 77,
102, 103, 104,


78, 79, 80,
105, 107, 108,


81,82, 83,
109, 111, 112.


84, 85, 86,
113, 114, 115,


87,88, 89
117
Equatorial tropic circles
16, 22,23,26,
63,64,66,67,
91, 92,93,94,

27, 29, 33,
70,71,72,
95, 96, 97, 98,

36,43,53,
73, 74, 75,
99, 100, 101,

56, 57
76, 77, 78,
102, 103, 104,


79,80, 81,
105, 107, 108,


82, 83, 84,
109, 111, 112,


85,86, 87,
113, 114, 115,


88,89, 90
116, 117
Lesser circles at 12° parallel
57
63," 83, 87
none
  to equator
Lesser circles at 20° parallel
none
none
93, 95"
  to equator
Lesser circles at 12° and 20°
33, 36. 53, 54
70, 72,''90
108, 109, 112,
  parallel to equator
Ecliptic polar circles
2, 16, 30, 36,
60,72, 89
116
96,'' 100, 101,

43, 56

102, 103. 104.
112, 113, 115
Ecliptic tropic circles
none
63,72, 77, 83
95, 96, 99, 100,
101, 102, 103,
114
Equinoctial colure
16, 31,43,56
63,64, 70,76,
91,92,93, 94,


78, 80, 83,
95,96,97, 98,


84, 85, 88,
99, 100, 101,


89
102, 103, 104,
105, 109, 114,
115,118
Meridians
33, 53, 54
63,^72,75,76,
83,' 90
93
Lacking ecliptic latitude
circles
none
72
none
Solstitial colure labeled
10
63, 79, 80, 81
91
Mayl kulli labeled
29, 57
63, 73, 77, 78,
81,83
91,92,93, 94,
96, 97, 99,
100, 101, 102,
104, 108, 110,
112, 114, 115,
116
Obliquity of ecliptic specified
none
73
94, 100
" Also parallel to ecliptic.
'' Ecliptic and equatorial.
' Incomplete.
"^ Plus circles 5° and 10° outside.

282

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY

TABLE 5.—Additional properties

Property
Class A
Class B
Class C

Lunar mansions
2. 11. 38
59. 60,
65, 74,
95"




76,84, 85


Lacking zodiacal names
32. 46,
56
47.48.
none

none

Star chart of Ptolemy cited
2, 3

none

none

Star chart of al-Suf T cited
6, 7, 8

62

none

Star chart of Ulugh Beg cited
14, 31

65

none

Calendrical eras cited other






than Hijrah






Coptic
31

none

none

Era of Flood
31

none

none

Alexandrian Era
7, 31

62,65

none

Yazdijird
7,9, 31

62.65,
74
none

Reign of Akbar
10

none

none

Reign of JahangTr
11, 12,
13
none

none

Reign of Shah Jahan
23. 24

none

none

Malikshah (JalalT) Era
31

none

none

Vikrama Era
33

none

none

AD
33

none

none

Named, but no stars indicated.
TABLE 6.—Construction of globes

Material/fabrication
Class A
Class B
Class C
Wood or papier mache covering
36, 56
72," 73, 76
93, 94
on wooden core



Hollow metal, no seam visible.
6, 9, 10., 11,
60,66,67, 68.
111, 113, 114,
and plugs visible
12, 13, 15,
69, 70, 71,
115, 117

18, 20, 22,
74,84, 85,


23, 24, 26,
86, 87, 88,


27, 28, 29,
89,90


32, 33, 37,



38, 39, 40,



41, 42. 43,



44,45,47,



48, 54


Hollow metal
1, 2, 3, 4, 5, 7,
59,61,62,63,
92, 95,96,98,
hemispheres
8, 16, 17,
64,65, 75,
100, 103,

30, 31, 34.
77, 78, 79,
105, 106,

35, 57
80,81, 82,
107, 108,


83
109, 110, 112
Metal, unknown whether
14, 19, 21, 25,
none
91,97, 99, 101,
hemispheres or seamless
46,49, 50,

102, 104,

51, 52, 53,

116, 118,

55, 58

119, 120,
121, 122,
123,124
Unclassified—125, 126



" Possibly painted ivory.

NUMBER 46

283

TABLE 7.—Globes with apparently contemporary rings and stand

Engraving of graduations
Class A
Class E

Class C
Graduations of horizon ring





1 ° with every fifth line
5, 7, 13, 18, 27,
62, 76


99,'
longer
28, 31, 34,
40




1 ° with every sixth line
3, 29, 33, 42,
none


114
  longer
2° with every sixth line
    54
32, 45
70, 71,
86

101, 102, 103,
longer




104, 105,
115
2° with every tenth line
6, 16
72, 77


none
longer





5° intervals
none
79


none
Line at every 30° interval
none
60


none
Nonconsecutive segments
3, 5, 7, 13. 18,
62, 70,
71,
72,
114
converging at notches
27, 28, 29,
77,
79,
84,

for meridian ring
31, 32. 33,
34, 45, 54
86



Nonconsecutive segments
42
59, 76


99, 101, 102,
beginning at notches




103, 104,
for meridian ring




105
Consecutive segments of 90°
6,40
none


none
in clockwise direction





Consecutive segments of 90°
16
none


none
in anti-clockwise direction





Notches at East-West points to
5." 13, 18," 27,
70," 71
79

none
accommodate a zenith ring
33,^' 53," 54




Gnomons on stand
3, 4,'' 6
none


115
Graduations of meridian ring





1 ° with every fifth line
18, 31, 40
none


97,' 113, 116
  longer
1 ° with every sixth line
3, 33, 53, 54
89


92
  longer
2° with every sixth line
longer
32,45
86


99, 101, 102,
103, 105,
115
2° with every tenth line
none
72,77


none
  longer
Nonconsecutive segments
beginning at poles
5, 31, 32, 33,
45, 53, 54
72. 79


94,97,99, 101,
102, 103,
105
Nonconsecutive segments
6
none


113, 116
    converging at poles
Total globes with contemporary
rings and stand
3, 5,6(?), 7, 13,
16, 18. 27,
28, 29, 31,
59, 60,
67,
71,
62,
69,
72,
66,
70,
76,
92, 94, 95,97,
99, 101,
102, 103,

32, 33, 34,
77,
79,
84,
104, 105,

40,42,45,
86,
89

108, 113,

53, 54



114, 115,
116
" Ring extant.
'' Rings and stand not contemporary.
' Every fifth degree line longer, but labeled as 6° intervals.

8.    Major Inscriptions on the Globes
  The Roman numerals in the margins refer to the catalog entry numbers
employed in Chapters 6 and 7. The editing conventions are those of the
Leiden System applied to the editing of Arabic inscriptions; see Burgoyne 8c
Abul-Hajj, and Dow.
284

NUMBER 46	285
ei«> '»    \sb^]eiA ^   \.^\]iss.\ >.<.«.>^   ^^ W^  ^Jw^t <w5t^t M'J-^    ^^^    •''■-*^    ^*
II
—   -	''^	-	—	"11



I I I




~CJ J>^       Jjt^     A*.M     ^   2
^^JJ^J^     a.^    ^>;

IV

^^1

U:JjJ    i>^    ^   Cr^    -^-^"^    ''■"^^   ^

\J^    ^\     U-^    ^^^J^   3
j>L   ^y^  jL	H*
UJ^;    A..;.A»-P^	I]5
U* ,h«v^ J   A.^^ [^	He

VI
VI I
VI
IX

NUMBER 46	287
3    \K\
^J («' '       i^ -i^       <>*-;'' ^^Ja-* if'      >U.>.l      s.>i-o.>.      A^^^Jba      J*, A»      fr _jJw      3



XI

XI
(>t'     15—^   l>;'      Jo-SN*      *:;W     «>Urt«J'     Jj9'      A.«>^^       ¥3 0 at       (J>>^       1  • f ©       ''^^^-^



XIII





XIV

000
XV

J

U,.^!      ^^i*^'      >,-^a.*, ■•>     1



c><   ij"^ i>   r^

2

^^>u.»    Jj^\    3

288	SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY
^ j^UJb   4
fJj>-^     ) • 1 1      «^^^    J   6
XVI
\jS>ji*       A*.J>^     v_^<^       AJLJ^-tO     1
) * o*      O^j     •^•■*^    2
XVII
XVIII

LT-!-*    i><



J*j>^     CK^-J'     *L>~^    1 :>■«>,* .•^    ^2fli'   '
jlo^l
cr;*   iyt'^   Cn^    '^'-^   f^^    cr;    2

(jj j^i      ^ yU^    ^j.ijJa^^   3

cJ>*-*'     ) • o o     <^"*~-    t/   4
pb    u.'
a»j>^      C'rf-tJ'      'L«-.J>    <-.a *  .^^     t;^^    1XIX



i^>>-ft     ^ . oY     <*^—'    (^  3
JA^^      ^te      ^^'      OkAi*      Os^-J'      *Lcw9     jL>rt«J\       Js\       Ja*    1
{Jj>^       ) •  1 •       AA—    (^ J _^ 2    3

XX





XXI
XXII

NUMBER 46	289
iSj>^     ) • "lY    A^—     J  <Sj>^^   tyjtf^   ,^^jk^^    j\jL^^     ^^
XXIII
tfUl    (j^**"    r T     «iJ;"-<-»  i_y y->-
XXIV
tfU    4;;L^^    Y- X     "^j^-   [; ur >^
XXV

XXVI
^L      ^j^t      J«A«     <^   wte    ^        j   .Jaj>«ia>	jj^t      Ja.JN«       ^j:;-iJ^      jL~i>      jL-jJ^       Ji*       J*P
XXVII
                                                                                                                XXVIII
^;^l   ^.,.|>,,fr    ^L    ^^.-*    -UA-    «fe    ^j^'    Oajs*    ^>;aH    j^~^   JU-«J'     J5*    J»s-
C^
XXIX
^ . Y i	tf .A..«'       (   ' 4-^    ]      (^ ^^>^       r(}\^     J«bA«      J.^
XXX

290	SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY
A; J^ui       tf^ >>-'*'        1   •  1  ♦	AA~*»      ^   5
XXXI
\,\\   Uj-ifrj   A:L.\Dt   ^»^4   Q^    V_J-AJ* JLSJLJ t    L;   tf^'   tfOA    J»p   J-S"
(JL^J^aJt     ^jiijt    ^3^L«      ^f-Ao    *^y»«»L3 jj    j_j-ujl    i>-«J*-    U'^^      »-«-j^    j_j -Uj?^;<c-J'
tf,A.jV9^    lA»r     A*—«    ^j,«      a jb*     Jj»       r^^'-^     dil J J    ^A-Bas.!'     ij^yCij      ^..uniJO' J
J»j^1)e,A^^4  ^j,   ^Ql   J^t   ^^   Jl  ^^»  j^\   J,^- ^:fc,  ybj
tftA,-M.   J -*    u*    1 •    .?   tf A..hA.i    ) j^ ^ Y   A^^-*«  '--''-t*-^   ^ 1 .?   •*:; >*^    MIX    tf-v-«>
o o
JJO^ I   ^.jtj J   ^ I   \.'n.>^-K J   tf^ ^^'^'^    tf^>. 1.11    J   ) 1 1 V    ^''^—^   ^-•>J ■    ^/uo    .^-f--»
jNi.r>^l ■   j>'^ '   <    J* .^ ''    y j>-   I     (jj J* »>■'   <     tf..^Al <    dJt; jjt.   t     ;_^ tfj« *.■»   < tww UjjMi       i   i3./^    )
<   J^w J (  ^ ac-   ^(.^jA..>. \  <  tf	iu uaJ e    < c Vi^ <  AO L^ <   LRJCS.4S I  iyy <  tfJ U^  i   ry^ji'
oooo
< JA-U'    ^^.no   «    ^jL*.^;.^'   «   A4«L>.   <   ^>f-»   «    (j   "* >**'        "'^   *   J—«AJ'     J">^->>   *    (3^J^
                                                             oo o o
(    A».*>>.   <   tf 4J>.   I   J^ j   <   tfJ—JLi J   <    ^jL jlfr t   y y>   t   jj, J^^      -'^   *   ,^ J >^
Lr^*-^^ *  0*i-/^ 'Lr^^* *  u* >* *  O*—^ 'LT-*^ ' cP*^  ( V.r^ )  *  j^^*^^
I    tfJL>>a   <   _^i<a«   <   JEIVA.^ J   (     .J0 ■ UJ   t   A y^   <   Js ^A..,^)    <    AJL>> y   t    tf—^ V   <   '. - . ^'
(     U.>>   <    A LlJ <     (3-^ "*   '        ■•*■* ■'^   <    <^1 A .La,y    (    ^ jj?   I     mJL ^    <    AxJL&   (    A} j j'    (     J>*-' '•A.o'
«   tf^ JA A.' ■    tfJw a^i    tf.wlf-    a^>Lw«   «   nu'j   (   . K«.   4   ^ U^   I   tf A,.r>.U   ( >■ Jj».   I   A.>- 9   </j£)*.>>
* CX 3 j—^   *   AA..A«JUJI   I   iS y   I     Jl JA/   4   <^..A._fc   4    ^'>^   (    Jr;y   <    JiOjaJ'      (    JU^ ^

NUMBER 46	291
XXXI I
*****	.1	l|_	ooo	ooo
o	o
XXXI II
^L|JuJt     (JutJ J     ^*      ^je* •^     <j	AIO-V A ■      tfJ      tf^^.<l&     tfA^Oa^
J-tA* J      -ii'     tf^    L5 .J^    ^      ^-t^.     ^     U^*	f**^->*     <'je*-',?    ^.i'j*-»     ^jt;   -'j   2
>-^Ua-3L     tf^l» J   u° ^ J^    X^S"* 3	r 3 J^^     tfiJaA-* J     ^L^jJl
,j^    y     tfu      ui'     »iJ^>Ji-»    tfirfUJLil    y3'^    c^      J>*-»	jit 3 "^ 3     J * ■^'      f^'^3    ^-"-ij    3
iV ^'^-^	^^   9       t-bw A  i<	V^_Au	*« ^"^^       O fc«l-<h   4	J«i ■< .^^	W      iiti     iS A   A ,» J        >■>■ »^	tf^
Ui'     J y>. yt     Jxmj     J J      T- 3 Ji     t^^3    rt ^   ^^      C J y^ J     y y.JXi ^     1^3*"^     ^'^ y\t 3   4
>'*^  •    Cj-Ju  (jl.^.-^«.»;    tfS    tf -><9^   v^.^   .? -'   ^.? ^ '-'■"■;■.' i    J j-<r^ 3
</*-f-^.>      1^^-?      t5^^3     C*^*     t^"*^     -^     >^-*-i
t..A.>.L0       tfi^A.M)       ^ ^»JcJ     (^A ii<       J yCL>.      y jjfcj»,w	J>..«\J      ^       AJ->      ^U.>-'        *^<       ^     5
^L ^      J*      ^J     (jOJ-OjiJ*      JP     j» >l*     ft*^       J-"       ^"^        (    *J—^)       <^*>f-'*
J<o j.^JL    tfjj    aLji-frL     oJ*    ^jli'    tf^^^ J    yJsu   ^    A^ ju-jt J   OJLA..>   jy—*rf   6
^ jj,  \      tJ     aLaJbj     <^ yc yt     L-*.>>Lo
c- JL^      vacat      j    ^^     "^ 3'' 3     '^^^ •"'■^ J    ^  ./"t^    c**^    _^ a     jujy
vj^buLfi) J     tf l^J^ J      U^yJ)     y 'j.^X>     ^U:L-« J
^^ J    ^ a    ^ >-<-^   J -^,?    J-t^ .^    ^ •'-«' /^ Ou-lA J    jl >.^    (i** .^ .?    Sf>^  ^
AJ-^J I ^A.j   OM« y ^'

292	SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY
CJL     ^_-WCJ y      9
                                                                  XXXVII
ooo  o
   \jJ>J)«,«.ij       U^   oe       ry      A.M.UtJ*      y^      AU <        J*        >AJL^&J •       Aj j.nA.«>' j      tfj J>. ./> A ^M» *
tfj^Mj    tf«,^_^    tfU'     U ^i«.>'j    (j—**•    j_>i    tjL*-^'      OA-*'     tfu'j'      aA->.j J      AJU j      A«jl*
XL
4     Jl JA;   I    dS y   I   y yj^   t    0^ \y   i    ^jl jj^.SJ      I     JC) y   I   y-^y^    l    J.i£> jo	JX;       WM'      1
4   a; ^L^ 4   ^  \ xs)    jj y j.^ 4   J Je « j^y^ 4   JL 'j ^ ^^    <  (j* -^'"^'-^	*  ^y^.
41^ a ,..4—.mJ ■     '■l.A,w   4   ^j^ j^^     J>^   *   «1« n g»    ^'   —   4   J 3-'j   A.W J   ^	J^w      VA.M>I     2
<    (^j*;;    <    jj ap  4    ^,^30    i    Aj> J iJ* \iii. I    4 j_y>J a^ I    4   ^-^jJUa     i>*-y '^Z*'   *    (j' J ./*"*    *	^ "*■■{■»
>-J->-    4     I* IA .l,B >    4     ^   *      - J    4    ^ L^ /yJu I^J?    4     tf^  y.jr*    4     ^ JUL>MP   4	JA^ ^
LIX

AI:

tfLA«w>    O    ./^     T^  "^   cj .^    ^     a.A^    fjfj    jJu    AJLJ^.^



o
iLl     JLsj^.. w     Jfr    a«j>M    ^;;^    tfJ^aJ*	JLP   ^^    AJ ^aJ'    aLt   2
oo	oo o o
) Y' • 1         (**^^)      ^-^■*-i      A:*	^^lAu      AJJI      Aa^j     a«-«J    3
>1-^    (J?    K/^ ^^     C^X   2
.?    ./'**'    u^    AA^.»   3
o	oo     o	o
J^     X   ci 3 '^- *	1
JA\-K,I'    J«S   iH	2
J ycu<^\      J y^.^\      li.Lufr	3
lrf\« W A>,w  _J        ^^.^      r".   .   »   .     tfA   ^      „	J 3-f^       i^	4
<^   >*'     f^ X	1

LX

NUMBER 46

293

S>\J\  o^-i*^* 2
iiL-   ^\    JaUH   3
LXI
i>i
^,*^'    (rt~^}    c^-^ 3
LXII
Cri
^^J:^^\     y^     Cy    2
J^Uw    .-.AU>   3
tfjL^f     aA.;   h-ot^f    tf aA   f 4>«^	1
^j^...>>n    ^^'^     J yo    J*    tfj     ^J'>kl*	2
ai^   tfJL*.   ^y    ^^^.^1*    u*-*-^*   -*^   i>;	3
tfL; J j>. J _)AJ 1    A jjo J    tfu yf^'	4




LXIII

^^'
-— 1


4>AJ:»XAJ»
oUaJL.   2

tf aA>{
J JUJL-J' J
^J^lA.    3


^V^  ^>
,   tfU *


oi
U*:U 5

W n   6
LXIV

O^ j

d " i' v

LXV
L|J \yl3\     J    tf JL J   jjo   w5t^ *   tf -U   cufc*-^ 1
      ^   >.J<    J-oj.    J*     ^*-'  2
afr^    tf^^p^'    y^   A-i—   ^y 3
J      d^ J y^ J ,)-w '   4
tf:,^a;cl-'^*     (*"^)    >^ 5
LXV
^»>.A,P      ^L    ^;;^*      aaj»«a      wfe      ^;^'      a«JK>      tjrfjjl      JLA-^       J«*   1
V^-'i^'^    v/>;^    ^^j'^A-*    -"^-HJ*     ^   6-^'   2
ii>*-*^   ) • 6A   *^-*- <y 3




LXVII
JL^-AJ*     t-i-g^i      tf-B ■*■>.<:'
1

Cr<»     AUI
aki
2

jj\iJi
JA-P
3

j; lyi^^     ^
,_>^n
4

AJUI
>
5


>;J
6

A^    7
                                                                   LXVIII
J A'L.B    Jo.*-    Q^}   aol^   jL-jJt    JAJ, , ^    -^  ^ " ^ ' ^  1
T 1       0^ -» '^*    (J?    je ^ 3
                                                                                                  LXIX
).y.      A^^    J     <Jj3^^      i^ 3:!^
                                                
NUMBER 46

295



.     ^   a»jN«   ....   oiL   . .   1

LXX



Cn^   cr^

JiiU ^\

.^   M. .-^l

4J-jJ*      *IA^     JL-JJI     JL>.\      J»P   1
^jytfcA       jljtfjjl       AA—1   2
> -AY    A^—   3

LXXI

LXXII
(-9   WUl'	lJ»  TN»	A» Mil	\
6j\ J     ^^J\sy    2
\\\ o     *-i—'    3
J-*   1
LXXIII
IkL)    J     ^\    cy\   o'^,*^^    f>»^S   >J'j cr-^*j    u:^>UiJ     UL-;^ 2
                       ^:^UJ»   ^j   ^\    d^Uj   ^'^'j    jLJ\     (^It.
^>-^»   AL»   J^LJ- ^U'il»  Jj^   ot^   ^^-J»j   cM»   eji:^j-^»  J   1} i
JWAJU*   >>«J*   je-i-A: vdJ j L,J ^>,Ja.«, -)   (J >>-: j_,«*-J*^   ^;^UJ^
aap    i ^JJUAJ    dJJ   *-a_-AJ'    ^;; ^>.^\5   aU   ^yL*.    JjlL.    tfL^a—3    ^*^ 4
ti^^-lJ*    »JUJI   ^^A.A..,.'K.M   ^   JUJ^   ^aUJt   ^y*.'^*   aA-«Jl   a*   ^   o*: 5
*lJLlJi ^    AA..»JUI    OU'    A«ja^    1^3  [y^ j''^^    A* aa»^

u J Uw  6
1 TTI

tV J^'   LJ •'*V    -'■t-"     U*     Oi J*^3    ^-*^

r>«   cy

tfJ I   ^	tfbAj*        tfll I        ^L.«>	tfw   >AJLJ t	<^>>-A

296

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



(jr-*-*     -'>»'-'>    Ml*

^   aik>^    JAP

LXXIV



1.5 >»-^     Mil      *^^
^;;A,J«..I<     ^»J   1

LXXV



.».i^«

VJIA .A. A .U^-V

 ^^ ^U—     tfy     j_^      jJL'   4
f    ?■*-; V    Mil     tfLA«*»   5
XCI
■^^A«o    tf a/bw   \
iJ y        ^y—c«.        ao.»«a	2
oo
^     J«*	3
1 TrA	4
XCII

-r' >
< v   tfllU 1

"Lo   ^^UoJU^

tS >>-^     Mil     ^^^-^   3
-      O^JJ-^^      «!L-«>U    wUc>.    ^yX>^      ^}i^    1

XCVI



X
Ai^

■^J=*^   J^

■\J_^\

'>UiJl

>-»   J



(J >^^     ^ ♦ 1 •     ^-*—

cxxv

NOTES

Chapter 1
   ' Neugebauer in HAMA, 577 notes that this is perhaps
not "natural" and certainly not universal, since the Egyptians
perceived the heavens to be a flat roof.
   ^ For general discussions of globes in antiquity see Fabri-
cius, second ed.. Volume in:465-466; third ed.. Volume
V:297-306; Thiele, 17-56; Schlachter; and Stevenson,
1:14-23 (very unreliable). See Boll [1903], 88-90, 153-
159, for reference to use of globes in astrological literature.
   ^ Diog. Laert., 11,1. Kirk Sc Raven, 99 at first interpret
this as a celestial globe, citing Suda, who says Anaximander
wrote a treatise on a celestial globe {aipaTpa) but later they
(103) reject it as improbable since inconsistent with his
cosmogony; see Diels, Frag., 1,81. Kahn, 89-90 accepts the
idea that Anaximander made a celestial globe, possibly a
"solid" type like that attributed to Eudoxus and Aratus or a
flat chart of the heavens. See also Gundel Sc Boker, col. 529.
   * Cic. Repub., l.xiv. Here he reports the statement of
Gaius Sulpicius Gallus through the figure of Lucius Furius
Philus who was born about 180 BC.
   * Cic. Repub., l,xiv; English translation is that of Sabine,
C\c. Repub. [trns.], 119.
   ^ For Eudoxus see Neugebauer HAMA, 573, 675-689;
Toomer [1], and Huxley DSB.
   ' For the life and writings of Aratus, see Knaack; Ludwig;
Barber; Taran; Lesky, 750-752; and Arat. Phen. [3], pp.
185-189.
*	See Walbank; Tarn.
   ^ Mair in Arat. Phen. [3], pp. 198-201 counts 47 constel-
lations but omits Corona Australis, which is clearly described
(lines 400-401); Lamb in Arat. Phen. [trns.], 21-29 counts
44 asterisms omitting Hydra, Lupus, Crater, and Corvus.
   '° Arat. Phen. [3], p. 210, note e; Arat. Phen. [trns.], 87;
Hipp. Comm. arat., 25, 28.
   " Boker, 8-9, 15 et passim. See plates 1 and 2 for inter-
esting stereographic projections of the stars in the northern
and southern hemispheres as seen on the Aratean globe.
   '^ For editions and translations of the poem see, Arat.
Phen. [1], [2], [3], [4], [5], Arat. Phen. [trns.], Martin
[1974], and Byvanck.
'^ See Maass [1898] and [1881]; Erren; and Knaack.
   '" For texts see Cic. Arat. [1], [2], and [3] (cf. Cic. Nat.
deo., 11,41-44); for Germanicus see Germ. Arat. [1], [2] and
[3]; and for Rufus Festus Aviencus see Avien. Arat. For the
Aratea in general see Martin [1956]; Beede; Williams, Maass
[1892]; Leuthold; and Kaibel.
   '^ The translation was instigated by Tahir ibn al-Husayn,
governor of Khurasan, who lived from AD 775-776 to 822-
823/159-207 H; the translation was cited several times by
2\-^\Tum Astrol, 160-161;al-BTrani/nrfia, 47, 192;al-BTrunT

India [trns.], 1:97, 383. See Honigmann; and Ullmann, 227-
228, 309.
'^ Lesky (751); Diels Frag., 1,41-42.
   "Erat. Catast. [1] and [2]. See also Lesky, 785-786,
Martin [1956], 58; Maass [1883]. Cf. Neugebauer HAMA,
287-288;577-578.
   "* The Augustan poet Manilius, writing during the reigns
of Augustus and Tiberius, in his didactic poem Astronomicon
drew upon both Aratus and pseudo-Eratosthenes, as did the
later poet Hyginus in the Poetica Astronomia. The latter poem
was frequently illustrated. It presented myths associated with
constellations and the positions of stars relative to each
constellation. See Thiele, 45-56.
'^ Martin [1956], 31-32; cf. Maass [1898], 561-562.
   ^^ Hipp. Comm. Arat.; among other things Hipparchus
criticises Aratus for confusing left and right in the descrip-
tions of the constellations, which was very common in all the
early descriptions of the constellations, perhaps resulting
from the images on a globe being reversed from those as
viewed in the sky, and from some people drawing the human
constellation figures on a globe so that they face in toward
the globe and some drawing them so that they face out from
the globe.
^' Toomer, [2]; Neugebauer HAMA, 274-343.
   22 Vogt; see also Neugebauer//AM/1, 340-341, 277-292;
and Pedersen (252 ff); cf. Boll [1901].
" Neugebauer//^M/l, 285-286.
" Neugebauer///IM/l, 292-298.
   "Schlachter, 14-29; Neugebauer HAMA, 931; Vogt;
Dijksterhuis, 24.
   2® Ptolemy refers to the configurations of the asterisms
{aanpLcrfios) on Hipparchus's "solid globe" {artpta a<paipa);
Ptol. Aim., Book VII, 1; see Ptol. Aim. [1], 2:11-12; Ptol.
Aim. [2], 2:9, Ptol. Aim. [trns. Ger.] 2:11-12; and Ptol. Aim.
[trns. Eng.], 226.
   2' Strabo Geogr., 11,5.10. Cf. Geminus Elem. astr.. Book
XVI; Geminus Elem. astr. [I], 172-173. See also Schlachter,
54-57. Strabo also refers to another globe {e^atpa, not
clear whether celestial or terrestrial) belonging to a certain
Billarus, which was taken as booty from Sinope in Armenia
on the coast of the Black Sea when the city was captured by
the Roman commander Lucullus (Lucius Licinius Lucullus
ca. 117-56 BC); Strabo Geogr., XII,3.11.
^^ Forbes.
   2^ Strabo Geogr., 1,3.3; present author's translation. Cf.
Strabo Geogr., 11,5.5.
^" Cic. Repub., I,xiv; Repub. [trns.], 119-120.
   ^' This "globe" by Archimedes is also mentioned in Cic.
Nat. deo., II,xxxv and Cic. Tusc. disp., I,xxv.63; by the first
century AD Ovid in Fasti, VI,269-280; the third- to fourth-
century Lucius Caecilius Firmianus LactantiusDtv. inst., 11,5,
  
297

298

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Div. inst. [trns.], 114; the fourth- to fifth-century writer
Claudian {Shorter Poems LI, no. LXVIII in Platnauer's trans-
lation); the fifth century Martianus Capella, Mart. Cap. Nupt.
phil., 11,212 and VI,583; and the sixth-century Cassiodorus,
Cass. Var., 1,45. See Dijksterhuis, 23-25; Price [1975], 56-
57; and Schlachter, 48-54. Archimedes according to Pappus
of Alexandria (VIII,3, who cites the authority of Carpus of
Antioch) is also supposed to have written a treatise On the
making of spheres {irtpl atpaipoTrouas), which may well have
described the intricate mechanical sphere or planetarium
rather than the simpler "solid" celestial globe. See also Lorch
[1980b] 290-291 and see especially Zhitomirskii.
   '2 Cic. Repub., l,xiv; English translation is that of Sabine,
Cic. Repub. [trns.], 119-120.
^^ Cic. Nat. deo., ll,xxxiv-xxxv.
''^ See Schlachter, 46-54.
   ^^ Plato Tim., 40d; after mentioning the movements of
the sun, moon, and planets, he says that "to describe all this
without the visible models of these same would be labour
spent in vain" (translation by Cornford, from Plato Tim.
[trns.], 135; cf. note 1). Such reference may not necessarily
imply a mechanically moved device of the type generally
associated with the Archimedean model.
   ^® Galen UP, XIV,5: "For just as there are those who
imitate the revolutions of the wandering stars [the planets]
with models which bv means of certain instruments they
endow with the principle of motion and who go away them-
selves while the instruments [continue to] act as if their
creator were present and always controlling them, so in the
same way . . ." (English translation that of May from Galen
UP [trns.], 2:627).
   ^' See Price [1975], 57-58 for a detailed estimate of the
gear ratios.
   '^ Hultsch; see also Tannery [1893] and Lorch [1980b],
291. For related Chinese devices of the second to eleventh
centuries see Lorch [1980b], 291-292 and Needham, Ling,
Sc Price, 44-47.
'^ Price [1975], 58.
   ''"See Lorch [1980b] for an edition and translation of
this text.
•*' See Dicks; and Neugebauer HAMA, 578-589.
   ■*^ Geminus book XVI, Elem. astr. [1], 168-169, Elem.
astr., [2] 54-55. See also Neugebauer HAMA, 581-582; see
Schlachter 46-54 for discussion of "ringed" spheres in anti-
quity.
   ■•^ The angular distances are given not only in the usual
360° {notpa) but in sexigesimal (e^T/Koara) divisions as well;
cf. Neugebauer//ylMy4, 582.
   '*'' For a discussion of the obliquity of the ecliptic in
antiquity see Gundel Sc Boker, cols. 539-540; and Mercier.
"■'* Neugebauer HAMA, 865-866.
'" Neugebauer///1M>1, 836-837.
   "' Ptol. Aim., book VII, 5 to Vlll, 2. {Alm.[\], 2:38-169;
[2], 2:32-83; [trns. Ger.], 2:32-64; [trns. Eng.], 234-258).
The number 1022 is sometimes given (e.g., Neugebauer
HAMA, 285, 890) as Ptolemy himself did {Aim. book VIII,

1; Aim. [1], 2:169), because the three stars of Coma Bere-
nices (5 wXoKafios) while described in the catalog under the
unformed stars of Leo, were not included in the subtotal for
the zodiacal constellations nor in the total for the entire
catalog (see Chapter 5). See Kunitzsch [1974] for Arabic
versions of the Almagest and in particular the star catalog;
Peters Sc Knobel; Tallgren; Neugebauer HAMA, 280-292,
834-838; and Ibn al-Salah for twelfth-century criticisms of
the catalog. For the Almagest and Ptolemy in general, see
van der Waerden; Toomer DSB; and Pedersen.
   "^ Neugebauer//y^Af/i, 593-594, 600, cf. 32-34. See also
Kunitzsch [1974], 150-160.
   ^^ Ptol. Aim., VII,4 {Aim. [1], 2:36). The reign did not
begin in AD 137 as frequently stated (e.g., Neugebauer
HAMA, 890 and Ptol. Aim. [trns. Ger.], 1:15,76; 2:31), for
Antoninus Pius was adopted by Hadrian and made a col-
league in the tribunate on 25 February AD 138 and was
made emperor following the death of Hadrian on 2 July AD
138. For Antoninus Pius, who ruled from 10 July AD 138 to
7 March AD 161 see Aurelius.
   ■'^° Ptol. Aim. [1], 2:179-185; [2], 2:92-97 (a poor draw-
ing of the precession globe is on the title page of vol. 2);
Ptol. Aim. [trns. Ger.] 2:72-76; Ptol. Aim. [trns. Eng.], 261-
263; very inadequate summary in Stevenson 1:20. See also
Neugebauer HAMA, 890-892 and figs. 79-80 to Part II,
where the interpretation of the design is slightly different.
Cf. Ptol. Geogr., 1,22.23 and Vll,6.7 for Ptolemy's comments
on terrestrial globes.
   ^' Neugebauer HAMA, 891 and figure 80A (to Part 11)
interprets the passage to mean a central slit is made along
one-half the circumferences; however, the text speaks of
lines (TJ ypap.nr]) not slits, and slits are in no way required
for the design.
'2 Neugebauer///iA/yl, 986 and 1027.
   ^^ Ptolemy also described an observational armillary
sphere to be used to make direct observations of celestial
coordinates (Ptol. A/?n., V,l; Pto\ Aim. [1], 1:350-354, Ptol.
Aim. [trns. Ger.], 1:254-258); cf. Aim., VII, 4. Ptolemy calls
such an armillary sphere an "astrolabe" {aerpoXajSov) even
though he was acquainted with what we today term a plan-
ispheric astrolabe. Compare Ptol. Tetra., 111,2, where aarpo-
Xa^a wpoaKOTia may refer to the same instrument or to a
planispheric astrolabe (according to Hartner [1939], 2532,
note 1). This observational armillary sphere (which has no
globe at the center and has sighting devices mounted on the
rings) forms a separate class of instruments distinct from
celestial globes or demonstrational armillary spheres or spher-
ical a.strolabes. See Lorch [1975] for later versions of obser-
vational armillary spheres related to the torquetum of Jabir
ibn Afiah of Seville in the early twelfth century. Compare
also Tycho Brahe's sixteenth-century design of the armillae
zodiacales (Brahe, 52-55).
   ^'* Ptolemy, as mentioned above, specifies wood for the
rings and presumably for the sphere itself, although he does
not specifically say so. Wood {^vXlvr]) is specifically men-
tioned by two later writers, Achilles Tatius, an astronomer
  
NUMBER 46

299



of the third century AD, in his Introduction to the Phaenomena
of Aratus (Maass [1898], 62) and the seventh- to eighth-
century Leontius in his On the Construction of the Aratean
G/o6? (Maass [1898], 565).
   ^^ Perhaps in some way related to the tradition of making
celestial globes is a white marble sundial in the form of a
globe, 262 mm in diameter, now at the Archeological Mu-
seum, Nafplion, Greece. This unique globe was discovered
in 1939 by Carl Blegen in the ruins of Prosymna and
attributed to the second century BC The sphere contains
three networks of arcs in addition to the horizon and merid-
ian circles and appears to be a uniquely designed sundial.
See Gibbs, 27-30, 376-377, and figs. 62-64.
   ■■^^See Price [1975] for study of remains of a geared
calendar computer from the first century BC showing the
relative movements of the sun and moon, but not the planets.
   ^' Now in the Museo Nazionale, Naples. See Passeri;
Thiele, 40-42 and plates II-Vl; Schlachter, 42-43; Raima
[1822]; Wernicke; Panofsky Sc SaxI, 232; Gundel & Boker,
cols. 614-615 no. 5; Wellesz [1964], 84, for reproduction
of eighteenth-century drawing; Stevenson, 1: figs. 7 and 8.
The best published photographs are in plate 23 of "Astron-
omy and Astrology" in Encyc. World Art, 2.
•'* Schlachter, 42-43; Gundel Sc Boker, col. 615, no. 6.
   •^^ Berlin Museum No. 1050A; see Thiele, 42; Gundel &
Boker, col. 615 No. 7.
   ^° Schlachter, 44, plate I, figure 17; Gundel Sc Boker, col.
613, no. 2.
   '^' Sala dei busti, 341; see Thiele 43; Gundel & Boker,
col. 613, no. 1.
   *'2 See Gundel 8c Boker, col. 614 nos. 3 and 4, for two
additional ones from the end of the second century AD.
   *'■'' Schlachter, 43-44, plate I, figure 1, suggests it is prob-
ably a gnostic or mithra-cult monument of the second or
third century AD; see also Thiele 43.
   " Schlachter, 58-105; Thiele, passim; Gundel Sc Boker,
cols. 613-620, 648-649, 659-660, nos. 1-28, 126-128,
and 160-162. See also Stevenson, I: figure 10, for photo-
graph of a fresco of about AD 50 found at the villa in
Boscoreale not far from Pompeii (now at the Metropolitan
Museum of Art in New York) showing what is probably a
celestial globe (although no stars are indicated) with five
parallels (equator, tropics, and "arctic' and "antarctic' cir-
cles) but with no ecliptic shown. The depiction is unique
among Greco-Roman artifacts in that the globe bears eight
or nine meridian circles.
   "•'^ See Schlachter, 61, plate II, figure 49 for drawing (the
gem was in a private collection in 1846; present location
unknown). Schlachter suggests the astronomer is intended
to be Hipparchus, while Halma on the frontispiece of Ptol.
Aim. [2], 1 reproduced the drawing, claiming the astronomer
to be Ptolemy.
   '^'^ For text see Leon. [1], [2], and [3]. See Stevenson,
1:21-23, for very free paraphrase; see also Schlachter, 34-
36.
•^'Leon. [2], 561 lines 15-17.
  
*'^ For this commentator upon Aratus who lived probably
in the second half of the second century AD, see Gudeman.
^^Leon. [2], 562, lines 1-15.
'°Leon. [2], 564, lines 16-18.
   '' In the first century AD the Augustan poet Manilius
described in his poem (Manil. Astron., I,xiv,56I-707) a
similar arrangement where a celestial globe having the five
parallels, two colures, and two oblique circles (i.e., the zodiac
drawn as three parallel lines and the Milky Way) was set into
two moveable circles or rings consisting of the meridian ring
marking the vertex and the horizon ring. The arctic and
antarctic circles are said to be 36° (6/60) from the poles,
which corresponds to the latitude of Rhodes rather than
Rome, which has a latitude of 41°53'. Hence the globes he
rather inaccurately described were probably imported into
Rome.
'Mbn al-Salah, 18.
   '* Dresden, Sachsische Landesbibliothek MS Dc 183, folio
13r. See Thiele, 43, fig. 7, and Lehmann, 25, fig. 66, for
reproduction of illustrations. A very similar illustration is in
a parallel manuscript in St. Gall, Stiftsbibliothek, MS 250
folio 472r. Both manuscripts have been edited by Maass
[1898], 154-171, cf. xxxii-xxxvi, who does not reproduce
the illustrations. See Gundel &: Boker, cols. 684-685, nos.
242-245, for other manuscripts with illustrations of globes.
   ''' For classical Chinese construction of celestial globes,
the earliest being AD 435, see Needham 8c Ling, 382-390;
and Needham, Ling, Sc Price, 94-99.
   "Creswell, 1(2):400. A later date of about AD 723 or
742 has also been suggested; see Maddison &: Turner, 135,
note 12.
   '® See Creswell, 1(1): plates 1 ib,c,d, and 16a,b. For a good
color photograph made after the recent cleaning and resto-
ration by a Spanish conservation team, see Baker Sc Stam-
ford, 23.
   " See SaxI for analysis from standpoint of art history and
Beer for the interpretation of an astronomer. For citations
in both classical Islamic literature and Greek writings of
other domes and ceilings that were decorated in a similar
tradition, see SaxI, 424 note 5, and Lehmann.
   ''^ Rome, Biblioteca Vaticana, MS Graec. 1087 (fifteenth
century) and Vatican MS Graec. 1291 (ninth century). See
SaxI, 426-428. In one of the manuscripts (Vatican MS.
Graec. 1087) there is a drawing of a celestial globe (no ring
or stand) showing the hemisphere about the winter solstice
with figures in reverse order as on a globe. The sphere is
divided horizontally by five lines representing the equator
and tropic and two additional lines half way between the
tropic and the poles, with a center vertical line indicating
the solstitial colure and a very wide band for the ecliptic.
   ^^ Such as the globe No. 5 of the present catalog. See
SaxI, 428-430.
   **° See Chapter 2 for the purpose and designation of these
circles.
   *" Boulogne-sur-Mer, Bibliotheque Municipale MS 188,
folio 20r, and London, British Library, MS Harley 647, folio
  
300

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



21v; both are Latin manuscripts, the first of the tenth to
eleventh century and the second possibly ninth century.
Photographs of the folios are in Encyc. World Art, 2: plate
21. See also Zick-Nissen, plates 52 and 53.
   ^2 Maddison [1966], 8 note 9 and Maddison Sc Turner,
66. Beer goes to some lengths to try to explain the positions
of the constellation figures and the concentric circles on the
dome, but unfortunately seems unaware of the planispheric
star charts of the type mentioned in note 81 above and
consequently does not consider them as a possible source of
design. Taking the suggestion of Maddison one step further
and conjecturing that the artist's model, which in fact he did
not understand very well, was a stereographic projection of
a celestial globe, then more details seem accounted for than
by any other explanation.
   ^^ See Gunther >IVV, 82-103, for an English translation of
his treatise on the astrolabe.
   **'* For the Paricasiddhdntikd XIV,23-26, see Varahami-
hira [1889], 80-81 and Varahamihira [1970], 1:128-129,
2:89.
"^ Kunitzsch [1974], 15-34, cf. 34-82.
   "*' Al-Sufi Suwar, 5; al-Harrani has been added to '^AIT ibn
■^Isa's name by the editors. Ibn al-Nadlm in the Fihrist, book
VII,2 (Ibn al Nadlm 1:284 and Ibn al-Nadlm [trns.], 671)
says "^AIT ibn "^Isa was an apprentice to Ibn Khalid al-Mar-
warrudhT the astrolabe maker to the caliph al-Ma^mun in
Baghdad; Ibn al-Nadim then goes on to list several instru-
ment makers who were in turn apprentices to "^AIT ibn "^Isa.
He also wrote a treatise on the use of the astrolabe (see "^AIT
ibn Tsa Asturldb).
   *^ Ibn al-Nadlm in the Fihrist, book Vll,2 (Ibn al-Nadlm,
1:284 and Ibn al-Nadlm [tins.], 670-671). See Price [1975]
for suggestion that Rhodes may have been the medium
through which complicated technical instrument making
passed from the Greek world to Harran.
   **** Kurdt kathlrah min '^amal al-harrdniyin (al-Siifi Suwar,
143).
   '*'' Ibn al-Nadlm in the Fihrist book VII,2 (Ibn al-Nadlm,
1:285; Ibn al-Nadlm [trns.], 672, note 167).
«° Ibn al-Salah, 20.
"^ Hartner DSB.
   ^^ Al-Battani Op. astr., 3:210-214 is Arabic text; 1:139-
142 is Latin translation, with notes on 1:319-321.
^^ Al-BattanT Op. astr., 3:210 lines 4-5; 1:139.
   ®'' This gnomon is not a sighting device, as has sometime >
been asserted (e.g., Lorch [1980a] 158 and [1980b] 295),
nor does it have two holes through which the light is aligned
as is shown in the drawing appended by the editor and
translator of the treatise (al-Battanl Op. astr., 1:320). The
globe with gnomon as described by al-BattanT could perhaps
be considered as an indirect observational instrument in that
the latitude is determined by observing the shadow cast on
the sphere by the moveable gnomon, but it is, of course, not
a direct observational instrument.
^^ For life and writings, see Harvey; Wiedemann [1]; Ibn

al-NadTm, VII, 3 (Ibn al-NadTm, 1:295; Ibn al-NadTm[trns.],
695).
   ^•^ Extant in two recensions; see Worrell for partial trans-
lation and summary. It is also sometimes referred to as Kitdb
fl ^amal bi-l-kurah dhdt al-kursl (Book on the Use of the
Globe Having a Stand). Princeton University, Garrett Col-
lection (Yahuda Section) MS 4901 has the title Risdlahfial-
^amal bi-l-kurah al-falaklyah. See Schnell for a German trans-
lation.
^'Steinschneider £uro/>., 1:77.
^^ See Harper.
^^ Narducci, 24.
   '°° Alfonso el Sabio, 1:163-208; for a general discussion
of the Libros del saber de astronomia, see Procter.
   '*» Kitdb al-"amal bi-l-asturldb al-kuri; see Harvey; See-
mann [1925], 46-49.
   '°2 Stern £/; Brockelmann G^L, 1:223, and 5M/)/?/., 1:398;
Hauber; Kunitzsch DSB[2].
'"^Storey, 41.
   "*'• Oxford, Bodleian MS Marsh 144; Paris, Bibliotheque
Nationale, MS arabe 5036, and Istanbul, Ayasofia MS 2595,
respectively. See Upton; Wellesz [ 1959], [ 1965]; and Winter.
An edition has been printed of the Arabic text based on the
Oxford and Paris manuscripts in al-Sufi Suwar. There are
two published versions of NasTr al-DTn al-TiasT's Persian
translation (al-Sufi ed. al-TQsT [1] and al-Sufi ed. al-TusT
[2]). For a French translation of the Arabic see al-Sufi 5ua;ar
[trns.]. For a discussion of the Latin versions see Kunitzsch
[1965].
   io.'5 PQJ, C u^arid ibn Muhammad al-Hasib, a ninth-century
mathematician and astronomer, see Ibn al-NadTm 1:278; Ibn
al-Nadlm [trns.], 658; and Ullmann, 316. Among other
writings he composed treatises on the use of the astrolabe
and the use of the armillary sphere {dhdt al-halaq).
'°^ Al-Sufi 5ui/;ar, 2-5; al-Sufi 5ua;ar [trns.], 29-31.
'"' Al-SQfi Suwar, 2; al-Sufi Suwar [trns.], 30 (cf. 37).
'"^ Al-SQfi Suwar, 1-2; al-Sufi Suwar [trns.], 29-30.
   '"^ Fr tashtlh al-suwar wa tabtlh al-kuwar (Suter [1922],
86).
"° Suter [ 1922], 81 -82; present author s translation.
   '" Kitdb al-"amal bi-l-kurah al-falaklyah, extant in Istan-
bul, Topkapi Sarayi MS 3505, item 1; Brockelmann GAL,
Suppl. 1:398. He quotes five authorities; Autolycus, Heron,
Philon, Theon of Alexandria, and Qusta ibn Luqa (Lorch
[1980b], 297). This treatise is possibly quite important and
should be studied and published.
   "2 Ibn al-QiftT, 440. The maker of astrolabes who saw
the globe, Ibn al-SinbadT, is described by Ibn al-QiftT as
follows: "This man was in Egypt and one of those people of
knowledge, learning and experience in the production of
astrolabes and celestial globes." According to al-QiftI, Ibn
al-SinbadT also saw in the library in Cairo a celestial globe of
brass {nuhds) supposedly by Ptolemy and on it was written
"this globe was taken from the AmTr Khalid ibn YazTd ibn
Mu'^awiyah." The latter was an Umayyad prince (died AD
  
NUMBER 46

301



704) with whom legend has associated an interest in alchemy
and other scientific and quasi-.scientific subjects.
   "^ Al-BTrQnT Qanun, 3:987-1126; alKtrum Astrol., 157-
164; al-BTrQnT Chron. [trns.], 335-365; see also al-BTrunT
Qanun 3:1139-1157 for lunar mansions. See Boilot; Ken-
nedy DSB; and al-BTrQnT Catalog.
   "'* London, British Library MS Or. Add. 7697, Kildb al-
tafhim li-awd^il sind"at al-tanjlm, copied in 685 H/AD 1286
by Ibn al-Ghulam al-QuniyavT (Ilkhanid style).
   "^The stars for Constellation 48 are there, but the
outline is not drawn and the name not inscribed.
   "<^ Sa-Md al-AndalQsT Tabaqdt, 934-935; Sa^id al-AndalQsT
[trns.], 138-139.
   '"The astiolabes are at Madrid, Museo Arqueologico
Nacional (dated Sha'T^Qn 459 n/June-July AD 1067, Gibbs
CCA, no. 117, Mayer [1956] 50-51; Oxford, Museum of
the History of Science (Shawwal 460 H/August AD 1068),
Gibbs CCA, no. 118, Mayer [ 1956], 51; Rome, Observatorio
Astronomica di Roma (one dated Rajab 463 n/April AD
1071 and a questionable one dated 493 H/AD 1099), Gibbs
CCA, nos. 1167 and 123, Mayer [1956], 51-52; Kassel,
Staatliche Kunstsammlungen (dated 478 H/AD 1086), Gibbs
CCA, no. 121, Mayer [1956], 52. For further details see
Brieux Sc Maddison.
"« Suter [1900], 117 no. 278.
   "^Chicago, Adler Planetarium (dated 525 H/AD 1130),
Gibbs CCA, no. 2557. Another astrolabe bears only the
signature of Hibat Allah ibn al-Husayn al-BaghdadT, al-
though Badr may have assisted since there is much similarity
in the signatures; present location of astrolabe (dated 513
H/AD 1119) is unknown, being at one time in a private
collection in Munich; Gibbs CCA, no. 3633, where date is
given as 518 H/AD 1124-1125. The author wishes to thank
Francis Maddison of the Museum of the History of Science,
Oxford, for bringing this material to her attention.
'2° Lorch [1980b].
'2' Gottschalk EI; and Gottschalk [1958].
   '22 On Qaysar see Sabra [1969], 8; and Destombes
[1966], 45-46.
'22 Mayer [1956], 80.
'2^ Barrett, xii-xiii; see also Pinder-Wilson BM.
   '2^ He also wrote a treatise on some of the instruments
used at the Maragha observatory, which does not, however,
mention celestial globes (see Seemann [1928]).
'2«See Minorsky; and Sayili, 187-207.
'2' Hartner [1950], esp. reprint, 220-221.
   '2" This globe will be fully analyzed in a forthcoming
study by Marcel Destombes. See Destombes [1957], 320 no.
7; and Destombes [ 1958], 307 note 2. For a guide to Ilkhanid
metalwork see EI [2], 3:1126-1127.
'2» Sayili, 187-188.
   '^"At Paris, Destombes Collection (dated 755 H/AD
1354), Gibbs CCA no. 1205; Massachusetts, S.V. Hoffmann
Collection (dated 774 H/AD 1372-1373), Gibbs CCA, no.
15, Mayer [1956] 54; Calcutta, Indian Museum (dated 790

H/AD 1388) Gibbs CCA, no. 16; Gunther y4M', 130, fig. 16;
present location unknown (appeared in a Sotheby sale cata-
log) dated 790 H/AD 1 388, Gibbs CCA, no. 2605. A forgery
under his name is at the (Greenwich National Maritime
Museum (dated 757 H/AD 1356); see Mayer [1956], 53;
Brieux Sc Maddison. Both of the globes byJaTar have holes
drilled along the solstitial colure on the opposite side of the
ecliptic poles from the equatorial poles and the same distance
away. The purpose of these holes is unknown.
   '^' In Rabat, Musee des Oudaia (dated 796 H/AD 1393-
1394), Gibbs CCA, no. 2710; and Copenhagen, Kunstindus-
trimuseet (estimated date, AD 1426) Gibbs CCA, no. 3595.
See Brieux 8c Maddison for further details.
   '■■'2 Storey, 62-72; Barthold; and Sayili, 272-282. See
Zenkert; Hookham; and Shcheglov for photographs of the
remains of the observatory.
   "^ Ulugh Beg Zij; Ulugh Beg Zij [trns. Fren.]; and Ulugh
Beg Zij [trns. Eng.]. See also Kary-Niazov.
   '^■^ Al-KashT Letter, 96; and al-KashT [trns.], 196. See also
al-KashT £(7ua^, 1-7; and Kennedy [1961].
   '"''^ A Hashimid is a family member of a line of Meccan
sharTfs who ruled Mecca almost without interruption from
the tenth century to 1924. See Rentz.
'^^See Guimbretiere Sc Hasan.
   "^ In 1935 the dated astrolabe was in the library of
Nawab Sir Salar Jung Bahadur in Hyderabad (see Nadvi
[1935], 627; and Gibbs CCA, no. 1120). The two undated
ones are in the Billmeir Collection, Museum of the History
of Science, Oxford; Gibbs CCA nos. 1089 and 2530. See
Brieux Sc Maddison for further details.
   "* European orientalists frequently write the name as
llahdad. See Storey, 157 note 3. The spelling with a single
lam is a common nasta"llq spelling and is used in all but two
signatures of the Lahore workshop. On two globes, however,
(Nos. 30 and 71) the name is written as two words with a
double lam, that is, as Allah-dad. See also Nadvi [1935] and
Nadvi [1937] and Abbott for other possible interpretations
of this name as well as further information on some of the
products of this family. See also Wiet [1936] and especially
Brieux 8c Maddison for a complete account of the family.
   '^'^ Abbott argues that HumayQnT means "royal" and the
third and fourth generations of the family were official
instrument makers to the successors of HumayQn. Nadvi
[1937], 537-538 rejects this interpretation, saying that it
clearly applies to Allahdad alone, and in any case never
means "royal" in Persian but only "auspicious." Note that
one grandson always signed his works as Qa^im Muhammad
ibn " Isa ibn Allahdad AsturlabT LahurT HumayunT, while
the other grandson and the great-grandsons signed their
names as ... . ibn " Isa ibn Shaykh Allahdad AsturlabT Hu-
mayQnT LahQrT. This change in order does not affect the
interpretation of the nisbah HumayQnT, however. On two
globes (Nos. 30 and 71) the great-grandson Diya^ al-DTn
Muhammad omits the nisbah LahurT from his signature,
perhaps indicating that by  that time the workshop had
  
302

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



moved to Delhi, the seat at the Mughal court at the time the
globes were executed.
   '"•"For AbQ al-Fadl's Akbarndma 1.241, the text is given
by Nadvi [1935], 622-623. See Nadvi [1935], 621-626 for
other anecdotes regarding Humayun's scientific interests,
including an account that when HumayQn reached Tabriz
in Persia he ordered his servant to search the city for a
celestial globe {kurah). The servant interpreted the word for
globe as the Persian word for a colt, and so brought before
HumayQn a selection of young horses. It is also reported by
AbQ al-Fadl in the Akbarndma that HumayQn died after
falling down the stairs of his library in Delhi which was on
the third floor of a building and had been converted into an
observatory. HumayQn was apparently hurrying down for
the evening prayer which had interrupted a discussion with
a group of mathematicians in the observatory while awaiting
the appearance of the planet Venus.
'""JahangTr Memoirs [2], 82.
   '■*2 0ne astrolabe is dated 1009 H/AD 1600 and two are
dated 1013 H/AD 1604, the former published in a Sotheby
sale catalog (present location unknown), Gibbs CCA, no.
3828, and the latter two at the Adler Planetarium in Chicago
and the Palais de I'Orient, Paris, Gibbs CCA nos. 1096 and
1604. Two undated astrolabes are in private collections in
London and Chiswick, Gibbs CCA, nos. 68 and 3825.
   '■'Mn Hannover, in the Kestner-Museum, and the Mu-
seum of Arab Antiquities, Baghdad, Gibbs CCA, nos. 1609
and 3820, respectively.
   '■''' See Brieux Sc Maddison for a complete listing. A partial
list includes the following: astrolabes at Delhi, Red Fort
Archaeological Museum (dated 1034 H/AD 1624 and 1047
H/AD 1637), Gibbs CCA, nos. 3723 and 3721; Delhi, Na-
tional Museum (1034 H/AD 1624), Gibbs CCA, no. 2700;
Baghdad, Museum of Arab Antiquities (1040 H/AD 1630),
Gibbs CCA, no. 3826; Oxford, Museum of the History of
Science (1044 H/AD 1634, 1053 H/AD 1643, 1051 H/AD
1641, and one undated one), Gibbs CCA, nos. 71, 72, 2531,
and 1013; present location unknown (listed in sale catalog
of Sotheby's, dated 1047 H/AD 1637, one undated, and one
of estimated date AD 1640), Gibbs CCA, nos. 2601, 3828,
and 2609; Karachi, National Museum (1047 H/AD 1637),
Gibbs CCA, no. 2704; Calcutta, Indian Museum (1048 H/
AD 1638), Gibbs CCA, no. 3730; Greenwich, National Mar-
itime Museum (1051 H/AD 1641) Gibbs CCA, no. 1054;
Smithsonian Institution, NMAH (1053 H/AD 1643), Gibbs
CCA, no. 86; London, British Museum (1070 H/AD 1654),
Gibbs CCA, no. 78. The other undated ones are at Leiden,
Rijtsmuseum voor de Geschiedienis det Natuurwetenschap-
pen, Gibbs CCA, no. 1097; the private collection of Dr.
Sottas, Gibbs CCA, no. 1119; Paris, Landau Collection,
Gibbs CCA, no. 3529; Paris, Destombes collection, Gibbs
CCA, no. 3537; New Haven, Connecticut, Yale University,
Gibbs CCA, no. 3807; Don Mills, Ontario, Ontario Science
Museum, Gibbs CCA, no. 3827; and Salem Mass., Peabody
Museum, Gibbs CCA, no.  3655. Two astrolabes though

unsigned have with confidence been attributed to him (now
at Oxford, Museum of the History of Science, Gibbs CCA,
no. 1014, and at Hannover, Kestner-Museum, Gibbs CCA,
no 1098). In the Charliat Collection in Paris there is a
forgery dated 1053 H/AD 1643, Gibbs CCA, no. 3504, and
in Haifa at the National Maritime Museum there is an
undated one generally conceded to be a forgery, Gibbs CCA,
no. 3565.
   '''^One of the astrolabes is of estimated date AD 1631 and
the other dated 1034 H/AD 1624-1625, now at Baghdad,
the Museum of Arab Antiquities (Gibbs CCA, no. 3821) and
in a private collection in Calcutta as of 1935 (Gibbs CCA,
no. 1128), respectively. See Brieux Sc Maddison.
'"^JahangTr Mmoiri [1], 260, 319; Memoirs [2], 81.
'^'Ibid. [1], 218, 319-320; Ibid. [2], 2, 1, 81, 215, 260.
   ■'**' We do know from written sources of one astrolabe and
celestial globe maker who enjoyed the royal favor of Hu-
mayQn (see above, p. 44), while JahangTr mentions in his
Memoirs [2], 98, two mastercraftsmen {ustadh) named PQran
and Kalyan "who had no rivals in the art of engraving" and
who were requested to make a dagger hilt, and also a
European goldsmith {Memoirs [2], 80).
   "•^An astrolabe dated 1103 H/AD 1691-1692 is in Istan-
bul, Turk ve Islam Muzesi, while an astrolabe signed by
Jamal al-DTn and bequeathed in 1094 H/AD 1682-1683 to a
mosque as a perpetual waqf by '^Abd al-GhafQr ibn Zayn al-
DTn al-Munajjim al-LarT is in Paris in the private collection
of A. Brieux (Gibbs CCA, no. 3519). A third astrolabe dated
1092 H/AD 1681 is attributed to him and is at the Indian
Museum of the Royal Asiatic Society (Gibbs CCA, no. 81).
See Brieux Sc Maddison.
   '■^"The astrolabes are in the Leonard Linton Collection
in Point Lookout, New York (dated 1102 H/AD 1690-1691);
Harvard University, Collection of Historical Instruments
(1038 H/AD 1628-1629), Gibbs CCA, no. 3624; Cambridge,
Whipple Museum of the History of Science (estimated date
of AD 1628) CCA no. 3563; Paris, Destombes Collection,
CCA no. 3538 (1071 H/AD 1660); London, Clockmakers
Company Museum (1099 H/AD 1687) CCA no. 1115; and
present location unknown (appeared in a sale catalog of
Sotheby 8c Co.) dated 1086 H/AD 1675, CCA no. 3822; and
another in the Destombes Collection in Paris. For details see
Brieux Sc Maddison.
   ''"'Collection of Alain Brieux, Paris. See Brieux Sc Mad-
dison.
   '"^2 There are three astrolabes at Chicago, Adler Planetar-
ium (dated 1047 H/AD 1637, 1057 H/AD 1747, and 1071 H/
AD 1660), Gibbs CCA, nos. 2558, 1095, and 2554; present
location unknown (appeared in a sale catalog of Sotheby's)
dated 1056 H/AD 1646, 1071 H/AD 1660, and one undated,
Gibbs CCA, nos. 2600, 2607, and 3651; Lucknow, library
of Nadwat al-'^ulama (1059 H/AD 1649) Gibbs CCA, no.
1126; Cardiff, National Museum of Wales (1062 H/AD 1651)
Gibbs CCA, no. 1107; Oxford Museum of the History of
Science, two dated 1069 H/AD 1658, and one dated 1064
  
NUMBER 46

303



H/AD 1653, Gibbs CCA, nos. 1002, 77, and 2533; Aligarth,
library of Nawab Sadar Yar Jung Maulana Habiba al-Rah-
man Khan Sherwani (1064 H/AD 1654 and 1074 H/AD 1680)
Gibbs CCA, nos. 1118, 1116; Cairo, Harari Collection (?Mu-
seum of Islamic Art) dated 1068 H/AD 1657, Gibbs CCA,
no. 3829; Edgertown, Mass., Hoffman Collection (1070 H/
AD 1659) Gibbs CCA, no. 87; Cambridge, Mass., Owen
Gingerich Collection (1073 H/AD 1662) Gibbs CCA, no.
3809; London, Victoria and Albert Museum (1074 H/AD
1663) Gibbs CCA, no. 1060; Rampur State Library (1074
H/AD 1663) Gibbs CCA, no. 2511; and Patna (Bihar) India,
Khuda Bakhsh Oriental Public Library (1074 H/AD 1663)
Gibbs CCA, no. 1117; Jaipur Observatory (1067 H/AD 1656,
1085 H/AD 1674, and 1091 H/AD 1680), Gibbs CCA, no.
2702, 2703, and 80; Paris, Landau Collection, two undated,
Gibbs CCA, nos. 3524 and 3525; and Paris, Alain Brieux
Collection, undated, Gibbs CCA, no. 3517; and Brooklyn,
New York, Brooklyn Museum, one undated, Gibbs CCA,
no. 3555. See Brieux Sc Maddison for further details.
   '""^See Kurz, plate XI, for an eighteenth-century engrav-
ing of an Islamicate demonstrational armillary sphere with
a small terrestrial globe at the center, taken from an edition
of the Jihdnnumd printed in Istanbul in 1732.
   ''^ Kaye [1918], 27-30, 118, and figures 19-22; Gunther
AW, 212-213. A zarqaliyah astrolabe, frequently called by
its Latin name saphaea Azarchelis, was so named after its
inventor Ibn al-Zarqellu, who attempted to redesign the
conventional planispheric astrolabe to compensate for the
inconvenience of needing a separate plate for each latitude.
Al-Zarqellu himself called the astrolabe al-'^AbbadTyah in
honor of al-Mu'^tamid ibn "^Abbad, King of Seville from AD
1068-1091. In the Libros del saber de astronomia of Alfonso
el Sabio there is a Spanish translation entitled Libro de la
agefecha of Ibn al-Zarqellu's treatise on this universal astro-
labe. See Hartner [ 1 ], 316-317; Michel [ 1947], 93-102; and
Gunther AW, 256-262.
   '^■"^ Nadvi [1935], 626 quotes text from the "^Ain al-Akbarl
by AbQ al-Fadl.
   "^■^ Istanbul, Topkapi Sarayi, dated 1006 H/AD 1597
(Gibbs CCA, no. 3707) and Washington, D.C, Smithsonian
NMAH, dated 1020 H/AD 1611 (Gibbs CCA, no. 70).
   '^^ Oxford, Museum of the History of Science, dated 1057
H/AD 1647 (Gibbs CCA, no. 73).
   ''^*'East Indian Company Museum, dated 1076 H/AD 1665
(Gibbs CCA, no. 23); see Morley, 37-39. Another astrolabe
giving Muhammad Salih as maker is dated 1077 H/AD 1666-
1667 (Oxford, Museum of the History of Science, Gibbs
CCA, no. 2502). A third astrolabe is in the private collection
of Edgerton. See Brieux 8c Maddison for further details.
  '■^"See, for example, Morley, 37-39; and Gunther AW,
139.
  "'''See the map accompanying the article "Hind" in EI
[2], 4:404-454.
   "'' The three undated astrolabes are in Paris, a private
collection (formerly Leonard Linton Collection, Point Look-

out New York, Gibbs CCA, no. 3677); Teheran, the collec-
tion of Dr. Jamileh Yeganeh (Gibbs CCA, no. 3700); and
Frankfurt, Johann Wolfgang Goethe Universitat (Gibbs
CCA, no. 3923); while the one dated 1031 H/AD 1621-1622
is in a private collection in London. See Brieux & Maddison.
   ""2 See Gascoigne, 140, for a color photograph of Jahan-
gTr's coins.
'•^^JahangTr Memoirs, [2], 6-7.
   "^"•One astrolabe dated 1071 H/AD 1659 is now at the
Mulla Firuz Library in Bombay (Gibbs CCA, no. 1153); one
dated 1085 H/AD 1674 has appeared in a sale catalog of
Sotheby Sc Co. (Gibbs CCA, no. 2608); one dated 1088 H/
AD 1677 made for Shah Subhan QulT Bahadur Khan is now
at Tashkent, the State Museum for the History of Uzbekistan
(Gibbs CCA, no. 1154). Undated astrolabes are in Berlin,
Staatliche Museum, Gibbs CCA, no. 1155; Paris, Landau
Collection, (Gibbs CCA, no. 3530; Washington, D.C, Freer
Gallery of Art, Gibbs CCA, no. 3545; and New York, the
Metropolitan Museum of Art, Gibbs CCA, no. 3550. See
Mayer [1956], 79. There is one astrolabe dated 1125 H/AD
1713 (Gibbs CCA, no. 2603) that appeared in a sale catalog
of Sotheby Sc Co., but is probably an incorrect attribution.
See Brieux 8c Maddison for fuller listings.
  "'"'At Oxford, Museum of the History of Science, dated
1057 H/AD 1647 (Gibbs CCA, no. 18).
   "^"^At London (Greenwich), National Maritime Museum,
dated 1070 H/AD 1659 (Gibbs CCA, no. 3813).
   "'^Chardin, 1:89. The section on astrolabe construction
was not included in the first three editions of his journal.
Text reproduced by Michel [1941]. This material was
brought to my attention by Silvio A. Bedini of the Smithso-
nian Institution.
"'"See Gingerich et al.
   "'^ An astrolabe was made for the Ottoman Sultan BayazTd
II (AD 1481-1512) by Shukr Allah Mukhlis ShirwanT in 891
H/AD 1486, now in the Harari Collection, Cairo (PMuseum
of Islamic Art). Three astrolabes signed by Mustafa AyyQbT
are known, in London, the Science Museum (dated 1114 H/
AD 1701-1702; Gibbs CCA, no. 1059); in Paris, the Jacques
Shumann Collection (dated 1110 H/AD 1698; Gibbs CCA,
no. 1218); and an undated one in Paris, the Destombes
Collection (Gibbs CCA, no. 3541). These three by Mustafa
AyyQbT the CCA does not consider to be Ottoman, but from
the Arabian penninsula; see Brieux Sc Maddison. There are
unsigned and undated probably Ottoman astrolabes in Ox-
ford, Museum of the History of Science (Gibbs CCA, no.
2537); the collection of Dr. Sottas (eighteenth century, Gibbs
CCA, no. 1111); and in a present unknown location (listed
in a sale catalog of Sotheby Sc Co.) Gibbs CCA, no. 2606.
These few Ottoman astrolabes are considerably simpler in
design and decoration than those produced in the same
period from other regions; see Brieux 8c Maddison.
""Sayili, 289-302.
"'Al-JabartT, 1:74, 114. and 347; Dorn [1865], 33-35.
'"Sayili, 244; al-JabartT, 1:99; Dorn [1865], 34-36.

304

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



'"Al-JabartT, 1:99; Dorn [1865], 33.
'^"Storey, 93-94. See also Kaye [1918]; and Hunter.
   '^^ Building larger instruments to ensure greater accuracy
of observation had a long history in the Islamic East; see
Hartner [1977].
   '^''The two astrolabes have an estimated date of AD 1724;
both are now at the Jaipur Observatory (Gibbs CCA, nos.
82 and 83).
'"Kaye [1918], 8-9.
   '^* In recent years they have been restored and preserved
by the Indian government; see Kaye [19181. For photo-
graphs of the JaipQr and Delhi observatories see Nasr
[1976], plates 66-69.
"^Griffin, 491-509; and Latif, 319-320.
   '^'^ Lalah BalhQmal made a large astrolabe for the same
patron after the patron had been made a Rajah by the
British government, for in a lengthy Persian inscription on
the instrument in which he hopes for a long reign for Queen
Victoria he states that he made it by order of the Rajah
Sahib Nihal Singh Bahadur AhluwalTyah [spelled correctly]
at KapQrthalah in AD 1851 and 1907 of the Vikrama era.
The present location of this astrolabe is unknown; it was
formerly in the collection of Hew Kennedy, Shropshire. A
second astrolabe bears a Persian inscription around the edge
to the effect that it was designed for the honorable Sir Henry
Elliott, K.C.B., Chief Secretary to his Lordship the Gover-
nor-General in Kapurthalah, Ahluwaliyah, by BalhQmal the
astronomer of Lahore in that gentleman's employ, AD 1849;
see Gunther y4W, 1:171-173, no. 60. This second astrolabe
was acquired in 1982 by the Science Museum, London. Two
unsigned astrolabes, both with estimated dates of AD 1850,
have also been attributed to Lalah BalhQmal; one is now in
the Egestorff Collection, Dublin, and the other in the pos-
session of Alain Brieux, Paris. Information on these astro-
labes was generously furnished by Francis Maddison of the
Museum of the History of Science, Oxford. See Brieux 8c
Maddison under the entry "BalhQmal LahurT" for yet a fifth
astrolabe. For a second celestial globe signed by BalhQmal
see globe no. 127.
   '*" Eight Sanskrit astrolabes are known to exist, one by
"^Abd al-Ghadur Moheb made about AD 1670 (London,
Collection of Messrs. Malcolm Gardner, Gibbs CCA, no.
1123); two anonymous ones of estimated date about 1670
(see Kaye [1921], Gibbs CCA, no. 94, and no. 93 at the
Jaipur Observatory); one of estimated date AD 1866 (Rock-
ford, Illinois, The Time Museum, Gibbs CCA, no. 2712);
and four undated ones (Chicago, Adler Planetarium, Gibbs
CCA, no. 2575; the Riva Collection in South America, Gibbs
CCA, no. 3921; and two in the A. Michaelis Collection in
London, Gibbs CCA, nos. 3913 and 3914).
   '"2 Howorth, 823-828. The Indian Section of the Victoria
and Albert Museum suggests Oudh in northeast India as the
place of origin. However, the nawdb of Oudh at that time
was GhazT al-Din Haydar, and ruled from 1814 to 1827;
there was never one named Muhammad "^AIT Khan (see
Majumdar,  1016). Furthermore, it being made in hemi-

spheres does not suggest an Indian origin.
   '"2 There is one astrolabe (a wooden one) known by a
maker named Muhammad KarTm, but it is dated 1133 H/AD
1720, so that it is unlikely that it is the same maker; it is now
at the Hermitage Museum in Leningrad (Gibbs CCA, no.
1149).
   '*■* However, Maddison Sc Turner, 118, note that Alex-
andrian and Yazdijird eras appear on Lahore astrolabes.
Such is not the case with the globes signed by the Lahore
family nor on any seamless (Phence Indo-Persian) globe
produced after the sixteenth century, only with the excep-
tion of the one by Muhammad KarTm.
   '^^Systematically mapped in AD 1750-1752 by Abbe Ni-
colas Louis de Lacaille (see Warner).
Chapter 2
   ' The precise number of stars on these globes varies
because of the difference in the star catalogs used, the
occasional inclusion of extra stars such as the "overlooked
star" in Ursa Major, and inadvertent omissions on the part
of the maker. See section on Classical Greek and Pre-Islamic
Sources in Chapter 5.
   2 A great circle is a circle whose diameter is also the
diameter of the globe. Thus a tropic circle or polar circle
would not be a great circle, but is referred to as a lesser
circle.
   ^ It is in fact possible to design a celestial globe on which
the relative positions of the stars are not reversed from that
viewed from earth, as witnessed by the paper model of a
celestial sphere (which is actually a truncated icosohedron
rather than a sphere) marketed in London under the name
Stardome. When using this model, which was selected for its
artistic design by the Design Center of London, the user
must remember that it is not representing a sphere of stars
surrounding the earth; he must imagine the convex surface
of stars at any given point to be bent back around him in a
concave manner. There is no evidence, however, that any
such design was conceived of or attempted in either the
Greco-Roman or Islamic contexts.
   '' There was also a tradition evident in the late ninth and
early tenth centuries of calling a celestial globe "the egg-
shaped" {al-baydah) instrument, for al-BattanT employed this
term (see Chapter 1). Furthermore, al-KhwarizniT says in his
encyclopedia written in AD 976 that "the celestial globe
{kurah) is known as one of the instruments of astronomers
and with it you learn the nature of the heavens and the
constellations of the stars, and it is also called al-baydah (the
egg)" (al-Khwarizml, 235).
   Dunas ibn TamTm in his introduction to his tenth-century
treatise on the armillary sphere refers to the celestial globe
as the "egg-shaped" {al-baydah) instrument (Stern [1956],
375). This was certainly not, however, the most common
term for it. Al-BTrQnT Qanun, 1:54-55, says the term falak
(sphere) sometimes "refers to the globe {kurah), in particular
when it is moveable {mutaharrik); falak thus does not apply
  
NUMBER 46

305



to the motionless [globe], and it is called falak only on
account of its similarity with the whorl of the rotating
spindle." See Hartner [2], 762.
   ^ See Bion, 176; Holland, 6; and Smart, 40. There is one
extant globe that does not have these circles. It is No. 72,
made in 1215 H/AD 1800-1801 by Muhammad Ashraf
TuqadT Zadah. 1 have preferred the term ecliptic latitude
circles for these circles because of the Arabic name for them,
rather than the term eclipfical circles of longitude employed
by Hartner [3], 503.
   ^ That is, meridians other than the colures. There are 10
known Islamicate globes which have meridians in addition
to the colures. All appear to be late eighteenth- or nine-
teenth-century eastern products. One (No. 72) was made by
Muhammad Ashraf TuqadT Zadah in 1215 H/AD 1800-
1801; four are all possibly products of the workshop of Lalah
Balhumal LahQrT (Nos. 33, 53, 54, and 90, illustrated in
Figures 24, 25, and 26); one is an Ottomon Turkish product
made in 1299 H/AD 1882 (No. 75); and two are undated
(Nos. 76 and 93, the latter illustrated in Figure 27). Globe
No. 63 made in Safavid Persia in 1012 H/AD 1603-1604 for
Shah "^Abbas I and the similar globe No. 83 both have some
meridians, but not complete sets.
   ' Celestial latitude and celestial longitude ought not to be
confused with latitude and longitude, which are equator-
oriented coordinates used for measuring terrestrial dis-
tances.
   "On globe No. 30, illustrated in Figure 17, which is
actually a cutout sphere by Diya^ al-DTn Muhammad of the
Lahore workshop, every ecliptic latitude circle is labeled. For
example, one has the inscription In dd^irat "^ardl janubi az
awwal burj al-dalw (this is the southern latitude circle at the
first of the House of Aquarius).
   ^ There are five exceptions that do not have the zodiacal
houses labeled: Nos. 32, 46, 47, 48, and 56.
   '" There is actually one globe. No. 75, an Ottoman Turk-
ish product made in 1299 H/AD 1882, which has the names
of the zodiacal houses written with the top of the writing
toward the north pole; in this case the names are not written
in an elongated fashion so as to fill out the full 30° of each
house.
   " See the section on Lunar Mansions in Chapter 5 for a
discussion of lunar mansions and their significance.
   '2 The seventeenth-century maker Muhammad Salih Ta-
tawT prefered the term qutb falak al-buruj (pole of the sphere
of the houses) for the ecliptic pole. This term was also
occasionally used by Diya^ al-DTn Muhammad of the contem-
porary Lahore workshop (e.g., on No. 30).
   '2 While Ptolemy gave a value for the obliquity of the
ecliptic of 23°5r20", various values were taken in the
Islamic world varying from 23°51'1" to 23°30'17". When
describing his celestial globe al-BattanT specified a value of
23°35', as did al-KhazinT (see Chapter 1). See Hartner [3],
503-504 for comparative tables of Islamic values; see also
Kennedy [1973], 32-53. Since only two globes have gradu-
ations smaller than a single degree, and many globes use

intervals of two degrees or more, the obliquity of the ecliptic
could only be approximated on a globe. Most globes appear
to use a value of about 23'/2° or 24°; one globe (No. 76)
has a measurable obliquity of 23'/2°, and two globes (Nos.
94 and 100) actually have the value of the obliquity written
near each solstitial point, in both cases 23'/2°. Globe No. 73
made by Muhammad "^AIT al-HusaynT in 1221 H/AD 1806,
states that the obliquity is "23 degees, 30 minutes, and 17
seconds," although the globe itself is technically non-func-
tional and inaccurate.
   '* Two globes appear to have no holes at the celestial
poles, but only at the ecliptic poles: No. 59, made in 535 H/
AD 1140-1141 by an apprentice to a well-known astrolabe
maker, and No. 9, dated 981 H/AD 1573-1574, which is a
poorly executed non-functional globe, suspended at the
north ecliptic pole by a cord. There are three globes that
have holes drilled at other positions along the solstitial colure
whose purpose is unknown; these are Nos. 7 and 8, both by
Ja'^far ibn "^Umar ibn Dawlatshah al-KirmanT, and No. 86.
In the case of globe No. 44, possibly unfinished, there are
no holes at either the equatorial or ecliptic poles but only at
points about 25° on the other side of the equinoxes; similarly
positioned holes are found on globe No. 45.
   '^ There are two known exceptions to this where standard
numerals are used instead of abjad numerals, both products
of the Ottoman empire: No. 31 made in Cairo by Ridwan
in 1112 H/AD 1701, and No. 75, made by Husayn Husnu
KanqarIT in 1299 H/AD 1882. Two globes employ both abjad
and standard numerals (Nos. 53 and the ring of No. 96).
For the abjad alphabet system of numerals see Weil Sc Colin.
Numerals in DevanagarT script are found on Nos. 29 and
54.
   "' These are the diameters of Nos. 37 and 57. The
diameter of No. 10, also graduated by 2° intervals, is unfor-
tunately not known. The ungraduated globes vary in size
from 147 mm to 68 mm, with a median size of 112 mm.
   " No. 11 of the catalog also has the ecliptic numbered
continuously, but at the same time it has the ecliptic num-
bered by 30° segments.
'" For the term mayl kulli see Morley, 12 note 2.
'^ See Chapter 2, note 13.
   20 Nos. 89, 99, and 115. See al-BTrQnl Astrol., 147, for
confusion in the use of the term dd^irat al-mayl for declina-
tion circle and for the ecliptic. On celestial globes the expres-
sion is never used for the ecliptic. One globe. No. 33, states
in its lengthy inscription that it is "adlm al-mayl, "lacking the
declination [circle]." It must be interpreted in this case as
referring to the equinoctial colure, since the ecliptic is pres-
ent and the equinoctial colure is missing from the globe,
which also has a set of six meridian circles that have been
shifted 6° westward so that none pass through the equinoxes.
   2' On the relatively few globes that have meridians other
than the equinoctial and solstitial colures (see Chapter 2,
note 6) the meridians are sometimes labeled dd^irat al-mayl
or dawd^ir muyul.
22 Kennedy [1956], 169.

306

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



   22 Of Class A globes there are 19 that have one or more
rings contemporary with the globe; there are 15 Class B
globes and of Class C a possible 15. See the Tables in Chapter
7 for details as to which globes appear to have contemporary
rings or stand.
   2* Only one globe of Class A (No. 6 whose authenticity is
somewhat questionable) and two of Class C (Nos. 113 and
116) have the order of the numbered segments of the
meridian ring beginning at the points alongside the celestial
equator and terminating at the poles. These are the only
extant globes to have such an ordering.
   2^ Nos. 66 and 69 by Diya^al-DTn Muhammad have rings
with holes, but they could not be examined sufficiently to
establish the pattern; No. 27, also by Diya^ al-DTn Muham-
mad has holes for six latitude settings, but since the ring is
ungraduated (Punfinished) the settings cannot be deter-
mined.
2^ See Miquel.
   2' For the localities named on the horizon rings of globes
Nos. 31 and 40 see their catalog entries. Globe No. 76,
which has 144 place names on the ring, was not available
for analysis of the names.
   28 Neugebauer HAMA, 934 and Gunther AW, 1:180. The
Fortunate Islands were thought to be 35° west of Green-
wich.
   2^ Numbering the meridian ring graduations from the
point alongside the equator toward the poles would present
the advantage of reading the declination of the star directly
from the ring, but would require that the globe be rectified
for a given terrestrial latitude by either placing the comple-
ment of the latitude (colatitude) at the north point of the
horizon ring or by placing the degree corresponding to the
latitude at the zenith. The latter arrangement of gradua-
tions, though it seems equally convenient, does not appear
to be a common pattern.
   ^° A somewhat obscure chapter entitled "De saber cuemo
se fazen las armillas dell atacyr en la espera, et egualar las
casas segund la opinion de Hermes, et cuemo obren con
elias,' by an unidentified author "Don Xosse nuestro alfa-
quin." See Alfonso el Sabio, 1:206. See also Lorch [1980b].
   2' In addition to copies of the Qusta ibn LQqa treatise
(see Worrell and above Chapter 1 note 96), the tract by al-
KhazinT (see Lorch [1980b]), and the treatise on the use of
celestial globes by al-SQfi (see Chapter 1, note 111), there
are a number of others, such as one entitled Risdlah fl al-
kurah dhdt al-kursl exfdYxt in several copies including Prince-
ton, Garrett Collection (Yahuda Section) MS 4967, and
Berlin, Arabic MS 5869. It is also of interest to look at some
of the western treatises on the use of celestial globes, such
as those by Holland, Bion, Hood, and Wright.
22 Worrell, 286-290. Cf. also Lorch [1980b].
   22 See Hartner [1939], 2552; al-BTrunT/l^^ro/., 150. See
also Qusta ibn LQqa, chapters 17-18 (Worrell, 291); cf. the
chapter appended to the Spanish translation of Qusta ibn
LQqa (Alfonso el Sabio, 1:206-208).
2"* See Wensink Sc King.
  
2^ For examples of small portable gnomons, see Rohr,
114. A simple needle or pin will do, especially with papier
mache or wooden globes. See also Lorch [1980a], 157.
   2*^ Such as the calculations described by Hartner [1939],
2551-2554 and by E.S. Kennedy and M. Destombes in the
introduction to a^-Sufl Asturldb, 13-44.
   ■" The zarqaliyah astrolabe was designed as a universal
planispheric astrolabe usable at any latitude, but it was not
that commonly employed. See Chapter 1, note 154.
2« See Hartner [1939], and North.
   2* For treatises on the use of spherical astrolabes by Qusta
ibn LQqa (the authenticity of this treatise has been ques-
tioned), al-BTrunT, AbQ al-'^Abbas al-NayrTzT (died ca AD 922),
and AbQ "^AIT al-Hasan ibn ''AIT TJmar al-MarrakushT (died
AD 1262), see Sedillot [1841] and Seemann [1925]. For a
treatise by al-RQdanT, see Pellat [1975] and Janin. See also
Maddison [1962] and Alfonso el Sabio, 2:113-122. For the
shdmilah, a hemispheric astronomical instrument, see Frank,
and Sedillot [1841], 148.
"" Hartner [1], 727-728.
"' Maddison [1962], 102 note 6.
^•2 Maddison [1962].
   ''2 Collection of Signor Ernesto Canobbio, Como. The
instrument, whose rete is missing, was displayed at the exhi-
bition "Science and Technology in Islam" held at the Science
Museum, London, as part of the Festival of the World of
Islam, April—August, 1976. For a description of it, see
Canobbio.
   '''* Globes Nos. 7 and 8, both fourteenth-century, have a
set of holes drilled on the other side of the ecliptic poles
from the celestial poles and at a distance equal to the obliq-
uity of the ecliptic along the solstitial colure. Globe No. 86
has sets of holes about 25° distant along the solstitial colure
from the ecliptic poles and from the equatorial pole. The
purpose of these holes is unknown.
'*^ Tannery, Hist., 53; see also Hartner [1], 727.
   '^^ From a personal communication he circulated among
various people interested in Islamicate astronomical instru-
ments. See also Price [1979].
   "^ Neugebauer [1949], 246-247; Diels [1924], 213-219,
246-247, and plate XVII; Maass [1902]; Rehm 8c Weiss;
Drachmann; Neugebauer HAMA, 869-970; Price [1975],
59.
   ^" Maddison [1962], 103 note 1 3, mentions the terrestrial
globes with revolving rete found on some European astro-
nomical clocks, where he notes they are not observational
instruments and that the rete does not fit closely over the
globe, which has no horizon line or almucantars.
   ■•^ Eliminating for this discussion Nos. 63, 73, 82, and 83,
which represent a separate tradition in globe making.
•"''Seemann [1925], 33-34.
   ^' For example, see Welch [1979], 186-187, which illus-
trates a brass bowl from India of the first half of the
seventeenth century. Compare the twelve roundels decorat-
ing the unusually thick edge of a thirteenth-century Persian
astrolabe with geared calendar movement made by Muham-
  
NUMBER 46

307



mad ibn AbT Bakr ibn Muhammad al-RashidT al-Ibari al-
Isfahani in 618 H/AD 1221-1222. This astrolabe is now at
Oxford, Museum of the History of Science, Inventory No.
IC 5 (see Gunther AW, 1: plate XXVI). The zodiacal figures
on these roundels are shown with planetary symbols that are
absent on the celestial globes. See also the 12 zodiacal figures
on a twelfth-century bronze mirror of Mesopotamian origin
made for an Urtuqid ruler (Fiirst Ottingen-Wallerstein Col-
lection; see Kuhnel, 164 fig. 131, and Du Ry, 174 fig. 141),
and the zodiacal signs on the border of a bronze mirror
inlaid with silver made in Syria about AD 1320 (Topkapi
Sarayi Museum; see Du Ry, 149). See also Fehervari, 73-
74, for twelfth- and thirteenth-century zodiacal medallions
from Iran. Of interest also are the designs by al-Jazari for
zodiacal signs to be put on a large copper disc for a water
clock (see, for example, Atil, 103).
   ^2 See Gascoigne, 140, for a color plate of the coins issued
by JahangTr.
" See Taddei; and also Warburg, 2:631-639.
** See Ettinghausen.
   " Globes Nos. 55 and 58 also do not fit the basic cate-
gories. The former is probably simply an unfinished product
of the standard Class A globe which was being made by
someone not an instrument maker, that is, the constellation
outlines were engraved first and then left unfinished without
the stars. The second globe was not examined sufficiently
to determine its nature.
   ^* There is need for a comprehensive and comparative
examination of constellation imagery in Islamicate decora-
tive and scientific illustration in an effort to see how many
different traditions there were and where they arose. The
study should include all the celestial globes with constella-
tions, those astrolabes which have constellation figures, the
numerous illustrated al-SQfT manuscripts, the illustrated trea-
tise by al-BTrunT (London, British Library MS Or. Add.
7697), along with the Aratus, Abu Ma'^shar, and al-QazwTm
illustrations, using Latin and Byzantine sources as well as
Islamic ones. An eleventh century Persian astrological hand-
book of Shahmardan RazT with pictorial representations of
the constellations has recently come to light in a Cairo
manuscript and is now being studied by E.S. Kennedy (see
Project. Med. Isl. Astron. 24, no. 45); this treatise may also be
of importance. See also Zick-Nissen.
   ■" Ptol. Aim. [1], 2:179-185; see Chapter 1, and Chapter
1 note 50.
   ^8 This according to Worrell, 287. In the two manuscript
copies of this treatise examined by this author (Oxford,
Bodleian Library MS Hunt. 584, folio 7a and MS Arab. e.
94, folio 6a), the text does not mention a "stone" but employs
the general word sphere {kurah). An early fourteenth-cen-
tury Latin treatise on a celestial globe {sphera solida) "com-
pilatus a magistro Johanne de Halebeke Flandrensi experto"
describes a wooden sphere, although metal is mentioned as
a less satisfactory alternative, which is shaped by turning it
on a lathe about one axis and then turning it again about a
perpendicular axis (Lorch [1980a], 156).
  
^^ Alfonso el Sabio, 1:163-169. The author wishes to
thank Judy Erickson von Gunten for preparing an English
translation of the Spanish text.
   *"" Ceni may perhaps be qdnr', meaning dark red, or per-
haps slnl, "Chinese" as suggested by Seemann [1925], 52.
   •*' Alfonso el Sabio, 2:125-216; Seemann [1925], 56-57
presents a German translation and reproduces diagrams
from the Spanish text.
   ''2 Compare the late twelfth-century Islamicate techniques
for casting by closed mold boxes with green sand as the
casting medium, and casting by the cire perdue process de-
scribed by al-JazarT in his treatise on mechanical devices
written in AD 1206 (al-JazarT, 191-195, 267-268, and 276;
Wiedemann & Hauser, 230-237).
   •^2 Alfonso el Sabio, 2:118-120; Seemann [1925], 55-56
gives a German translation with diagrams.
   '''' As early as the first century BC there is a reference in
the Geography by Strabo to a sphere being shaped on a
turning-lathe (see Chapter 1, note 29). See also Cave.
   " The thirteenth-century astronomer al-MarrakushT
compares the procedure for drawing the circles on a spher-
ical astrolabe where the maker begins with the horizon and
meridian to the apparently usual practice of beginning a
celestial globe with the solstitial colure and celestial equator
(Sedillot [1841], 142).
   '*'^ Alfonso el Sabio, 2:118-124; Seemann [1925], 54 and
fig. 12.
   •^^ Lorch [1980b], 296-297; omitted by Sedillot in his
edition of al-MarrakushT.
   ^" The traveler Chardin detailed the equipment used in
Safavid Persia for making the remarkably steady and precise
circles and lines on a planispheric astrolabe, but the tech-
niques do not seem to be applicable to a celestial globe. See
Chapter 1, note 167.
   ""^ As described in the chapter appended to Librio del
alcora, Alfonso el Sabio, 1:206-208.
'"Lorch [1980b], 297.
" Kennedy [1956], and Ibn al-Salah.
'2 See Mercier.
   " The epoch of al-SQfT was taken to be 364 H/AD 974.
See Hartner [1939], 311 of the reprint, where an addition
has been made to the original article.
   '''* For the problems involved in dating an instrument on
the basis of the position of the vernal equinox and position
of stars see Poulle (1956], and Michel [1947]. See G.R. Kaye
[1921], 16-19 for an analysis of the star positions on globe
No. 71 made by Diya^ al-DTn Muhammad in 1087 H/AD
1676-1677; see also Destombes [1957].
   "Al-BattanT Op. astr., 3:210-214, Arabic text; Latin
translation in 1:139-142.
^Ubid., 3:213.
" Sedillot [1841], 110-115 and plates 4 and 5.
   '" In contrast to al-MarrakushT's theoretical sophistication
but practical inexperience is the design for an instrument to
measure the central point of three points of unknown posi-
tion on the surface of a sphere, by the author of the Book of
  
308

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Knowledge of Ingenious Mechnical Devices, Ibn al-Razzaz al-
JazarT. This late twelfth-century writer demonstrates enor-
mous acquaintance with metallurgical techniques but quite
rudimentary knowledge of mathematics and, curiously, no
interest in the problems and methods of producing spheres
(al-JazarT, 196-198, 268-274).
" See Allan, 52.
   ^° For example, Nos. 18, 28, 32, 62, 65, 77, 85, 86, 89,
100, and 115.
**' See, for example, Mukherjee, 167-176 et passim.
"2 Ibid., 204-210; and Allan, 38-39.
"2 Smith, 124-128; and Hodges, 73-76.
   **'' In a communication from the Assistant Keeper of the
Department of Technology, the Royal Scottish Museum,
Edinburgh, dated 27 June 1972, Mr. G.W. Anderson sug-
gested after consultation with a silversmith that globe No.
27 of the present catalog must have been raised, adding that
the thin cross-section is a strong argument against casting
"quite apart from the almost impossible technical feat of
casting a seamless globe." In November of the same year,
however, he reported that after examining the inside of the
globe with a tiny electric lightbulb he concluded that the
rough and pitted interior surface and protruding small brass
stumps were overwhelming evidence that it had been cast
rather than raised.
   **•" Only the seamless globes Nos. 9, 27, 28, 38, 42, 48,
60, 86, and 88 have been inspected.
   ""Hodges 70-73; and Mukherjee, 213-225, 230-240.
Compare the chapters on seventeenth-century Chinese cast-
ing by the lost-wax method in the treatise by Ying-Hsing,
159-170. See also Smith 105-115.
   "' The analysis was carried out by Maurice Salmon then
with the Laboratory, but for reasons beyond my control the
results have yet to be published. As a result I must simply
summarize in very general terms some of the results.
"" Allan, 52 et passim.
«^ See Steinberg, 103-138.
   ^° A preliminary attempt was made to compare volume
and weight of some of the globes to see if a clear cut pattern
emerged when compared with construction methods. Un-
fortunately, while weight was available for 33 of the metal
globes, the wall thickness was known for only 5 globes, one
of which was a globe whose weight was unknown. The four
globes for which both weight and wall thickness were known
were: No. 38, with the thickness varying from .47 mm to 1
mm; No. 9, with a wall thickness of 2.5 mm; No. 41, with a
thickness of the shell of 2 mm; and No. 48 whose wall
thickness is 0.43 mm. For these four globes, all of which are
known to be seamless, the resulting densities are 20.12
(taking a figure of 1 mm for No. 38), 11.35, 4.81, and
33.75, respectively. The resulting densities or specific grav-
ities of the alloys for these four globes cover such a large
range of values as to be quite inconclusive. This may be due
to the fact that the walls of many globes seem not to be
uniform in thickness and frequently incorporate thick plugs,
and the fact that for the last globe (No. 48) the weight

included that of the metal axis rod. That of globe No. 9 is
so low (4.81) as to make one question the reliability of the
measurements. Thus no conclusive results can be drawn
from these four sets of measurements concerning the den-
sities of alloys in metal seamless globes. When the weight
and diameter of the other 29 globes are compared, taking
an arbitrary value of 1 mm for the average shell thickness,
the results are slightly more suggestive for a density in the
range of 11.2 to 22.7 appears to be the rule for nearly all
the seamless items. The exceptions to this are globes No. 87
and No. Ill, which are known to have very thick walls,
although no exact measurement has been taken. Globes
known to be in hemispheres range in density from 5.4 to
24.1. One might conjecture that those globes showing a
lower density in the range of 5.4 to 9.1 were made by raised
hemispheres, while those in the higher values of 1 1 to 24
are of cast halves. Since, however, the shell thickness varies
from globe to globe and even within one globe, it is not
possible to draw any definite conclusions concerning the
densities of the alloys employed in the globes until many
more measurements are available for comparison.
   ^' Mukherjee, 213-241 et passim. Unfortunately this sur-
vey did not include modern Pakistan where Lahore, the seat
of the most productive workshops through the nineteenth
century, is located.
^2/itrf., 230-240, 265.
^2 Al-JazarT, 191-195, 267-268, and 274.
^^ Mukherjee, 357.
^^ See, for example, Allan, 19; and Mukherjee 433.
Chapter 3
   ' The designation of lunar mansions along the ecliptic is
not a frequently encountered feature of celestial globes and
is not known to occur on any other certain products of the
Lahore workshop other than No. 11 and the Smithsonian
globe (No. 38). On all the globes where lunar mansions are
indicated along the ecliptic (e.g., Nos. 2, 59, 60, 74, 76, 84,
and 85) the order begins with the first lunar mansion at the
vernal equinox, with only the two exceptions of Nos. 11 and
38.
Chapter 4
   ' I would like to thank Emilie Savage-Smith for encour-
aging me to undertake an examination of the Smithsonian
globe in the spirit of cooperative scholarship. I am grateful
to Hanns-Peter Schmidt, Martin J. Powers, and Estelle Whe-
lan for criticism and advice.
2 See, for example, Wellesz [1965], plate 3.
2 AI-SQn 5wi/;ar, 195.
   '• Although as a rule Cepheus appears as a beared male
both on globes and in the al-$ufT manuscripts, Centaurus is
sometimes shown beardless, as in Oxford Bodleian MS
Marsh 144 (see Wellesz [1965], plate 21).
  
NUMBER 46

309



   ^ For lithographs of the figures illustrating the latter two
manuscripts, see al-SQfT Suwar [trns.], plates I-VI.
   ^ For lithographs of the constellation figures on globe
No. 5, see Drechsler, plates 1-lV.
'Wellesz [1964], figure 2.
   " The Gemini appear as clothed figures in the al-§QfT
manuscript in the Metropolitan Museum of Art, New York
(Upton, fig. 33), on the Malcolm globe (No. 4) where the
twins appear to wear knee-length leggings (Pinder-Wilson
BM, fig. 146), and in an unpublished al-SQfT manuscript in
the Freer Gallery of Art, Smithsonian Institution, Washing-
ton, D.C. (MS No. 07-626, folios 53b and 54a).
   ^ This device appears only once on the Smithsonian globe,
and resembles the rendition of shoulder joints on figures
such as Centaurus, Andromeda, and Hercules on globe No.
1, dated 473 H/AD 1080. For a lithograph of the figures on
globe No. 1, see Meucci, plate 1.
'" See Chapter 4, note 6.
" For lithographs of globe No. 26, see KlQber.
'2 A\-Suf\ Suwar [trns.], plate V, figure 13.
   '2 This sort of necklet also appears to be in evidence on
globe No. 26 and in Oxford Bodleian MS Marsh 144; for
an illustration of the Andromeda figures in the Oxford
manuscript see Wellesz [1965], figures 12-13.
   '■* Perseus' winged footgear, as well as a cap of darkness
and a wallet, were given to him by the Hyperborean nymphs,
and appear as his standard attributes as early as the fifth
century BC (see Avery, 854).
   '^ In Greek mythology, Bootes was "supposed to repre-
sent a man holding a crook and driving the Bear (Ursa
Major)" (Avery, 221). This identification does not account
for his winged footgear on the Smithsonian globe, which
may simply reproduce the type of footgear worn by Perseus
by a sort of visual analogy.
   '^ Cepheus's boots appear to be part of the costume that
identified him as an Ethiopian king, the father of An-
dromeda and husband of Cassiopeia.
   " The adaptation or transformation of attributes such as
these individualized forms of headgear from pre-Islamic
dmes via copies of the al-SufT treatise into later Islamic art
promises to be a rich field for future study. Such obviously
Muslim modifications as Auriga's turban become standard
and are repeated, while the headgear of figures such as
Cepheus show a much broader range of adjustments.
   "* The easy identification of these beasts depends in large
part on the rendition of salient characteristics of their heads,
paws, and tails. Once again, certain features, such as the
length and width of Ursa's tail, show greater variation than
others, a factor that might have been due to artisans's direct
observation (or lack thereof) of certain animal types in nature
or in captivity.
   '^ An alternate practice was followed on at least two of
the globes (Nos. 4 and 5), where the fur of animals was
indicated by overall patterns of hatching arranged in rows.
See, for example, Pinder-Wilson BM, especially figure 136.
2° For a comparable Draco in Oxford Bodleian MS Marsh

144, see Wellesz [1959], figure 20. The head of the Draco
figure on the Smithsonian globe is similar to that of the
constellation as illustrated in an al-$QfT manuscript in Paris
(see Blochet, plate LXXXVIII). The same type of head was
used for Serpens in the same manuscript (Blochet, plate
XCI).
   2' Globes No. 12 and No. 14 are also signed by Qa^ im
Muhammad, but photographs are unavailable for study or
too poor to permit analysis. The brother of Qa^ im Muham-
mad, Muhammad MuqTm, also made a globe with constella-
tion figures (No. 15), but this was also unavailable for study.
   22 Globes Nos. 19, 21, and 24 signed by Diya^ al-DTn
Muhammad were also unavailable for study or the available
photographs were not detailed enough to allow comparison.
   22 The Paris al-SQfT manuscript mentioned above (chapter
4, note 20) shows Auriga in the same posture as on the
Smithsonian globe; see Blochet, plate XC
2^ Al-SQH Suwar, 93.
   2^ On the Malcolm globe (No. 4), for example, the attri-
bute more closely resembles a horse whip than it does on
the Smithsonian globe; see Pinder-Wilson BM, fig. 141.
   2** See, for example, Blochet, plate XC; Marfin [1912], 2,
plate 35; and the unpublished Freer al-SQfT MS (see chapter
4, note 8), folios 19b and 20a.
   2' This attribute can be seen, for example, in Oxford
Bodleian MS Marsh 144 (see Wellesz [1959], fig. 8), and on
the Malcolm globe. No. 4 (see Pinder-Wilson BM, fig. 141).
2" Al-Sufi 5uwar, 93.
   2^ The facial type and clothing of this figure mark it as
coming from a Mughal milieu in a fashion more immediately
obvious than is the case with any of the figures on Qa^ im
Muhammad's globes. This interest in details of contempo-
rary dress and close kinship with contemporary painting
styles are identifying characteristics of Diya^ al-DTn Muham-
mad's oeuvre.
   "^" An example of an even more radical anatomical trans-
formation on one of Diya^ al-DTn's globes occurs on No. 22,
where Capricorn is rendered as a reindeer-like creature.
   2' A remark should be added here concerning the signif-
icance of signatures on celestial globes and other metalwork.
We still do not know, in the vast majority of cases, the
professional identity of the individuals who signed objects,
although it is usually assumed, as here, that they were
overseers (possibly master craftsmen) who took ultimate
responsibility for the quality of a product that was in fact
produced by a group.
   22 The term "inverted" refers to the drawing of the ears
within the circumference of the head instead of protruding
outside of it.
   22 In affirming the femininity of this figure by articulating
her anatomy, the maker of the Smithsonian globe joined a
small group of artisans who departed from the usual practice
of differentiating between male and female figures by means
of hairstyles and attributes only. The maker of the Malcolm
globe, No. 4, represented Andromeda as bare-breasted with
a calf-length skirt (see Pinder-Wilson BM, figs. 140, 144);
  
310

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



both Andromeda and Virgo are distinguished by hairstyle,
attributes, and feminine costume only in Oxford Bodleian
MS Marsh 144 (see Wellesz [1964], figs. 4, 5, and 8).
   2'' This hairstyle may be a more elaborate variation of
Orion's scalloped coiffure or headdress on the Smithsonian
globe.
   2^^ Diya^ al-DTn's Virgo, with her sinuous, slender figure,
jewelry, and more elaborate sash, is dressed more like a
court lady or dancing girl than like an angelic figure. Her
garments and attitude may perhaps be related to represen-
tations of court women in Rajput painting; see, for example,
a painting dating from about AD 1660, which shows Prince
Murad Baksh receiving a lady at night (Archer, plate 46).
   2^ The anatomically correct representation of feline hind-
quarters posed a problem for Islamic artists in various media.
To cite one example, a group of leonine incense burners in
bronze from twelfth- to thirteenth-century Persia exhibit a
range of leg proportions and postures that indicate some
difficulty in coming to grips with the proper placement of
knee and ankle joints (see Aran).
2' Wellesz [1959], fig. 33; Martin [1912], 2, plate 38.
   2* For an example, see an illustration from the sixteenth-
century copy of a manuscript attributed to a certain Muham-
mad al-Sooudi (Blochet, plate CLXXIV).
2^ See Chapter 4, note 18.
   '"' See Upton, fig. 34; and al-SQfi Suwar [trns.], plate II,
figure 26, respectively.
   ■" There is also a large vase-like form at the side or center
of the vessel, the identity and function of which is unclear.
   "'2 These bracket-like forms resemble cartouche motifs
used widely in Islamic leatherwork, in particular on book
bindings. See, for example, Kuhnel, figs. 41 and 44.
   "2 See Pinder-Wilson BM, fig. 155; and Wellesz [1965],
fig. 20.
"*'* See text for Figure 81 (Argo Navis).
   ''"'' Upton, fig. 48; al-SQfi Suwar [trns], plate IV, fig. 40,
respectively.
■"* See text for Figure 81 (Argo Navis).
   ■^ An exception is globe No. 1, where all the human
figures are naked except for Andromeda.
■*" On variants of headgear, see Chapter 4, note 17 above.
   ■"^ Generally, shorter (knee-length) tunics appear to have
been worn by common laborers, farmers, hunters, or lowly
attendants—i.e., men whose professions demanded the
greatest freedom of movement. Calf-length tunics, on the
other hand, appear as the typical dress of princes and their
courtiers in Mughal painting. An interesting example of the
juxtaposition of these costume types occurs in the marginal
illustrations on a page from an imperial album of JahangTr
dated ca. AD 1605; see Beach, 49, 51, folio 7v.
   •'^" See Welch [1978], plate 17 (a miniature from a dis-
persed copy of the J ahdnglr-ndma showmg Jahangir in Dar-
^ar and dated ca. AD 1620).
   "'^' See Beach, 60-61, no. 13 {Jahangir Entertaining Sheikhs
from a J ahdnglr-ndma dated ca. AD 1615-1620).
"^2 See Welch [1978], plates 18 (a miniature from a dis-

persed copy of the Tuzuk-i-Jahdnglri, ca. AD 1615) and 21
(an album painting er\t\t\edjahdngtr's Dream, ca. AD 1618-
1622, which shows JahangTr himself embracing Shah "^Abbas
Safavi of Persia).
   •'2 The earliest of these works, dated AD 1583, is a page
from a copy of the Kitdb-i-Sa"at, an astrological treatise
written at Hajipur. The particular miniature in question,
folio 24v, illustrates Mars in Aries; the figure of Mars,
dressed in a short tunic with single-looped sash, and carrying
a sword over his right shoulder and a human head in his left
hand, advances resolutely toward the right. His somewhat
squat body bears an uncanny similarity to several of the male
figures on the Smithsonian globe. On this manuscript, see
Falk, Smart, Sc Skelton, 19-23, especially page 21. For
examples of later paintings depicting male figures wearing
tunics and single-looped sashes, see Ibid., p. 80; Beach
[1978], 37, no. 1 (folio 41r); 50, no. 7; and 158, no. 56; and
Welch [1973], 11, no. 76, and 103, no. 62.
Chapter 5
' Tibbetts [1965], [1971]; Kunitzsch [1967], [1977].
2 Kennedy [ 1956]; al-BTrQnT Qaniin, 3:987-1126.
2 Ptol. ^/m. [1], 2:167-68.
   * Ptol. Aim. [trns. Ger.], 2:48, 66, 67; Baily; Peters 8c
Knobel.
•'Al-SQfT 5wu;ar, 199.
" Ulugh Beg Zij [trns. Eng.]; Peters; Baily.
' Ulugh Beg Zij [trns. Eng.], 8-9.
^ A\-SufI Suwar, 94.
® Ulugh  Beg Zij [trns.  Eng.],   10;  for comparisons of
catalogs see Knobel [1875]; Moesgaard; Peters: Tallgren.
'° A\-Kash\ Letter, 96.
" Gundel Sc Boker, cols. 522-528; Hartner [3]; Hommel;
Encyc. World Art, 2:34-55 and plate 20 ("astronomy and
astrology"); see also Brown; and Gleadow.
'2 Kuntizsch[1961], 22.
   '2 See Ideler 214-218; and Kunitzsch [1961], 24. for
various interpretations.
   '" Kunitzsch [1961], no. 166b; al-BTrunT Astrol., 163; al-
Sufi Suwar, 289.
'■■^Wissmann, 880-881.
'" This latter image is actually illustrated in two extant
manuscript copies of al-SQfi's treatise (Wellesz [1964], 91).
"Wiedemann [1926].
'" Benhamouda; Hartner [3]; Ideler; Mesnard; Samaha;
Jurdak; Knobel [1897]; Filers. See also the general works of
Allen; Higgins. For bibliography see Kunitzsch [1959], 3-9.
'^ Al-SQfi Suwar (all references to al-SQfi in the following
two sections are to the Hyderabad edition); al-BTrunT Astrol.,
157-164; al-QazwTnT; see also Ideler.
20 Neugebauer OCD; Buttmann; Boll-Gundel; Aujac; La-
bordus; Aurigema; Gundel; Schadewaldt; Ramsey Sc Wilkins.
2' Kunitzsch [1961] and [1974].
22 Whitney; Chu.
22 Yampolsky.

NUMBER 46

311



2" Hommel; Kunitzsch [1959], 53.
2^ Weinstock; Mercier, 42.
2^ Wissmann, 873, 880-881.
   2' The cosmical setting of a star or asterism and the
heliacal rising of the opposite star-group, in the ordering of
the 28 lunar mansions, marked the beginning of a period
called naw^, hence the term anwd^ used for this early system.
The term anwd^ is also used generally for all pre-Islamic
terminology, conceptions, and significances applied to star
groups by the early Bedouins. See Pellat EI, and Pellat
[1955]. In much of the literature on lunar mansions and the
anwd^ system, a confusion in the use of the term acronychal
has occurred. The cosmical setting of a star or asterism in
the west as the sun rises in the east has frequently, but
incorrectly, been called the acronychal setting. The heliacal
rising, sometimes called cosmical, occurs in the east at sun-
rise.
   2* Al-BTrQnT Chron. [trns.], 335 and al-BTrQnT ^.5<ro/., 164,
says the Hindus had 27 divisions; see also Gundel & Boker,
cols. 511-514; cf. Kaye [1924], 22-23. Others such as
Whitney; Yampolsky; Hommel; and Mercier, all give 28.
2^ Al-BTrQnT Chron. [trns.], 335-336.
   2° Kunitzsch DSB. Ibn Qutaybah is quoted extensively in
Kunitzsch [1961].
•" See Pellat [1955].
   22 See, for example, the calendar written in AD 961
(see Calendrier de Cordoue) and the thirteenth-century
one by Ibn al-Banna^.
^•' Sabra DSB.
   2"* Al-BTrunT C/iron. [trns.], 338 (translation of Sachau;
words in brackets added).
  
^^ As seen, for example, in the treatises of the thir-
teenth-century authority on the occult, al-Bunl (see al-
BQnT, 18-24). See Savage-Smith &c Smith. 39-42. espe-
cially the chart on pages 40-41. for a detailed discussion
of these patterns.
   ^^ For further discussion of lunar mansions see espe-
cially Kunitzsch [ 1961 ] 1 3-20 et passim; Gundel &: Boker,
cols. 511-519; Steinschneider [ 1864]; Hommel; al-BTrOnl
Astrol, 164, Chron. [trns.], 343-353, Qanun, 3:1139-
1157; al-Sufi Suwar, passim; al-QazwTnl, 1:42-54; see
also Ideler. See also al-BunT, 10-25; Ruska; Griffini;
Musil. For a brief discussion of the numerical symbolism
associated with the lunar mansions (e.g., their sum is
equal to the sum of all the integers of the number of
planets, 7 -I- 6 -I- 5 ... -I- 1 = 28) see Nasr [1965], 162-
163.
   2' Al-BTrunT Chron. [trns.], 353-354; see also Savage-
Smith & Smith, 40-41.
   2" One might expect qlfd^us; however, al-SiifT reads
qiqd^us, as do all globes that have diacritical points, except
for No. 4, which reads qifd^us. The two earliest globes.
No. 1 and 34, as well as the later No. 31, bear the title
al-multahib (the angry one), which al-SQfi also uses, as
well as qiqd^us. (See Kunitzsch [1974], 173-174, for a
discussion of the versions of Cepheus's name in Arabic.)
   ^^ It is of interest that in the diagrams illustrating the
oldest extant text of al-Sufi (Oxford, Bodleian MS Marsh
144), both no. 30 of Centaurus and no. 11 of Lupus are
eliminated from the drawing. See Wellesz [1959], fig. 17
and [1965]. fig. 21.
  
SELECTED BIBLIOGRAPHY

Abbott:
Nabia Abbott. Indian Astrolabe Makers. Islamic Culture,
1937, 11:144-146.
Alfonso el Sabio:
Alfonso el Sabio. Libros del saber de astronomia del rey D.
Alfonso X de Castilla. Compiled, with commentary, by Don
Manuel Rico y Sinobas; 5 volumes. Madrid: Don E.
Aguado, 1863-1867.
•^AIT ibn "Isa Asturlab:
"^AIT ibn ^sa. Kitab al-'^amal bi-1-asturlab li-'^AIT ibn "Isa.
Edited by P.L. Cheikho. Al-Mashrlq, 1913 (16):29-46.
[German translation by C Schoy. "^AlTb. ^sa. Das Astrolab
und sein Gebrauch. Isis, 1927, 19:239-254.]
Allan:
James W. Allen.    Persian Metal Technology, 700-1300
A.D. Oxford Oriental Monographs, no. 2. Oxford and Lon-
don: Ithaca Press for the Faculty of Oriental Studies and
The Ashmolean Museum, University of Oxford, 1979.
Anderson:
R.G.W. Anderson. Science in India: A "Festival of India"
Exhibition at the Science Museum, London, 24 March-1
August 1982. London: Science Museum, 1982 [exhibition
catalog].
Allen:
Richard Hinckley Allen. Star Names, Their Lore and
Meaning. 1899. Reprint, New York: Dover, 1963.
Aran:
Andrea P. Aran. Medieval Persian Incense Burners in
Leonine Form. Master's thesis, University of California,
Los Angeles, 1976.
Arat. Phen. [1]:
Aratus. Arati Phaenomena. Edited by J.T. Buhle; 2 vol-
umes, Leipzig: Weidmann, 1793, 1801.
Arat. Phen. [2]:
	. Arati Phaenomena. Edited, with introduc-
tion, by Ernst Maass. Berlin: Weidmann, 1893. [Second
edition, Berlin: Weidmann, 1955; third edition, Berlin:
Weidmann, 1964.]
Arat. Phen. [3]:
	.    Aratus. Edited and translated by G.R. Mair
(Loeb Classical Library). London and Cambridge, Massa-
chusetts: William Heinemann Ltd., and Harvard Univer-
sity Press, 1921.
Arat. Phen. [4]:
	.    Arati Phaenomena. Translated into French
by Jean Martin. Biblioteca di Studi Superiori, no. 25. Filol-
ogia Greca. Florence: La Nuova Italia Editrice, 1956.
Arat. Phen. [5]:
	.    Aratus Solensis, Phainomena: Sternbilder und

Wetterzeichen. Edited with German translation by Manfred
Erren. Munich: Heimeran Verlag, 1971.
Arat. Phen. [trns.]:
	.    Aratus Solensis, the Phenomena and Diosemeia.
Translated by John Lamb. London: J.W. Parker, 1848
[Aratus Solensis, Phenomena. Translated by John Lamb;
notes by Paul O. Holbrook. Lexington, Kentucky: King
Library Press, 1975.]
Archer:
W. Archer. Indian Miniatures. Greenwich, Connecticut
and London: Studio Books and New York Graphic Soci-
ety, 1960.
Aristotle PA:
Aristotle. Parts of Animals. Translated by A.L. Peck
(Loeb Classical Library). London and Cambridge, Massa-
chusetts: William Heinemann, Ltd. and Harvard Univer-
sity Press, 1937.
Arnold:
Sir T. Arnold. The Library of A. Chester Beatty: A Catalog
of the Indian Miniatures. Three volumes. London: privately
printed by J. Johnson at Oxford University Press and
published by E. Walker, Ltd., 1936.
Ashfaque:
Syed M. Ashfaque. i] Apxv r^s 'Aarpovoplas Kal i]
'Zivp-jioXr] Tcbv 'Apxoil(i>v 'EXXijvwp [The beginning of as-
tronomy and the contribution of the ancient Greeks].
Doctoral dissertation. Thessalonika: 1977 [in Greek].
Assemani:
Simone Assemani. Globus coelestis cufico-arabicus Veliterni
Musei Borgiani. Padua: Patavii Typis Seminarii, super-
iorum facultate, 1790.
Atil:
Esin Atil. Art of the Arab World. Washington, D.C:
Freer Gallery of Art. Smithsonian Institution, 1975.
Aujac:
Germaine Aujac. Representation du ciel des fixes en
grece ancienne. Revue d'histoire des sciences et de leurs
applications, 1976, 29:289-307.
Aurelius:
Aurelius (138). T. Aurelius Fulvus Boionius Arrius An-
toninus. In Pauly and Wissowa, editors, Real-Encyclopddie
der klassischen Altertumswissenschaft, 1896. II: columns
2493-2510.
Aurigema:
Luigi Aurigema. Le signe zodiacal du scorpian dans les
traditions occidentals de I'antiquite greco-latine a la Ren-
aissance. Civilisations et Societes, no. 54. Paris: Mouton,
1976.

312

NUMBER 46

313



Avery:
CB. Avery, editor.    The New Century Classical Handbook.
New York: Appleton-Century-Crofts, 1962.
Avien. Arat.:
Rufus Festus Avienus. Rufi Festi Avieni Aratea. Edited by
Alfred Breysig. Leipzig: Teubner, 1889. [For an earlier
edition, see Arat. Phen. [1], volume 2.]
Baily:
Francis Baily. The Catalogues of Ptolemy, Ulugh Beigh,
Tycho Brahe, Halley, Hevelius, Deduced from the Best
Authorities: With Various Notes and Corrections and a
Preface to Each Catalogue. Memoirs of the Royal Astronom-
ical Society, 1843, 13:1-248.
Baker & Stamford:
Patricia Baker and Vanessa Stamford. The Frescoes of
^Amra. Aramco World Mag-azine, July-August (I980):22-
25.
Barber:
Eric Arthur Barber. Aratus (1). In Oxford Classical Diction-
ary, second edition, p. 92.
Barrett.
Douglas Barrett. Islamic Metalwork in the British Museum.
London: The Trustees of the British Musuem, 1949.
Barthold:
Wilhelm Barthold [Vasilii Vladimirovich Bartol'd]. Ulug
Beg und seine Zeit. Abhandlungen fur die Kunde des Mor-
genlandes, Band 21. Leipzig: Deutsche Morgenlandische
Gesellschaft, 1936. [Reprinted in Nendein, Liechtenstein:
Kraus Reprint, 1966.]
Barthold EL
	.    TadjTk. In Encyclopaedia of Islam, first edi-
tion. 4:598-599.
al-BattanT Op. astr.:
al-BattanT. Al-Battanl sive Albatenii, Opus astronomicum. Ed-
ited by CA. Nallino; 3 volumes. Pubblicazione del Reale
Osservatorio di Brera in Milano, no. 40 parts 1-3. Milan:
Reale Osservatorio di Brera, 1899-1907.
Bazin:
Louis Bazin. Uber die Sternkunde in altturkisher Zeit.
Abhandlungen der Geistes- und Sozialwissenschaftliche
Klasse, no. 5. Mainz: Akademie der Wissenschaften und
der Literatur, 1963.
Beach:
Milo Cleveland Beach. The Grand Mogul. Williamstown:
Sterling and Francine Clark Art Institute, 1978 [exhibi-
tion catalog].
Beede:
Grace Lucile Beede.    Virgil and Aratus; a Study in the Art
of Translation. Chicago: The University of Chicago Li-
braries. 1936.
Beer:
Arthur Beer. The Astronomical Significance of the Zodiac
of Qasayr "^Amra. In K.A.C Creswell. Early Muslim Archi-
tecture, l(2):432-440.
Benhamouda:
A. Benhamouda.    Les noms arabes des etoiles. Annates de

I'Institut d'etudes orientates (Algiers Universite). 1951,
9:76-210.
Bion:
Nicolas Bion. The Construction and Principal Uses of Math-
ematical Instruments; Translated from the French ofM. Bion,
Chief Instrument-Maker to the French King: To Which Are
Added the Construction and Uses of Such Instruments as Are
Omitted by M. Bion; Particularly of Those Invented or Im-
proved by the English. Edited by Edward Stone. London: J.
Richardson, 1723. [second edition, London: Richardson,
1758. Reprinted, London: Holland Press, 1972.]
al-BTrunT Astrol.:
al-BTrunT. The Book of Instruction in the Elements of the Art
of Astrology, Written in Ghaznah, 1029 A.D.: Reproduced
from British Museum MS Or. 8349. Translated by R.R.
Wright. London: Luzac, 1934.
al-BTrQnT Chron.:
	.    Chronologic orientalischer Volker von Alber-
uni. Edited by C Eduard Sachau. Leipzig: Gedruckt auf
Kosten der Deutschen Morgenlandischen Gesellschaft, in
Commission bei F.A. Brockhaus, 1878. [See also Docu-
menta Islamica inedita. Edited by J. Fuck, pp. 45-98.
Berlin: Akademie-Verlag, 1952.]
al-BTrQnT Chron. [trns.]:
	. The Chronology of Ancient Nations—an Eng-
lish Version of an Arabic Text of the Athdr-ul-Bdkiya of
AlberHnl or "Vestiges of the Past" Collated and Reduced to
Writing by the Author in A.H. 390-1, A.D. 1000. Edited and
translated with notes by C Eduard Sachau. London: W.H.
Allen Sc Co., for the Oriental Translation Fund of Great
Britain and Ireland, 1879.
al-BTrQnT Qanun:
	.    Kitdb al-qdnun al-mas'^udl {Canon Masudi-
cus). Three volumes. Osmania Oriental Publications, no.
116. Hyderabad: Osmania Oriental Publication Bureau,
1383H[AD 1954-1955].
al-BTrunT India:
	.    Fl tahqiq ma li-l-Hind min maqulatin maqbu-
latin fl al-"aql wa mardhulatin. Edited by E.G. Sachau,
1887. [Reprinted in Leipzig: O. Harrassowitz, 1925.]
al-BTrQnT/ntfifl [trns.]:
	.    AlberunVs India: An Account of the Religion,
Philosophy, Literature, Geography, Chronology, Astronomy,
Customs, Laws, and Astrology of India. Translated by E.G.
Sachau; 2 volumes. 1888. [Reprinted in London: K. Paul,
Trench, Trubner, 1910.]
al-BTrunT Catalog:
	.    Zvezdnyi  katalog al-BTrQm's prilosheniem
Katalogov Khaiyama i al-Tusi [Stellar Catalog by al-BTrQnT
with Supplemental Catalog by Khayyam and al-TQsT].
Istoriko-Astronromicheskie Issledovaniia, 1962, 8:83-192.
Blochet:
E. Blochet. Musulman Painting, Xllth-XVIIth Century.
Translated from the French by Cicely M. Binyon. London:
Methuen & Co., 1929.

314

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



Boker:
Robert Boker. Die Entstehung der Sternsphaere Arats.
Berichte iiber die Verhandlungen der sdchischen Akademie der
Wissenschaften zu Leipzig, mathematisch- und naturwissen-
schafftliche Klasse, Band 99, Heft 5. Berlin: Akademie-
Verlag, 1952.
Boilot:
D.J. Boilot. L'oeuvre d'al-BerunT: essai bibliographic.
Melanges de I'Institut Dominicain d'Etudes Orientates, 1955,
2:161-256 [compare 1956, 3:391-396].
Boll [1901]:
F. Boll. Die Sternkataloge des Hipparch und des Ptole-
ma'ios. Biblioteca Mathematica, 1901, 2:185-195.
Boll [1903]:
	.    Sphaera: Neue griechische Texte und Untersu-
chungen zur Geschichte der Sternbilder, mit einem Beitrag von
Karl Dyroff Leipzig: Teubner, 1903.
Boll Gundel:
F. Boll and G. Gundel. Sternbilder, Sternglaube und
Sternsymbolik bei Griechen und Romern. In Ausfdhrliches
Lexicon der Griechischen und Romischen Mythologie, edited
by W.H. Rosher, Nachtrage zu Band 4 (in Band 6): 867-
1071. Leipzig and Berlin: Teubner, 1937.
Brahe:
Tycho Brahe. Tycho Brahe's Description of his Instruments
and Scientific Work as Found in Astronomia instauratae
mechanica (Wandesburgi 1598). Edited and translated by
H. Raeder, E. Stromgren, and B. Stromgren. Copen-
hagen: I Kommission hos E. Munksgaard, 1946.
Brockelmann GAL:
C Brockelmann. Geschichte der arabischen Litteratur.
Two volumes. Leiden: Brill, 1898-1902. Supplement.
Three volumes. Leiden: Brill, 1937-1942.
Brieux 8c Maddison:
Alain Brieux and Francis R. Maddison. Repertoire des
facteurs d'astrolabes et de leurs oeuvres: Premiere partie,
Islam. Paris: Centre national des recherches scientifiques,
in press.
Brown:
Robert Brown. Researches into the Origins of the Primitive
Constellations of the Greeks, Phoenicians and Babylonians.
Two volumes. London: Williams and Norgate, 1899-
1900.
Bulwer-Thomas:
Ivor Bulwer-Thomas. Conon of Samos. In Dictionary of
Scientific Biography, edited by CC Gillespie, 3:391.
al-BQnT:
Ahmad ibn '^AIT al-BQnT. Shams al-ma'^drif al-kubrd wa
latd^if al-"awdrif. Cairo [1945].
Burgoyne Sc Abul-Hajj;
Michael Hamilton Burgoyne and Amal Abul-Hajj,
Twenty-Four Mediaeval Arabic Inscriptions from Jerusa-
lem, L«van^ 1979, 11:31-51.
Buttmann:
Buttmann. Uber die Entstehung der Sternbilder auf der
griechischen Sfare. Abhandlungen der Koniglichen Akade-

mie der Wissenschaften, historisch-philologische Klasse
1826, pp. 19-63. Berlin: Koniglich Akademie der Wissen-
schaften [1829].
Byvanck:
A.W. Byvanck. De platen in de Aratea van Hugo de Groot,
with a Summary: the Illustrations in the Aratea of Hugo
Grotius, and a List of Illustrated Astronomical Manu-
scripts. Mededelingen der Koninklijke Nederlandsche Akade-
mie van Wetenschappen Afd. Letterkunde, new series, 12(2).
Amsterdam: Noord-Hollandsche Uitg. Mij., 1949.
Calendrier de Cordoue:
Le calendrier de Cordoue.    New edition; accompanied by
annotated French translation by C Pellat. Medieval Iberian
Peninsula Texts and Studies, vo\ume 1. Leiden: Brill, 1961.
Callimachus:
Callimachus. Aetia, Iambi, Lyric Poems, Hecate, Minor Epic
and Elegiaic Poem, and other Fragments. Edited and trans-
lated by CA. Trypanis (Loeb Classical Library). London
and Cambridge, Massachusetts: William Heinemann, Ltd.,
and Harvard University Press, 1958.
Calvert:
H.R. Calvert. Astronomy, Globes, Orreries and Other
Models. Science Museum Illustrated Booklet. London: Her
Majesty's Stationery Office, 1967.
Canobbio:
Ernesto Canobbio. An Important Fragment of a West
Islamic Spherical Astrolabe. Annali dell'Istituto e Museo di
Storia della Scienza di Firenze, 1976, 1(1):37-41.
Carra de Vaux:
B. Carra de Vaux. Ta^rlkh. In Encyclopaedia of Islam, first
edition, 4:672.
Casanova:
Paul Casanova. Une sphere celeste de I'an 684 de
I'Hegire. Memoires de la Mission archeologique frangaise du
Caire, 1892, 6:313-330.
Cass. Var.:
Cassidorus. Cassiodori Variae. Edited by T. Mommsen.
Monumenta Germanicae historica auctorum antiquissorum,
no. 12. Berlin: Weidmann, 1894.
Catullus:
Catullus. Catullus. Edited by Elmer T. Merrill. Cam-
bridge, Massachusetts: Harvard University Press, 1893.
[See also Fordyce.]
Cave:
J.F. Cave.    A Note on Roman Metal Turning. History of
Technology, edited by A.R. Hall and N. Smith, 2:77-94.
London: Mansell, 1977.
Chadenat:
Collection Ch. Chadenat: Instruments de mathematiques an-
ciens . . . dont la vente aux ench'eres publiques aura lieu:
Hotel Drouot ... le jeudi, 24 mai 1956. Commissaire-
priseur E. Ader, assiste par H.-D. Fromanger. Paris, 1956
[sale catalog].
Chardin:
Sir John Chardin. Voyages du chevalier Chardin en Perse,
et autres lieux de I'Orient, enrichis d'un grand nombre de

NUMBER 46

315



belles figures en taille-douce. Edited by E. Langles, fifth
edition. Paris: Le Normant. Imprimeur librarie, 1811.
Chu:
Coching Chu. The Origins of the Twenty-Eight Lunar
Mansions in Astronomy. Popular Astronomy, 1947, 55:62-
77.
Cxc.Arat. [1]:
Cicero.    M. Tuttis (jiceronis Arateorum quae super sunt. Ed-
ited by E. Baehrens. In Poetae Latini Minores,  1:1-28.
Leipzig: Teubner, 1879.
Cic. Arat. [2]:
	.    The Poems of Cicero. Edited, with introduc-
tion and notes, by W.W. Ewbank. London: University of
London Press, 1933.
Cic. Arat. [3]:
	.    Ciceron, Les Aratea. Edited, with translation
and commentary, by Victor Buescu. Institut Roumain
d'Etudes Latines, Collection d'Editions Critiques, no. 1. Paris
and Bucharest: Societe d'Edition "Les Belles Lettres,"
1941. [Reprinted in Hildesheim: G. Olms, 1966.]
Cic. Nat. deo.:
	.    Cicero De natura deorum. Edited and trans-
lated by H. Rackhan (Loeb Classical Library). Cambridge,
Massachusetts/London:   Harvard   University   Press and
William Heinemann Ltd., 1933.
Cic. Repub.:
	.    Cicero, De republica, de legibus. Edited, with
English translation, by Clinton Walker Keyes (Loeb Class-
ical Library). London and Cambridge, Massachusetts: Wil-
liam Heinemann, Ltd., and Harvard Unviersity Press,
1928.
Cic. Repub. [trns.]:
	.    Marcus Tullius Cicero: On the Commonwealth.
Translated, with introduction, by George Holland Sabine
and Stanley Barney Smith. Columbus, Ohio: Ohio State
University Press, 1929.
Cic. Tusc. Disp.:
	.    Cicero   Tusculan  Disputations.   Edited  and
translated by J.E. King (Loeb Classical Library). Cam-
bridge, Massachusetts and London: Harvard University
Press and William Heinemann Ltd., 1966.
Claudian:
Claudian.    Claudian. Edited and translated by Maurice
Platnauer (Loeb Classical Library); 2 volumes. Cambridge,
Massachusetts, and London: Harvard University Press and
William Heinemann Ltd., 1922.
Creswell:
K.A.C. Creswell, editor. Early Muslim Architecture,
Umayyads A.D. 622-750. Second edition; 3 volumes. Ox-
ford: Clarendon Press, 1969.
Czenakal (see also Tchenakal):
V.L. Czenakal.    The Astronomical Instruments of the
Seventeenth and Eighteenth Centuries in the Museums of
the U.S.S.R. Vistas in Astronomy, edited by Arthur Beer,
9:53-77. Oxford and New York: Pergamon Press, 1968.,

Destombes [1957]:
Marcel Destombes. Globes celestes et catalogues d'e-
toiles orientaux du moyen-age. Actes du VIII"" Congres
international d'histoire des sciences, Florence, 3-9 septembre
1956, pp. 313-324. Florence: 1957.
Destombes [1958]:
	.    Un  globe  celeste arabe du   XII'"" siecle.
(Comptes rendus de I'Academie des inscriptions, 3:300-313.
Paris, 1958.
Destombes [1959]:
	.    Un globe celeste inedit de I'epoque seljou-
kide (539 de I'Hegira). Actes du IX""' Congres international
d'histoire des sciences (Barcelona and Madrid), (1959): 447-
452.
Destombes [I960]:
	.    Les chiffres coufiques des instruments as-
tronomiques arabes. Physis, 1960, 2:197-210.
Destombes [1966]:
	.    La diffusion des instruments scientifiques
du haut-moyen age XV"'"" siecle. Cahiers d'histoire mondi-
ale, 1966, 10:31-51.
Dhama:
B.L. Dhama. Delhi Fort Museum. Archaeological Survey
of India, Annual Report 1928-1929:143-144, plate
LVIII. Calcutta: Superintendent of Government Printing,
1933.
Dicks:
D.R. Dicks. Geminus. In Dictionary of Scientific Biography,
edited by CC Gillespie. 5:344-347.
Diels, Frag.:
Hermann Diels. Die Fragmente der Vorsokratiker. Greek-
German edition; eighth edition; edited by W. Kranz.
Berlin: Weidmann, 1956.
Diels [1924]:
	.    Antike Technik, sieben Vortrdge. Leipzig and
Berlin: Teubner, 1924. [Reprinted in Osnabruch: Zeller,
1965.]
Dijksterhuis:
E.J. Dijksterhuis. Archimedes. Acta Historica Scientiarum
Naturalium et Medicinalium, of the Bibliotheca Universi-
tatis Hauniensis, no. 12. Copenhagen: E. Munksgaard,
1956.
Diog. Laert..
Diogenes Laertius. Lives of Eminent Philosophers. Edited
and translated by R.D. Hicks (Loeb Classical Library); 2
volumes. London and New York: W. Heinemann Ltd.
and Harvard University Press, 1925.
Dizer 8c Meyer:
Muammer Dizer and W. Meyer. The Celestial Globe of
Kandilli Made by JaTar ibn Dawlatshah al-Kirmani. Inter-
national Congress of the History and Philosophy of Science.
Proceedings of conference in Islamabad, 8-13 December
1979 [6 unnumbered pages].
Dorn [1830]:
Bernhard Dorn. Description of the Celestial Globe Be-
longing to Major-General Sir John Malcolm . . .   Depos-

316

SMITHSONIAN STUDIES IN HISTORY AND TECHNOLOGY



ited in the Museum of the Royal Asiatic Society of Great
Britain and Ireland. Transactions of the Royal Asiatic Society,
1830, 2:371-392,2 plates.
Dorn [1865]:
	. Drei in der kaiserlichen offentlichen Bib-
liothek zu St. Petersburg befindliche astronomische In-
strumente mit arabischen Inschriften. Memoires de I'Aca-
demie imperiale des sciences de St. Petersbourg, series 7, 1865,
9(1):1-150, 2 plates.
Dow:
Sterling Dow. Conventions in Editing. Greek, Roman and
Byzantine Scholarly Aids, No. 2. Durham: Duke University
Press, 1969.
Dozy:
R. Dozy. Supplement aux dictionnaires arabes. Two vol-
umes. Leiden: Brill, 1881 [Reprinted in Beirut: Librairie
du Liban, 1968.]
Drachmann:
A.G. Drachmann. The Plane Astrolabe and Anaphoric
Clock. Centaurus, 1954, 3:183-189.
Drechsler:
Adolph Drechsler. Der arabische Himmels-Globus angefer-
tigt 1279 zu Maragha von Muhammed bin Muwajid Elardhi
zugehorig dem Koniglichen mathematisch-physikalischen Salon
zu Dresden. Dresden: Koniglichen Hofbuchhandlung von
Hermann Burdach, 1873.
Du Ry:
Carel J. Du Ry. Art of Islam. New York: Harry N.
Abrams, Inc.. 1970.
Rt[n
The Encyclopaedia of Islam. First edition, edited by M. T.
Houtsma, A.}. Wensinck, et al.; 4 volumes. Leiden: Brill,
1911-1938.
EI [2]:
The Encyclopaedia of Islam. Second edition, edited by
H.A.R. Gibb, CE. Bosworth et al. Leiden: Brill, 1962-
present.
Eilers:
Wilhelm Eilers. Stern-Planet-Regenbogen: zur Nomenk-
lature der orientalischen Himmelskunde. In Der Orient in
der Forschung, Festschrift fiir Otto Spies, pp. 92-146. Wies-
baden: Harrassowitz, 1967.
Encyc. World Art:
Encyclopedia of World Art. Fifteen volumes; edited by
Ma.ssia Pallottino, et al. New York: McGraw-Hill, 1960.
Engelmann:
Max Engelmann. Sammlung Mensing: Altwissenschaftliche
Instrumente. Amsterdam: Frederik Muller and Co., 1924
[catalog].
Erat. Catast. [I]:
[Pseud.] Eratosthenes. Eratosthenis Catasterismorum. Ed-
ited by C Robert; Berlin: Weidmann, 1878; reprinted
1963.
Erat. Catast. [2]:
	.    Pseudo-Eratosthenis, Catasterismi. Edited by
Alexander Olivieri. In Mythographi Graeci, volume 3, fasc.
1. Leipzig: Teubner, 1902.

Erren:
M. Erren. Die phainomena des Aratos von Soloi: Untersu-
chungen zum Sach- und Sinnverstdndnis. Wiesbaden: F.
Steiner, 1967.
Ettinghausen::
R. Ettinghausen. Hilal In Encyclopaedia of Islam, second
edition, 3:381-385.
Fabricius:
Johann Albert Fabricius. Bibliotheca Graeca, sive notitia
scriptorum veterum graecorum, quorumcumque monumenta
integra, autfragmenta edita exstant: tum plerorumque e MSS.
ac deperditis. Second edition; 14 volumes. Hamburgi:
1712-1740. [Third edition; edited by Gottlieb Christo-
phor Harles; 12 volumes. Hamburgi and Bohn: C Lie-
bezeit, 1790-1809. Reprinted Hildesheim: G. Olms,
1966-1970.]
Falk, Smart, 8c Skelton:
Toby Falk, Ellen S. Smart, and Robert Skelton.    Indian
Painting. London and Bradford: P. Sc D. Colnaghi Sc Co.
Ltd., 1978 [exhibition catalog].
Fehervari:
Geza Fehervari:
Islamic Metalwork of the Eighth to the Fifteenth Century in the
Keir Collection; with a Foreword by Ralph Pinder-Wilson.
London: Faber and Faber, 1976.
Fiorini:
Matteo Fiorini. Sfere terrestri e celesti di autore italiano
oppurefatte o conservate in Italia. Rome: Presso La Societa
Geografica Italiana, 1899.
Fiorini & Gunther:
	.    Erd- und Himmelsgloben, ihre Geschichte und
Konstruktion. Edited by Sigmund Gunther. Leipzig: Teub-
ner, 1895.
Forbes:
P.B.R. Forbes. Crates (3). In Oxford Classical Dictionary,
first edition p. 240.
Fordyce:
C.J. Fordyce.    Catullus, a Commentary. Oxford: Oxford
University Press, 1961.
Frank:
Joseph  Frank.    Uber zwei astronomische arabische In-
strumente.    Zeitschrift   fiir    Instrumentenkunde.     1921,
41:193-200.
Galen UP:
Galen. Galeni, De usu partium libri XVII ad codicumfidem.
Edited by G. Helmreich; 2 volumes. Leipzig: Teubner,
1909; reprinted Amsterdam: Adolf Hakkert, 1968.
Galen UP [trns.]:
	.    Galen on the Usefulness of the Parts of the
Body. Translated, with introduction and commentary, by
Margaret Tallmadge May; 2 volumes. Ithaca, New York:
Cornell University Press, 1968.
Gascoigne:
B. Gascoigne. The Great Moghuls. New York and Lon-
don: Harper and Row, 1971.
Gennnus Elem. astr. [1]:
Geminus.    Gemini, Elementa astronomiae ad codicumfidem.

NUMBER 46

317



Edited, with German translation, by Carolus Manitius.
Leipzig: Teubner, 1898.
Geminus Elem. astr. [2]:
	.    Gemini, Elementorum astronomiae capita I,
III-VI, VIII-XVI. Edited by EJ. Dijksterhuis. Textes Mi-
nores, volume 22. Leiden: Brill, 1957.
Germ. Arat. [1]:
Germanicus Caesar.    Germanici Caesaris, Aratea cum scho-
liis. Edited by Alfred Breysig. Berlin:   1867; reprinted
Hildesheim: G. Olms, 1967.
Germ. Arat. [2]:
	.    Germanici, Arateorum quae supersunt. Edited
by E. Baehrens. In Poetae Latini Minores, 1:142-200.
Leipzig: Teubner, 1879.
Germ. Arat. [3]:
	.    The Aratus Ascribed to Germanicus Caesar.
Edited, with translation, introduction, and commentary,
by D.B. Gain. University of London Classical Studies, num-
ber 8. London: University of London, 1976.
Gibbs:
Sharon L. Gibbs.    Greek and Roman Sundials. New Haven
and London: Yale University Press, 1976.
Gibbs CCA:
Sharon Gibbs, Janice A. Henderson, and Derek de Solla
Price.    A Computerized Checklist of Astrolabes. Photo-
copied. New Haven, Connecticut: Yale University, De-
partment of the History of Science and Medicine, 1973.
Gillespie DSB:
CC. Gillespie, editor. Dictionary of Scientific Biography.
Sixteen volumes; New York: Scribner's for the American
Council of Learned Societies, 1970-1976.
Gingerich, et al:
Owen Gingerich, David King, and George Saliba.    The
"^Abd al-A^imma Astrolabe Forgeries. Journal for the His-
tory of Astronomy. 1972, 3:188-198.
Ghirshman, et al:
R. Ghirshman, V. Minorsky, and R. Sanghvi. Persia the
Immortal Kingdom. New York: New York Graphic Society,
1971.
Gleadow:
Rupert Gleadow.    The Origins of the Zodiac. New York:
Atheneum, 1969.
Gottschalk EI:
H.L. Gottschalk. Al-Kamil. In Encyclopaedia of Islam,
second edition. 4:520-521.
Gottschalk [1958]:
	.    Al-Malik al-Kamil von Egypten und seine Zeit.
Wiesbaden: Harrassowitz, 1958.
Griffin:
Lepel H. Griffin. The History of the Principle States of the
Punjab and their Political Relations with the British Govern-
ment {The Rajas of the Punjab). London: Trubner and Co.,
1873; reprinted Lahore: Sheikh Mubarak Ali, 1976.
Griffini:
E. Griffini. Intorno alle stazione lunari nell'astronomia
Aegharabx. Rivista degli Studi Orientali, 1907, 1:423-438.

Grotzsch:
Helmut Grotzsch. Dresden Mathematisch-Physikalische Sa-
lon. Leipzig: E. A. Seemann, 1978.
Gudeman:
Gudeman. Sporos. In Pauly and Wissowa, Real-Encyclo-
pddie der klassischen Altertumswissenschaft. zweite Reihe.
HI: columns 1879-1881.
Guide to the Adler Planetarium:
A Guide to the Adler Planetarium Operated by the Commis-
sioners of the Chicago Park District and the Adler Planetar-
ium. Chicago: 1970.
Guimbretiere & Hasan:
A. Guimbretiere and Mohibbul Hasan. KashmTr. In En-
cyclopaedia of Islam, second edition. 4:706-710.
Gundel:
G. Gundel. De stellarum appellatione et religione Romana.
Gissae: Impensis Alfredi Toeplemanni, 1907.
Gundel Sc Boker:
H. Gundel and R. Boker. Zodiakos. Der Tierkreis in der
antike Literatur und Kunst. In Pauly and Wissowa, Real-
Encyclopddie der klassischen Altertumswissenschaft. XA: col-
umns 462-707.
Gunther AW:
Robert T. Gunther. The Astrolabes of the World. Two
volumes. Oxford: Oxford University Press, 1932.
Gunther [1923]:
	.    Early Science in Oxford. Volume 2. Oxford:
Oxford University Press, 1923.
Guye & Michel:
Samuel Guye and Henri Michel. Time & Space: Measur-
ing Instruments from the Fifteenth to the Nineteenth Century.
Translated from the French by Diana Dolan. New York
and London: Praeger, 1971.
Halma [1821]:
N. Halma. Les Phenomenes d'Aratus de Soles, et de Ger-
manicus Cesar, avec les Scholies de Theon, les Catasterismes
d'Eratosthenes et la Sphere de Leontius traduit. Paris: Merlin,
1821.
Halma [1822]:
	.    Examen et explication du zodi