S M I T H S O X I A 4 N C O N T R I B U T I O N S T O B O T A K Y N U h l B E R 1 9
The Genus Thrinax
(Falmae: Coryphoideae)
Robert W. Read
SMITHSONIXN INSTITUTION PRESS
City of IYashington
1975
A B S T R A C T
Read, Robert W. The Genus Thrinax (Palmae: Coryphoideae). Smithsonian
ContriOutions to Botany, number 19, 98 pages, frontispiece, 57 figures, 1975.-The
history, distribution, ecology, morphology, cytology, anatomy, and taxonolny of
the four species of Thrinax are presented. The objective is to demonstrate that
quantitative characters, formerly believed to be important in distinguishing
species, merely represent random selections from clinal patterns and are therefore
of little value in distinguishing taxa. Certain characters described for the first
time-leaf sheath, blade, color and puberulence of the inflorescence-have proved
to be of great taxonomic value. Particular attention is given to tlie variability
of Thrinax parvifiora Sw. over a wide range of climatic conditions and to its
unusual phenotypic behavior, in special situations, in the Cockpit Country and
on the slopes of Mt. Diablo in Jamaica. The haploid chromosome number (determined
at pollen-tube mitosis) of all four species is 18. Little difference in size
or morphology of the chromosomes exists among these taxa. Anatomically all
four species can be distinguished by comparison of leaf segment sections.
The taxonomic portion is a thorough revision of the genus. It includes literature
references, synonymy, complete descriptions, and specimens examined.
Thrinax parvipora Sw. and T. excelsa Lodd. ex Griseb. are endemic to Jamaica
while T. mdinta Lodd. ex J. A. & J. H. Schult. occurs in the littoral of the Greater
Antilles (except Puerto Rico), tlie Bahamas, Florida, and Mexico. Th1-inax
morrisii H. Wendl. occurs on most of the islands from Anegada (east of the
Virgin Islands) to the islands off the coast of British Honduras, and Florida, but
not in Jamaica.
OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recorded
in the Institutions annual report, Smithsonian Year. SI PRESS NUMHFR 5125. SI.RIES COWR n m m :
Leaf clearing from the Katsura tree Cercidiphyllurn jnponicunz Siebold and Znccarini.
Library of Congress Cataloging in Publication Data
Read, Robert L$’., 1931-
The genus Thrinax (Palmae: Coryphoideae)
(Smithsonian contributions to botany, no. 19)
1. Thrinax , I. Title. 11. Series: Smithsonian Institution. Smithsonian contributions to botany,
QKl .S274’i no. 19 [QK495.P17] 581’.08s [584‘5] 74-6489
no. 19.
For sale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C. 20402 Price $2.40 (paper coler)
Contents
Page
Introduction 1
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
History of the Genus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Distribution and Ecology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Thyinax radiata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Thrinax excelsa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Thyinax mowisii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Thrinax parviflora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Morphology 17 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Selection of Samples for Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Caudex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Leaf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Inflorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Germination of Seeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Pollen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Anatomical Key to the Species of Thyinax Using Laminar Characters . . . 41
Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Lamina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Material and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
48
52
59
Cytology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Phytomorphosis and Zoomorphosis in Thrinax parviflora . . . . . . . . . . . . . . .
Breeding System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Variation of Thrinax pai-viflo~.n., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Transect from Portland Ridge to Hollymount, Jamaica . . . . . . . . . . . . .
Donkey Trail Transect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Thyinax Swartz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Key to the Species of Thl-inax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Typification of Thrinax ParuifEora . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Diagnostic Key to the Subspecies of Thrinax parvipora . . . . . . . . . . . .
68
1 Thyinax paruiflora 69
70
74
75
76
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Thrinax parvifiora Swartz subsp . parvipom . . . . . . . . . . . . . . . . . . . . .
Th?~inax pat-uiflora Swartz pziberula, new subspecies . . . . . . . . . . . . . . . .
2 . Thrinax radiata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3 . Thrinax excelsa 84
4 . Thrinax monisii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Appendix 11: Nomina Incerta et Dubia . . . . . . . . . . . . . . . . . . . . . . . . . 95
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix I: Binomials Published in the Genus Thrinax . . . . . . . . . . . . . . 94
Literature Cited . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . 96
... 111

The Genus Thrinax
(Palmae: Coryphoideae)
Robert W. Read
Introduction
Thrinax is a genus of true palmate-leaved palms
in the subfamily Coryphoideae. The genus occurs
only on the islands of the northern Caribbean and
the nearby continental coastal areas bordering on
the Gulf Stream. Its nearest allies are Hemitlzrinax
and Coccothrinax, which share the same general
range of distribution.
The genus Thrinax may be recognized by the
unarmed split leaf-petioles, flowers with a single
whorl of perianth, and nonruininate seeds which
have the center partially or completely intruded by
the testa. Stamen number varies from 5 to 15
(rarely more) and the abaxial surface of the leaf
may or may not be white or silvery. Fully mature
fruits are always white or ivory colored.
At the beginning of this study five species of
Thrinax were attributed to Jamaica with an additional
seven or eight distributed throughout the
Caribbean region apart from Jamaica. With the
exception of a few quantitative Characters of dubious
value, no convincing differences could be
found between Beccari’s T. parviflora, T. tessellata,
and T. harrisiana. This was also true concerning
T. mowisii of FVendland, T. ponceana and T. psaeceps
of Cook, T. drtidei and T. puncttilata of
Beccari, and T. miclocarpa ol‘ Sargent. Similarly,
Bailey’s T. parvifloi.a and T. excelsa were separated
in a key based solely on the size of the fruit and
leaf segments. During the present study it has been
found that, with the exception of Beccari’s T. parviflora
(in part), none of the names in common use
Robert Tt’. Read, Department of Botany, Smithsonia?~ I t i s t i -
tution, Washington, D.C. 20560.
1
have been applied correctly, and all of the names
of the species not occurring in Jamaica are to be
relegated to synonymy or to subspecific status of
the least commonly used specific epithet. The reasons
for the previous confusion are inadequate
herbarium material and a lack of understanding
of the variability of Thrinax species in natural
populations.
My interest in Thrinax developed as a result
of an attempt to understand the palms of southern
Florida, and the necessity of clearing up nomenclatural
problems of the plants used in my cytological
studies. After several trips to the islands of the
Caribbean, and to Jamaica in particular, it was
soon obvious that the same taxon known in Florida
as T. parviflora was known under a different specific
epithet on each of the various islands. In
Jamaica, the type locality of T. parzri%ora Sw., the
Florida taxon was known as T. excelsa and the
epithet T. parviflora was applied to an entirely
distinct but very poorly understood local taxon. It
was clear that a study of the Jamaican taxa of
Thrinax was essential before an understanding of
the taxa elsewhere was possible.
Before revisionary work in palms could be accomplished
at the species level, it was essential to
understand the range and type of variation possible
within each taxon and to have some understanding
of the effects of environment on the
variation. It was also important to determine the
kinds of characters (key characters) which unquestionably
distinguish taxa and which remain recognizable
regardless of environment. Far too often
descriptions of palms drawn from incomplete and
and inadequate collections have been relied upon
2
for identification. The situation clearly stated by
Joseph D. Hooker and Thomas Tlionison (1855)
has been found true today in the study ol‘ palms.
They wrote that “the niore specimens we examined,
and especially if taken from different indivicluals,
the greater the difficulty in framing diagnosis. . . .
iVe are firmly convinced that no species can be
properly defined, until it has been examined in all
the variations induced by those differences in climate,
locality, and soil.’’ It seems superfluous to
restate these well-knoTvn facts, but this basic information
lias yet to be acquired in most tropical
families, particularly those with large cumbersome
plants such as the Palmae.
No amount of statistical analysis, numerical
taxonomy, or classical taxonomy for that matter,
Trill solve the problems of species definition when
based on inadequate collections, incomplete herbarium
specimens, and insufficient knowledge of ecological
tolerances. A study of tlie scope undertaken
here is too laborious and time-consuming to be
applied Fvithin the foreseeable future to niore than
a fraction of the palm species described. Proof of
conclusions regarding genetic differences is not
possible without transplant experiments; but the
obstacles to long-term experiments ivitli such slolvly
developing plants as paliiis preclude this approach.
Proper study of palms lias until recently also
been hampered because ol their size, complexity,
general unuielcliness, and inaccessibility in tropical
regions. In addition, portions of the plants \vliicli
did not conl‘orni to the limits set by the standard
herbarium sheet or “pigeonholes” were often tlisregarded
by collectors. The opportunity ~ v a 5 taken
during this study to make intensive observations of
Tkiirznx in tlie field and to analyze antl present the
data in a comparative Tvay for taxonomic purposes.
This has led to detailed examination of the development
and mature condition 01 all organs Tvliose
nature could in any Tray assist the accurate antl
complete definition of the genus Tlzrinnx antl its
species in Jamaica.
The result is a statement of tlie natural variability
of Tlz~iiznx against the background ol the pliysical
factors 01 tlie environment, leading to a
complete taxonomic revision of the genus.
. C K ~ . O ~ ~ L E L ) G ~ l T ~ T s . - T h e basis of the present
monograph is a thesis for the P1i.D. degree. The
author rrishes to express his sincere appreciation to
5hlITHSO\I4\ COATRIBLTIOXS TO ZOTANY
Dr. C. Dennis Adams of the Botany Department
of tlie University of the ’1Vest Indies for his valuable
advice, guidance, and patience during the
investigations and preparation of tlie original manuscript;
and to Dr. Harold E. Moore of the Liberty
Hytle Bailey Hortorium, Cornell University, who
has contributed immeasurably by providing information,
suggestions, and reading tlie manuscript.
Very special thanks are extended to the Botany
Department of tlie University of the West Indies,
and to George R. Proctor of the Institute of
Jamaica, for providing transportation, assistance,
and facilities during the studies in Jamaica; and
especially to Dr. and Nrs. Ivor Cornman for their
contribution to the success of many trips into the
“bush.”
Sincere appreciation is also extended to tlie directors
and staff of numerous herbaria and libraries
for photographs, loan or use of herbarium material
and freely given information, assistance, and encouragement;
especially Dr. William Dress of Corriel1
University; Dr. George Bunting, formerly at
Cornell University; Dr. Richard A. Howard, Director
of tlie Arnold Arboretum; Drs. Lyman B. Smith
and Velva Rudd of the National hluseuni of Natural
History, Smithsonian Institution; Dr. Howard
Irwin of the New York Botanical Garden; Mr.
George W. Rosner, Richter Library, University of
Miami, Florida; Dr. T.V. T. Stearn of the British
hIuseuiii (Natural History); hlr. ’Iy. hlilne-Redhead
of Kew Royal Botanic Gardens; Dr. C. E. B. Bonner
of the Conservatoire et Jardiii botaniques,
Geneve; Prof. Tycho Norlindh and Barron Sparre
of the Naturhistoriska Riksniuseum, Stockholm;
Drs. H. AIerxmuller and A%. Schreiber of the Botanische
Staatssammlung, hlunich; antl to hlr.
Harold Loomis, former director of the United
States Plant Introduction Station in Xliami, Florida,
lor his early encouragement and knowledge of the
palms at the time I was starting to get interested
in the family.
Additional research ~ v a s made possible for me as
a National Research Council Visiting Research
Associate in the Department of Botany, Smithsonian
Institution, later as a Research Associate of
tlie Department, and now as an Associate Curator
in the same Institution. This has made it possible
to change “A Study of the Genus Thiinnx in Jamaica”
into a monograph of the genus as a whole.
AUXIBER 19 3
History of the Genus
Although Linnaeus did not include any palms
of the Caribbean islands in his Species Planlarum,
their existence had been recorded by various authors
during the century prior to 1753. In tact Sir
Hans Sloane (1696) described ten true palms in his
Catalogiis Plantarurn, of which one has been determined
to be a member of the genus Thrinax. It is
important to understand that while Sloane ~ v a s
writing specifically of palms he observed in Jamaica,
he also attempted to identify them with descriptions
already in the literature. This association of
earlier accounts Tvith the Jamaican palms n’as often
incorrect; therefore in order to understand the
application of Sloane’s descriptions, it is essential
to ignore these misleading synonymies. Nor did
that practice end with Sloane for it has continued
to plague the study of palms to the present.
The palm to be discussed most fully in the
present paper appeared for the first time in the
literature as the last true palm mentioned by
Sloane (1696: 178, 179). Sloane unfortunately associated
the Jamaican “thatch” with Ray’s “4. Palnin
Biasiliensis p ~ z i n i f e ~ n ” (1688: 1368 ) because of the
similarity of the leaf form. Ray’s palm, liolvever, is
Brazilian antl has been determined to belong to the
genus Copemicia (Noore, 1963). Since no Cofiernick
has been found in Jamaica, the plicate-leaved
palm growing “In sylvis saxosis k collibus sterilioribus
Jamaicae Insulae ubiyue reperitur” (Sloane
1696: 179) can only be one of those presently knoivn
as Thi.innx or Coccotkiinox. In 1525, Sloane again
cited “Palma Bmsiliensis” as the “thatch” (1). 121)
of Jamaica and again the sole indication of its
identity, in the text, is that, “This grows on all tlie
Honey-comb Rocks of this island” (p. 122). Sloane’s
pl. 213: fig. 2 of tlie same lvork, holvever, unquestionably
illustrates a Thrinax-like palm leaf. The
lack of teeth on the petiole, the nonacuminate
hastula, and the irregular fusion of the segments,
it‘ this xvas intended to be diagnostic, clearly intlicate
that this illustration depicts the plant no~v
determined to be Thrinax pawij7oi.a Sw.
The first complete account of a palm readily
identifiable as Th~inax parvij7orn Sw. in the literature
appeared in Patrick Browne’s Civil ai7d X n t -
ziral Histoyy of Jamaica, published in 1756. The
Latin diagnosis of “Coryphn I. Palniacen” 1 is accurate,
but the association of Kay’s Brazilian palm
again apparently contributed to Grisebacli’s (1 864)
belief that a Copemicia was to be found in Jamaica.
In 1‘784, Olof Swartz left Sweden at the age of
24 and spent tlie next two years traveling through
the West Indies and Jamaica. Follolving his trip
through the Caribbean, Swartz \vent to England to
study the Hans Sloane Jamaican collection and the
herbarium of Sir Joseph Banks at the British Museum.
Within two years of his return to Siveden he
published (1788) the first of his major works 011
LVest Indian plants, Souti Genera et Species Plantal-
unz sell Prodronzzis. In the Prodromus, as it is
known, appeared the first published description of
the palm Thi.inax parviflo~a. A diagnosis of the
genus appeared on page 4, number 29, and tlie
single species was described on page 57.
It is possible that Swartz went to Jamaica with
some foreknorvledge of the palm lie later described
as Tlzrinax pawifiora. Although he gave no indication
in the P?.odromiis of the source of the name
or its meaning, he attempted to rectify his debt to
Linnaeus the younger by inserting a note ol‘ acknowledgment
in his Flo1.a (1797). That the younger
Linnaeus was preparing to publish the binomial,
before his death in 1783, is now an establishetl
fact. =\ complete and accurate description of T h ~ i -
nax, under the binomial T. pat~viflo~a, has been
found in an unpublished manuscript belonging to
the younger Linnaeus. The manuscript reproduced
in part in Figure 2 is clearly the source ol
the binomial, antl whether Sivartz saw the manuscript
before he ivent to Jamaica or upon his return
is not known. Linnaeus’ description is a
composite of Browne’s Coqpha I., Sloane’s illustration
(pl. 213) and a specimen of an inflorescence
preserved in the Banksian herbarium at the British
Museum. Since the Linnaean manuscript was never
published, we must consider only Swartz’s speci-
Correspondence between Dr. Bailey and the Linnean
Society of London in 1937, in a lcttcr signed by 5. Savage,
Assistant Secretary, revealed that “no specimen of the species
of Corjl~ha described by Patrick Browne in Hist. Jni?!. p. 190
exists in the Linnaean Herbarium. In Linnaeus’ copy of
Browne’s book there is no annotation by Linnaeus to show
that a specimen of this plant was received by him with
Browne’s herbarium.”
’ Proritled by the courtesy of the Linnean Society, London,
through Dr. H. E. Moore of Cornell I*niversity (List no. 40,
“ I h n. /i 1. Palmarti in De icript iones”).
4 SMITHSOSIAN COSTRIBUTIONS 10 BOTAKY
FIGURE I.-Tlfrinas pnruipotn, cut and removed froin the forest for photographing, being
examined 117 Dr. Itor Cornman.
KUMBER 19 5
men and his descriptions based on personal experience
in Jamaica in typifying Thrinax pawiflova.
The finding of the Linnaean manuscript is, however,
of historical significance and provides answers concerning
the source of the binomial and its meaning.
Thrinax according to the manuscript is derived
from the Greek 6giva5, which can be translated as
“trident” or as indicated by Linnaeus “ventilabrum”
meaning, “winnowing fork.”
Thrinax was the tenth genus described in the
Palmae after the publication of Linnaeus’ Species
Plantarum in 1753. It should not then be surprising
that a number of dissimilar and unrelated
palms have been included in the genus (see Appendix
I). The original description by Swartz in the
Prodromus (1788) clearly sets down characters that
are still diagnostic for the genus. The six-lobed
perianth cupule, unilocular ovary, and infundibuliform
stigma apply exclusively to the Thrinnx
alliance.
Early in 1789, W. Aiton reported in his Hortits
Kewensis that there was a palm growing at Kew
under the name of Thyinax parvipoia, which was
introduced by Dr. William Wright upon his return
FIGURE P.-Facsimile of a portion of Linnaeus the younger’s unpitblished manuscript on
Thrinax paruipora.
6 SMITHSOSIAK CONTRIBUTIONS TO BOTANY
from Jamaica in 1777. The palm at Kew ~ i a s mentioned
again by IV. T. Aiton in 1811, and again
seventy-one years later by Hooker in the Kew Report
for 1882. Thiinax was treated in the eighth edition
of Linnaeus’ Genera Plantarum (edited by C. D.
Schreber, 1791), and T. parviflora was treated later
in the fourth edition of the Species Plantamm,
edited by C. L. Willdenow (1799).
In his magnum opus on West Indian plants,
Swartz (1797:613-615) published a much amplified
account of the genus and species in the Flora Indiae
Occidentalis; flowers and fruit were illustrated on
plate 13.
The first mention of additional species in the
genus Thrinax appeared in 1829 in Desfontaines’
Catalogus Plantariim (3rd ed.) where T. radiata
and T. argenten appeared in a list without descriptions.
The father and son team of J. A. and J. H.
Schultes (1830) took up the Desfontaines names for
which, with the addition of T. piimilio, they published
diagnoses based on specimens sent by Loddiges,
a nurseryman, to the botanical garden at
Munich. hlartius (1838) elaborated on tlie clescriptions
of T. parviflora, T. radiata, and T. al-gcntea
and added two new species, T. miiltiffol-a and T.
barbadensis. In 1845, nurseryman Loddiges, froiii
whom Alartius also obtained several specific epithets,
published a catalog of palms; there are no
descriptions, thus the names that Tiere not already
provided with diagnoses are nomina nuda. During
the one hundred years between treatments by
Rlartius and Bailey (1838 and 1938 respectively)
forty additional epithets were attributed to the
genus Thrinax.
In 1866, A. H. Grisebach attempted to erect a
new genus, Porothrinax, based on lh1inax piimilio
of J. A. & J. H. Schultes. The lack of a separate
diagnosis or description for a monotypic genus
based on an old species does not satisfy the requirements
of the present International Code of
Botanical Nomenclature (see Appendix 11), so I
am treating the genus as a nomen nudum.
Charles S. Sargent (1899:88) made tlie first clear
division of the genus when he recognized that specimens
collected in Florida, and attributed to
Thrinax, could be separated from that genus on
the basis of the seeds being “vertically sulcate by
the infolding of the testa into the ruminate albumin.”
Thus he described Coccothrinax, a new genus
based on C. jiiczinda Sargent. He also transferred
Thrinux garberi Chapman into the genus to make
a second species, C. garberi (Chap.) Sarg. He further
suggested that Thrinax radiata and T. argentea
both of Lodd. ex. J. A. & J. H. Schultes might
also belong in the newly described genus, but it
remained for K. Schuinann to make the formal
transfers in 1901. Sargent (1899) divided the species
remaining in Thrinax into two sections, the Peditellatae
and Sessiliflol-ae, and described two new
species, T. floridana and T. keyelzsis.
In 1883, Bentham and Hooker included Thrinax
along with their newly founded Hemit hl-inax in
the tribe Corypheae, but Drude (1887) reduced
Hemithrinax and Porothi.iizax along with Euthl-inax
as subgenera of Thi.inax. Thus it remained
until Beccari (1907) produced his work on American
Coryphoid palms. Beccari raised Hemit hrinax
to full generic status again and divided the Corypheae
into two subtribes: Eucorypheae, characterized
by the perianth having two whorls and the
ovary being three-locular or with three carpels; and
Thrinaceae, with a periantli of a single six-lobed
whorl and the ovary consisting of a single uniovulate
carpel. He divided the genus Thrinax into two
subgenera based on the degree of intrusion of the
testa into the seed: Ezithrinax, including the earlier
Porothrinax of uncertain identity, and Typh10-
thririnx. No mention was made of Sargent’s earlier
division of the genus based on pedicels. The genus
as outlined by Beccari (1907) contained ten species
that he considered good, including four new names,
plus three names of dubious application. Beccari
(1912) published a series of articles on the palms
of Cuba in which he treated three species of
Thrinax. A supplement to his earlier work, in
English, was included in a larger posthumously
published xiork in 1933 [“1931”].
’1 new genus, Sitnpsonia, based on Thyinax microcarpa
Sarg., it-as published by Cook (1937); however,
the requirements of Article 36 of the International
Code of Botanical hTomenclatzil-e, 1966,
were not satisfied and the genus was not validly
published.
Bailey (1938), in an attempt to clear up some
identification problems among palms in horticulture,
produced a comprehensive study on the genus
Thrinax. He accepted ten species, which he arranged
in two subgenera as suggested earlier b~
Sargent. In a review of Beccari’s work Britton
wrote (1908:240) that, “This is probably the most
SUhiBER 19 7
difficult of the American palm genera to untlerstand,”
to which Ba.iley (1938: 133) replied thirty
years later:
Any subsequent student of the genus will be sympathetic
with this estimate; but many of the difficulties banish if
one knows the plants in the field, has really good herbarium
material, and does not assume he must necessarily recognize
all the species that have been described. Differences one
draws diligently from poor herbarium sheets may disappear
in the living plant.
The present author must agree with this, antl with
Bailey’s earlier statement (1938: 133) concerning
Thyinax paixipoi,a Sw. and T. nzicrocai,pa Sarg.
that, “If the student understands these two, all tlie
other species fall readily into place.” Unfortunately
Bailey did not have an adequate understantling
of T. payvifloya in Jamaica, hence the species
there did not fall so readily into their proper
places.
Distribution and Ecology
Palms of the genus Thi,inax occur only on alkaline
substrate, specifically coralline sands or limestone
outcrops, from sea level to about 1200 meters
elevation. They are distributed from areas of strong
seasonal drought with rainfall averaging 100-1 12
cm per year (with no more than 6 months having
rainfall less than 8 cm per month) to regions where
the rainfall may reach 762 cm per year. Two species
endemic to Jamaica-Thyinax pni-uipol-a antl T.
excelsa-are restricted entirely to exceedingly welldrained
sites far from the influence of salt spray.
Thyinax radiata, in addition to Jamaica, is widely
distributed in tlie littoral throughout the northern
Caribbean and nearby shores of the Gulf Stream,
and tolerates considerable exposure to salt-laden
winds. The fourth species, T. mowisii, not found in
Jamaica, is fairly tolerant of salt and, although
usually at sites similar to tlie aforementioned, is
commonly found farther inland beyond the worst
effects of salt spray.
Each of the three species in Jamaica is confined
to one of three major environmental situations (see
Asprey, 1959): e.g., Dry Evergreen TVoodlantl or
Thicket, Lower Rlontane Rain Forest, and Littoral
Woodlantl or Thicket. Species of Thyinax have not
been found naturally in areas lacking limestone or
alkaline sand, they are strict calciphils.
-rHKINAX RADIhTA.-one Of the tWO most widely
distributed species in the genus, T. ~ a d i a t n , is found
only in coastal areas on limestone and alkaline
sand, under tlie influence of salt-laden breezes,
throughout the Greater Antilles (excluding Puerto
Rico). It is also found on many of the smaller
islands to the north and wesr such as the Cayman
Islands and tlie cays off the coast of Cuba and
British Honduras, the Yucatan Peninsula of Xlexico,
the Bahama Islands, and southern Florida. In
nature this palm is confined to saline conditions of
the littoral but where there is adequate rainfall
(127-254 cm per year) throughout the year. This
is also the most commonly cultivated species and is
grown under various conditions of soil and moisture
far from the influence of salt.
In Jamaica, Thyinax yadiclta may occur in dense
stands but is more coinmon in small scattered populations
along most of the coastline from extreme
eastern Morant Point along the north coast west
to Negril lighthouse and south to Little Bay. A
small population also occurs at Long Acre Point,
near Black River, and another is believed to be
on Big Pelican Cay some distance off the southern
coast of Jamaica.
In 1958, G. F. Asprey and A. R. Loveless described
a new addition to their list of Littoral Dry
Evergreen formations in Jamaica based on a population
of Thi inax 1-ndinta situated at Galina Point
on tlie north-central coast in the parish of St.
Mary. This formation, which they called “Littoral
Palm thicket,” occurs on a raised coralline limestone
shelf, 3.05 to 7.62 in (10 to 25 ft) above sea
level. The shelf is relatively flat, devoid of soil, and
is exposed to the full force of strong salt-laden
trade winds. The sea is frequently rough and
breaks over tlie low cliffs to produce a splash zone
up to 12 meters broad. The species forms a characteristic
feature of the littoral hedge, becoming
increasingly plentiful on the landward side, “where
it would appear to take the place of Littoral Evergreen
bushland, but no explanation as to why this
occurs can be offered” (p. 560). The authors further
pointed out (p. 560) that
Asprey and Loveless (1958) failed to distinguish between
T. pnrzli/7orn Sic.., a conspicuous element in their studies on
the Dry Evergreen thicket of Portland Ridge on the south
coast, and the \ery different T. rndintn at Galina Point.
SMITHSONIAN COSTRIBUTIOSS TO BOTANY
The raised limestone beach is dominated by Thrinns pnrviflora
[sic] to the exclusion of other plants. The trces grow
out of fissures in the bare soilless rock and vary in height
from 73-12! m (25-40 ft). The) are spaced, on the average,
about 1.2 m (4 ft) apart and form a closed canopy. The
palm reproduces actively since seedlings abound and thc
plant is evidently in a most favourable environment.
4 somewhat similar situation occurs at Morant
Point near the lighthouse where the palms form
small stands in the thicket along the windward
coast. Throughout the rest of its range in Jamaica,
T. mdintn occurs as scattered individuals in coastal
thickets within the range of salt spray. The dense
stand at Galina Point is not unique, for of Florida’s
upper Keys, J. K. Small (1925:53) wrote: “There
the groLvth is phenomenal, or at least it was, before
the lust for destroying every living thing reached
these islands with the advent of civilized man.
Thatch hammocks clothed large and small areas.”
In some small low areas and on outlying ke)s the
palms still dominate the landscape.
In the paper by Xsprey and Loveless, a site at
Fort Point on the eastern edge of Discovery Bay
FIGURE 3.-The Cockpit County of western Jamaica, in the vicinity of Barbecue Bottom.
NUMBER 19 9
(Dry Harbour) was also described. Fort Point is
approximately 36 miles west from Galina Point.
Meteorological data used for both localities were
based on records for Richmond Llandovery Sugar
Estate, which is nearer to Fort Point than to Galina
Point. The rainfall figures quoted were 55 to 70
inches (139.7-177.8 cm) per year which, when compared
with evaporation figures over the same period
of seventy years, presents a condition where there
are only 5 months in an average year when precipitation
exceeds evaporation. It must be realized,
however, that the utilization of such figures is
indeed tenuous in a region where abrupt changes
can occur in very short distances depending on the
exposure and topography. According to the most
recent rainfall figures, Galina Point, at the western
edge of one of the wettest regions of the Island,
has between 60 and 80 inches (152.4 and 203.2 cm)
of rain per year, while Fort Point is in a region of
considerably lower rainfall not exceeding 50 inches
(127 cm) per year.
The preceding observations are brought out be-
FIGURE 4.-Thrinax radiata, base of the caudex and low mound of surface roots, growing on
exposed and eroded limestone in the Florida Keys.
10 SAlITHSONIz4N CONTRIBUTIOSS TO BOTANY
cause even as Asprey and Loveless noted (1958:
561): “The only major anomaly in the data is
brought about by the megapliyllous palm Thiinax
parviporu Sw. [sic].” Had they recognized that the
palm from Galina Point was distinct from the palm
in their Portland Ridge studies, a very important
fact would have been realized and quite different
conclusions might have been reached. Characters
for separating the two species are in the key to the
species in the taxonomic section: however, a very
important characteristic, not in the key, is the fact
that T. radiata is a true halophyte and T. parviflora
is not. Further, T. mdiata occurs only where
regular deposition of salt is accompanied by regular
high rainfall. Thiinax paiviflora on the other
hand occurs not at all where saline conditions prevail
but can survive very dry environments. The
present author agrees with the observation made
by Asprey and Loveless (1958:562) “that the vegetation
reflects the climate far more accurately than
any meteorological data.”
It seems appropriate here to point out a few very
important differences between the two localities
studied by Asprey and Loveless on the north coast
of Jamaica, which bear on the ecology of Thrinux
radiata. Galina Point is exposed directly to the full
impact of strong trade winds and heavy seas. Ample
rainfall is indicated by its proximity to the highrainfall
zone of Portland and by the character of
the vegetation in the immediate vicinity beyond
the reach of the salt spray where herbaceous ground
cover is plentiful. Asprey and Loveless noted (1958:
561) that, “The presence of Philodendron sp.
[ Iacertim] among the fringing woodland plants
indicates a moist habitat,” from which they concluded
that “the relationship of Palm thicket to
rainfall distribution might repay further investigation.”
A further observation by the authors (1958:
562):
In view of the fact that the rainfall along the stretch of
coastline in which Fort Point and Galena Point are situated
is more than twice as much as at Port Henderson, although
the monthly distribution patterns are similar, it might at
first sight seem surprising that Evergreen bushland is found
at both centres.
FIGURE 5.--Thrinax radiata in Jamaica: A, near the lighthouse at Morant Point; B, plants exposed
along the limestone shelf at Galina Point.
SUhIBER 19 11
This is not surprising, for they did not in fact find
bushland at Galina Point, it was displaced by palm
thicket, and the bushland described at Fort Point
is in a much drier area than Richmond Llandovery,
where the rainfall records were made. The authors
relied too much on interpolation of rainfall data
and included too extensive a stretch of north coast
in their data. Furthermore the additional effects
of strong constant winds in conjunction with salt
would produce drought conditions in the spray
zone more intense than might be indicated by rainfall
figures, no matter how accurate.
A short distance away, to the west of Discovery
Bay in the vicinity of the Queen’s Highway,
Thrinax radiata occurs under what appear to be
pessimum conditions for the species. Several widely
spaced individuals are growing on the seaward
side of the highway on a flat limestone terrace
where there is an accumulation of sand and decaying
vegetation. Evidence of low rainfall can be
seen in the abundant xerophytic vegetation containing
plants such as Agave hart isii, Eztphot bin
punicea, and Plitrneria jamnicensis, which become
increasingly abundant on bare rock toward the east
in a thicket of semideciduous trees and shrubs.
Coccothrinax jamaicensis, a palm otherwise limited
to the drier thicket vegetation along the south coast
(Read, 1966), occurs in abundance along the eastern
portion of the Queen’s Highway. To.ivarc1 the western
end of the Highway, halfway between Discovery
Bay and Bengal Bridge, the climate changes
rapidly to a more moist environment containing
none of the species mentioned as characterizing
the dry thicket toward the east. A single plant of
T. pat-viflora stands out conspicuously in the ecotone
on the inland side of the highway. On tlie
seaward side of the highway Coccothrinax persists
farther west among the plants of T. 7aclintc2, .\\.here
it appears to tolerate increased moisture under
saline conditions. It has been the author’s experience
with palms in cultivation that many species
which occur naturally in a very dry climate also
have a high tolerance for saline conditions; e.g.,
species of Waslzingtonia, Erythen, Hyphnene, Phocnix.
That is not to say that the converse is also
true, for it is not.
At the westernmost end of Jamaica between
Orange Bay and Negril there occurs a variety of
habitats in a region of fairly regular, often high
rainfall. On Orange Point, Thi inax radinta thrives
on a limestone ridge a few feet removed from the
weatherit orn, absolutely soilless, clinker-type dogtooth
limestone. Farther south along the shores of
Bloody Bay and inland from Kegril Beach the land
is low, almost flat but slopes imperceptibly into
the nearby Great Morass. The substrate is white
sand, with a very high water table, and T h r ~ n n x
tcidzata occurs as scattered individuals along tlie
better chained rise, only a few feet above sea level.
The species is fairly common along the high cliffs
near the Kegril lighthouse and occurs as scattered
individuals southeastward to Little Hay and Long
Acre Point.
THRINAX ExCELsA.-This is a Jamaican endemic
which is iery restricted in its distribution, occurring
only in the region of the John Crow Mountains
01 the easternmost parishes of Portland and
St. Thomas, and on Uncommon Hill, a limestone
outcrop on the lower slopes of the Blue Mountains
above Fruitful Vale in Portland. Its environmental
requirements are rather unusual in that it is limited
to the driest, best drained, most exposed situations
in the wettest part of Jamaica. According to
Asprey and Robbins (1953), the John Crow Mountains
are clothed with lower montane rain forest
on the lower slopes, Mhere the rainfall may be as
much as 762 cm per year.
That the vegetation is very characteristic of the
substrate and drainage is strikingly illustrated by
the situation in the John Crow AIountains. In the
vicinity of TVinchester Peak at the southern end
of the range, the deep heavy soil on the lower
slopes supports a tall forest with a dense canopy.
In the relatively low illumination, epiphytes are
rare and the undergrowth consists mainlj of seedlings,
some ferns, and a few herbaceous plants.
Higher on the steep slopes below the peak, the
well-drainetl substrate consists of a mixture of soil,
decajing matter, and loose broken rock. The forest
here is lower and more open with smaller more
slender trees but with occasional large emergent
trees. Ferns and herbaceous plants are more common
in the increased illumination and abundant
mcisture. At elevations between 304 and 762 m on
the slopes and ridges throughout the range the
foundation rock is frequently exposed, particularly
along water courses. It is peculiar that in such a
wet environment the exposed rock is usually devoid
of vegetation other than the palm Thrinax exrelsn,
which dominates the forest on nearly all exposed
12 SMITHSOSIAN COSTRIBUTIONS TO BOTANY
solid rock. Winchester Peak, at the southern end
of the John Crows, receives less rain than elsewhere
in the range. The southeast face of this peak
is dominated by a palm brake of T. excelsa at between
457.2 and 533.4 m elevation.
Throughout the rest of the range of the John
Crow Mountains the forest is considerably more
dense and more typical of lower montane rain
forest, than that just described, because of the
greater rainfall. On Uncommon Hill in the northeastern
foothills of the Blue Mountains and in the
John Crows, where the rainfall is consistently high,
T. excelsa occurs only on rock outcrops at the
higher elevations. An exception was more recently
FIGURE 6.--Thrinax morrisii in the pinelands of New Providence in the Bahamas. This species
is found in almost identical associations on Big Pine Key in Florida.
KUXIBER 19 13
found on a rocky hill near the mouth of the Rio
Grande.
THRINAX MORRISII.-~~ith much the same range
as Thrinax radiata, this species extends farther
east to the Virgin Islands and Anguilla, but it is
conspicuous by its absence on both Hispaniola and
Jamaica; however, the tiny island of Navassa, situated
immediately between tlie two much larger
islands, supports a small population of T. mol-risii.
In Puerto Rico and Cuba the species is common on
limestone hills of the interior where it occupies
habitats apparently paralleling those of T. parviflora
in Jamaica. Occupying a niche which seems
to be intermediate between T. pai.niflol-a and T.
~adiata, the Morris palm apparently cannot compete
with either. Although somewhat tolerant of
saline conditions, T. morrisii does not readily compete
with T. radiata except on the smaller islands
where annual rainfall probably falls below the
minimum necessary for survival of the latter. Intermittent
periods of rigorous drought may also restrict
competition from T. mdiata. Thrinax morrisii
exhibits a very wide range of tolerance for
extremes of climatic conditions.
As might be expected from a species that has
two synonyms in Puerto Rico, two in Cuba, and
two more in the Florida Keys, Thrinax mowisii is
morphologically very variable; however, there does
appear to be a pattern suggestive of more than a
single genotype. Considerably more research is
needed before anything more definite can be written
concerning the delimitation of subspecies.
One of the earliest reports of ecological consideration
and competition between palms and other
plants appeared in a discussion of T. ponceana by
Cook (1901:536):
Many of the palms are scattered among the taller shrubs
and trees wherever there is sufficient soil and water to permit
these to grow and yet not enough to give them exclusive
possession, but on many of the drier and more sterile higher
slopes the advantage is with the palms.
THRINAX PARVIFLORA.-h the broadest sense this
species occurs on nearly all exposed and eroded
“Jamaican Hard White Limestone” throughout the
western two-thirds of Jamaica. Along the southern
part of its range Thl-inax pa?-viflom is a conspicuous
element of the formation described by Loveless
and Asprey (1957:813) as “Dry Evergreen Thicket,”
typified by the vegetation of Portland Ridge in
the parish of Clarendon. Thrinnx parviflorn ’isas
described as “The commonest member,” of the
shrub layer, “. . . which although capable of growing
much taller, does not do so in this community
except where there is a gap in the canopy.” These
authors included the species as the principal representative
of the Palmae among the five most important
families of arborescent species in the Dry
Evergreen Thicket. Apart from the presence of the
“thatch,” the formation is characterized by the
absence of herbaceous ground cover and terrestrial
ferns. The substrate is generally deeply eroded bare
rock with occasional pockets of shallow soil and
very little humus. “Epiphytes are not well developed
but are represented by tivo xerophytic species,
the bromeliad Tillandsin fascicitlafa antl the orchid
B1-oztghtonia sangitinen.” Dry Evergreen Thicket is
a widespread formation on most of the hills and
ridges of low to middle elevations from sea level
to about 457.2 m (1500 ft) where the limestone is
naturally exposed. Local variations in floristic composition
result from local changes in topography
and edaphic conditions.
Areas of dry evergreen thickets on Dallas hlountain,
Long Mountain, lower llolynes hIountain,
antl Ferry Hill in St. Andrew represent the easternmost
part of the range of T. pnwifloio. The Hellshire
Hills of St. Catherine; Portland Ridge, Brazilleto
hlountain, antl Round Hill in Clarendon;
and the low irregular hills which rise along the
northern edge of the St. Catherine and Clarendon
plains, all support a dry evergreen thicket formation
which continues northward past Juan de Bolas
and llocho hlountains, hlende7 and Point Hills
to the slopes of hlt. Diablo. This same formation
containing T. painiflorn continues along the base
of the May Day Mountains in Manchester froin the
vicinity of Clarendon Park south past Round Hill
and west toward Alligator Pond, then north along
the southivestern escarpment of tlie Don Figuero
lfountains in the vicinity of Gutters and Spur
Tree. In St. Elizabeth a small area near Bull Savanna
in the Santa Cru7 Mountains represents the
only significant population in that parish.
In a few places at low elevation such as Cane
River Gorge, where cool moist air flows down from
the Blue Mountains, the physical environment is
modified by the lower temperatures and increased
atmospheric moisture causing the formation of
clew and fogs or increased rainfall. Somewhat similar
conditions seem to exist along the Rio Cobre
14 ShfITHSONI.4S COKTRIBUTIONS TO BOTAKY
in the gorge between Bogwalk and Spanish Toivn
and on the eastern slopes of the mountains near
Clarendon Park. In the extreme southwest of
Jamaica lies Negril Hill, a long low ridge running
parallel to the south coast in a region of moderate
precipitation 127-190.5 cni (50-75 inches) per year.
Thrinax parvifloiu occurs here at the westernmost
limit of its range as scattered individuals between
New Hope Estate and Rivival. Similar populations
occur to the north along the eastern edge of the
Great Morass near Silver Spring, Hanover.
Areas of frequent drought, described as Dry Evergreen
Busliland by Loveless and Asprey (1957,
support only the more drought resistant palm
Coccothrinax jamaicensis in the ecotone between
the two vegetation formations where i t replaces
FICURF 7.--Thrinax paruiflora growing in a siyal field near
Palmer's Cross, Clarendon, Jamaica. Notc the heat) texture
and coirupated armearance of the IeaFes.
T. puivifloia. This is the situation east of Cane
River in St. Andrew; south of May Pen, Clarendon;
the western end of Portland Ridge; and between
Gut River and Alligator Pond in Mancliester. Between
Black River and Savanna-la-mar, an area
with less than 88.9 cm (35 inches) of rain per year
and subject to long drought, T. pa?vzflora is absent
and Coccothrinax jamaicensis occurs only near the
parish boundary shared by St. Elizabeth and Westmoreland.
At Lances Bay in Hanover, an area of high rainfall
with over 190.5 cm (75 inches) per year, a few
individuals of T. pnrviflora have been found growing
on the low rocky slopes overlooking the sea
only a short distance from a population of T.
7adiata. While the cliffs supporting the latter face
northeast and are exposed to frequent salt-laden
winds, however, T. parvifloia occurs only on the
lee side of the ridge facing west. Along the north
coast, T. pnruiflora has been found at only two
localities near the sea: at White Bay, Trelawny,
and along the Queen's Highway near Bengal Bridge
in St. Ann. The region around 'IVhite Bay apparently
experiences occasional drought since the vegetation
consists mainly of logwood and thorn scrub.
Rainfall is normally under 127 cm (50 inches) per
year. The solitary robust plant of T. pnrviflora
growing along the Queen's Highway stands among
T. radiata and Coccothrinax jamaicensis.
Ddphin Head, south of Lucea in Hanover, is a
somewhat isolated mass of limestone rising to about
548.64 m (1800 ft). Although dry evergreen thicket
occurs on the lower cliffs, the upper ridge and
peak are swept by clouds and support a dense
woodland. Th?.inax paruifloru is a conspicuous element
of the more exposed and higher levels of the
mountain. A short distance to the east in the parish
of St. James, T. parviflora has been collected at
Bubby Hill under conditions similar to those on
Dolphin Head. A single plant was seen near Hastings
along the road to Maroon Town. It is curious
that although conditions appear to be favorable for
Thrinau, over the western Cockpit Country the
species, with one exception, has not been collected
or recorded anywhere between Deeside and Catadupa.
The species no doubt occurs as widely scattered
individuals, but the fairly evenly distributed
and high rainfall of 190.5 to 508.0 cm (75 to 200
inches) per year seems to be the limiting factor
" I I determining its exclusion.
SUhIBER 19 15
In the vicinity of IVindsor Cave, Trelawny,
sharply eroded limestone hills form both the northern
edge of the Cockpit Country and the headwaters
of the Martha Brae River. The annual rainfall
in the area averages between 127 and 190.5 cm
(50 to 75 inches) and appears to be fairly evenly
distributed throughout the year. The ridges are
heavily wooded and T. pal-vipora occurs as widel)
spaced individuals; however, the palm reaches its
best development on exposed cliffs. 4 short distance
to the north where the rainfall is much less,
about 127 cm (50 inches) per year, the hills and
ridges support larger populations of the species
near N t . Pantrepant and Sherirood Content nhere
typical dry evergreen thicket again prevails.
The eastern portion of the Cockpit Country is
characterized by a region of jagged limestone pinnacles
and interconnecting ridges 457.2 to 548.6 m
(1500 to 1800 ft) in elevation which set off huge pits
or “bottoms” often having vertical walls 60.96 to
152.4 m (200 to 500 f t ) high. Thi-znax pn?vlffo~n
again assumes a conspicuous place on the top of
each pinnacle, ridge, and cliff as at Mango Tree
Hi1l.l Frequent rains permit a rather dense wootlland
at all places where soil or humus can collect
in the “bottoms” and on the slopes betTveen the
pinnacles. A peculiar antl most interesting situation
exists in the huge cockpit extending from Barbecue
Bottom along the road to Burnt Hill. This is fully
described elsewhere.
East of the Cockpit Country is a region called
the Dry Harbour Mountains which closely resembles
the Cockpit Country but differs principally in
the more extensive scrubby to thicket vegetation
and higher elevations. This somewhat drier, rather
inaccessible region supports a large population of
Thrinax on the numerous ridges and pinnacles,
particularly in the neighborhood of Albion Mountain.
The lower elevation in the vicinity of Pedro
and Grierfield supports a heavily wooded vegetation
and relatively fewer Thl-inax. Scattered plants
occur in the vicinity of Browns Town where limestone
is exposed, but along the northern escarpment
the annual rainfall increases steadily toward
the east where 190.5 to 254 cm (75 to 100 inches)
of rain per year may prevail. Thyinax p a ~ v i f l o ~ a is
\$’hen questioned concerning a name for hlango Tree
Hill, the local farmers gave the name “Carambi Hill”; however,
“Mango Tree Hill” is the name coined b~ G. R. Ploctor
of the Institute of Jamaica.
found only occasionally near Bamboo and is rare
on the cliffs of Fern Gdly. North and east of a
line from Bamboo through Claremont, Moneague,
Guys Hill, Glengoffe to Stony Hill, Thyinax does
not occur except as individual outliers such as that
at Fern Gdly, no doubt because of a lack of proper
substrate antl the effects of increased rainfall. It is
now clear, after comparing the distribution of T.
parviflora with a rainfall distribution map, that
the greatest concentration of plants of the species
is to be found in the region classified as having
127 to 190.5 cm (50 to 75 inches) of rainfall per
year with numbers decreasing rapidly as rainfall
increases above 215.9 cm (85 inches) per year or
decreases below 101.6 cm (40 inches) per year.
Optimum rainfall for T. p a w i p o ~ a must then be
101.6 to 203.2 cm (40 to 80 inches) per year.
In the center of the Island betlveen the fertile
rolling cattle country of Claremont and the rich
valley of St. Thomas in the Vale (Ewarton-Linstead)
lies hlt. Diablo, a mass of limestone rising gratlually
from the north to over 914 m (3000 ft) near
the parish boundaries of St. A-inn antl St. Catherine.
An extension running to the northeast forms a
range of irregular limestone pinnacles (herein referred
to as the Devils Backbone) about 762 ni
(2500 ft) in elevation along the parish boundary.
The principal mass of 5It. Diablo terminates in a
very steep escarpment about 609 in (2000 f t ) high
along the southeast and west, overlooking Ewarton
and Worthy Park respectively. X spur projects to
the south and connects with a system of hills, of
which Point Hill and RIendez Hill are a part, and
encircles the valley of St. Thomas in the Vale to
the gorge of the Rio Cobre. Thrinax pn~viflorn
occurs on almost all exposed limestone and great
numbers cling to the walls of the gorge.
The vegetation above 548.64 m (1800 ft) elevation
on hlt. Diablo and associated slopes, which are
frequently bathed throughout the year in morning
clouds and mists, most closely conforms to the formation
described by J. S. Beard (1955) as “Dry
Evergreen Woodland.” Perhaps a new formation
to be called Montane Dry Evergreen IVoodland
would be preferred, since Beard’s formation applies
primarily to the lower elevations and not to those
influenced by clouds and mists. Thrinax parvipol-a
occurs on the more exposed cliffs and nearly all
of the higher peaks, e.g., Hollymount, Grier Mount,
Blue AIountain, Devils Backbone. The environ16
SMITHSONIAS CONTRIBUTIONS TO BOTANY
ment is perpetually moist and temperatures are
normally in the 20s ("C) or often lower. Softleaved
epiphytes are quite numerous on the tree
trunks and rocks where a lush growth of moss,
miniature orchids (Pleurothallis, Lepanthes), and
delicate ferns often obscure the substrate. Hohen-
Dergia polycephala and Vyiesea platynema, bromeliads
requiring a somewhat drier environment, are
restricted to the upper level of the canopy. Gzumania
erythrolepis, G. lingzilata, and Vriesea ringens
require cool moist conditions and are common
on the rocks and lower parts of the tree trunks.
The limestone outcrop upon which mature
plants of T. parvipora are localized, occurs as
ridges and small peaks with intervening valleys of
deep heavy soil or scree slopes. Although Thl.inax
FKXRK 8.--Thritinx parviflot-a growing at about 914 m elevation on Jlt. Diablo. Sote the open
crown and peculiar form of the leaves. Thesc plants and those in Figure 7 probably represent
the two greatest extremes of phenotypic variability of the species.
KUhIBER 19
seedlings are plentiful on all types of soil, they are
not known to reach maturity except on the stable
rock outcrop where similar lithophilus species such
as Euphorbia punicea and Plzimeiia marchii grow.
Similar environmental conditions are met with on
nearly all high limestone peaks and ridges such a5
Albion hlountain, Top Hill, Somerset, and Dolphin
Head.
The foregoing is a general account of the distribution
and ecology of Th1inax paruzfloin in Janiaica.
Further details of the behavior of this species
in two series of habitats where special ecological
features can be recognized are described elsewhere.
Morphology
The genus Thyinax, in common with Heirzitlri inax,
Coccothrinax, Zombia, and Haitiella of tlie
Thyinax alliance, is distinguished from all other
palms by having bisexual floivers in ivliich the
perianth consists of a single irregularly lobed cupule.
The genus is further distinguished fro111 Corcothrinax,
Zombia, and Hnitielln by the splitting
of the base of the petiole in the region ol' the
sheath (Figure 9c). From HemiLhi.innx it is zipparently
separated solely on the character 01 the
stamens. Hemithrinax with subsessile anthers, a
broad connective, and extrorse dehiscence is in need
of additional intensive study, but adequate material
of the different species is lacking at present.
During this study certain characters, mainly of
a quantitative nature and formerly believed to be of
diagnostic importance, were given particular emphasis
in order to test their useiulness as diagnostic
characters. The literature is full of descriptions in
which the size of the various organs (usually l'roiii
a single incomplete specimen) is all one has for
comparison of taxa. Good qualitative characters
were all but lacking.
mature flowering size plants were considered in tlie
data presented. Whenever possible, specimens were
taken from all accessible plants in a particular
locality (measurements and specimens were collected
from every plant exhibiting an inflorescence
at some stage from anthesis onward even if the
fruit had fallen already). Local populations consisted
of all the plants in a particular niche or
environment, e.g., a rocky pinnacle, gully, cliff, or
hillside. These are usually of limited si~e; how-
SELECTION OF SAkZPLES FOR h~EASLJRE1\IEXT.-Onl).
ever, if the population was extensive the collections
consisted of all the material that could possibly
be collected during the time available in one
or more trips to the area.
In order to study the species in every conceivable
environmental situation at all elevations in Jamaica,
a great deal of time was devoted to exploring
the island in search of new localities. This was
necessary in order to describe the range of distribution
and the range of tolerance of each species.
It was also necessary to collect representative specimens
from as many different localities as possible
in order to eliminate the bias caused by information
from classical collecting areas.
As much information as possible was obtained
from herbarium specimens early in the study, but a
concerted effort was made to re-collect in the localities
represented by existing specimens in addition
to the numerous new localities. It should be noted
that herbarium specimens rarely if ever contain
complete leaves or inflorescences. The most conimonly
occurring materials on herbarium sheets
usually comprise one or more of the lower primary
branches of the inflorescence at some stage betneen
anthesis and maturation ot the fruit but rarel) at
either of these stages, and fragments of a single
leaf, most commonly juvenile, Jvithout petiole or
sheath. Most frequently if a petiole is present it
and the hastula, and the blade, are split in two
rendering them fairly useless for measurements.
hIeasurements of the length antl diameter of the
caudex were made at the time and on the plant
froin which leaf and inflorescence specimens were
taken. The specimens consisted of one complete
inflorescence (usually cut at the sheath level, if
the plant \<as not cut down), two fully expanded
leaves with full petiole, two fully mature sheaths,
antl one unexpanded bud leaf. The leaves comprised
the most recently expanded leaf blade and
one of the older but still whole leaves. About half
of the segments were removed from the blade and
discarded by cutting them several inches above the
liastula and slightly to the side of the central
group of segments.
In order to achieve as much uniformity as possible
in the data from both field and herbarium
collection, where the apical portion of the inflorescence
was usuall) either badly damaged or missing,
it was necessary to restrict tlie collection of
data to the five lowermost primary branches on
18 SMITHSOKIIAN COSTRIBUTIONS TO BOTANY
FIGURE S.--Thrinnx pawi/lorn: A, bud region showing indumentum on the leaf sheaths and
internodes; R, the same as A but demonstrating how the inflorescence emerges from the split
sheath and arches out beneath the subtending leaf petiole; c, thc same as in A but a heavier
caudex showing more clearly the broad split which distinguishes the genus from Coccothrimx;
D, a single primary branch with its numerous flowers at anthesis.
NUMBER 19 19
each inflorescence. In the calculation of inflorescence
length (length of exposure above the sheath
apex), number of primary branches per inflorescence,
primary branch rachis length, number of
ultimate branches per primary branch, and the
length and diameter (at the point of insertion on
the rachis) of the ultimate branches, all parts of
the five primary branches were measured; however,
since the number of flower or fruit pedicels per
primary branch is excessively large a random sample
was necessary. This was achieved by randomly
selecting five ultimate branches from each of the
five lowermost primary branches per inflorescence.
The ultimate branches on each primary branch
were numbered from the base to the apex along
the rachis. These branches were randomly selected
by placing a pointer on some digit in the table
(1.2) of Snedecor’s (1948) table of random digits.
That digit and the one following it, be it “01” or
“25,” indicated the branch to be sampled. If the
number was greater than the number of branches
present, a second combination of digits was necessary.
Once the branches were chosen all pedicels
on each branch were measured.
CAvDEx.-The caudex of Thrinax is columnar
and nearly uniform in diameter throughout its
length but tapers almost imperceptibly toward the
apex, while the base is often noticeably expanded
2 to 5 dm above the substrate by the production
of masses of tough wiry roots. It is not uncommon
to see plants with large mounds of exposed roots.
The caudex may be gray or tan and is roughened
by alternating smooth internodes and slightly raised
leaf scars. As the palm ages the length of the internodes
being produced is much reduced, eventuall)
becoming a mere line encircling the apical portion
of the caudex in tall plants. The leaf bases are
persistent for some time following the death of the
leaf and are always conspicuous below the crown.
Within the tightly congested overlapping leaf
bases the internodes are velvety tomentose (Figure
A striking feature of plants of T. parvipoin when
growing at the higher elevations is tlie very tall
slender caudex; however, measurements from a
number of populations throughout the range of
the species show that there is great variability in
caudex height and diameter even lvi thin local populations.
The range of variation in caudex height
of 86 individuals of Thiinnx pni-riiflo~cl from among
9A).
19 different localities ,j is 1.0 to 10.0 (-13.0) in. The
mean heights for each locality varied from 1.5 to
10.0 m. These figures are to be compared to
(1.0-)1.5 to 7.5 (-12.0) ( Y z 4 . 0 ) m for T. radiata
(51 samples) and 3.5 to 9.0 (-11.0) (F=G.5) m for
T. excelsa (25 samples).O The range of variation
in caudex diameter of tlie same individuals of T.
paruiflora mentioned above is 5.0 to 14.0 (-15.0)
cm. The mean diameters for each locality varied
from 6.0 to 12.5 cm. These figures are also to be
compared to (7.5-)8.0 to 11.0 cm (X=9.0 cm) for
T. radiata and 12.0 to 16.0 cin (Y=14.5 cm) for
T. excelsa.
LEAF.-sheclth: The fully expanded leaf of
Thrinax is composed of four principal parts: blade,
hastula, petiole, and sheath. The mature sheath is
at first tubular and encloses the younger tleveloping
organs. The sheath has no doubt been the
most neglected portion of the palm in the making
of herbarium specimens and in studies of palm
morphology. Although it has not been fully exploited
as a diagnostic character, the forin of the
leaf sheath is of utmost importance in distinguishing
Thrinax and Hernithiinax from their closest
allies Coccothrinax and Zombia. Even very young
seedlings can be recognized by examination of the
leaf sheath. Coccothrinax and Zornbicl have leaf
sheaths which at maturity are still completely tubular
with the solid petiolar base continuing unchanged
to the point of insertion on the caudex.
The netlike tubular portion of the sheath retains
its appearance long after the blade has died and
broken off.
The mature leaf sheath of Thiinnx and Hemithrinnx
is, however, quite different from that described
for Coccothrinax and Zombia. The petiole
on entering the sheath portion is divided basally
in the form of an inverted wedge or ,I, (Figure 9c).
The fibrous portion is netlike at first as in Coccothl-
inax but soon breaks down and becomes
ragged or loosely fibrous. The conspicuously split
5The localities mentioned are from among those cited in
the study of pedicel length and mapped in Figure 41.
Thriiinv morrisii is escluded from these figures as it was
not included in the original intensive study. Considering the
extensive range of the species, the small amount of data
available would perhaps be misleading by comparison. More
complete data and bargraphs depicting the range of variation
and means of the individual localities for Thrinnx
f m i ~ i f l o r n can be found in Read (1968).
SMITHSOSIAS COSTRIBUTIONS 7 0 BOTANY
petiole bases persist below the crown long after
the fibrous wrapping has deteriorated.
The apical portion of the Thl-inax sheath is of
great diagnostic importance within the genus. That
of T. parviflol-a, T. excelsa, and T. morrisii is linguiforni
(Figure IOA,B) extending upward from
the point of attachment on the petiole and becoming
long pointed. This tonguelike extension is clothlike,
often separating between the fibers and then
appearing as a net. The sheath apex of T. 7ndiatn
on the other hand is not at all linguiform (Figures
l O c , l l ~ ) but takes an exactly opposite form with
the margins descending from their attachment
along the solid petiole in a V-form. The sheath of
T. l-adiata soons tears away from the petiole, even
as the leaf expands (Figure l l ~ ) , and the fibers
readily separate into ragged fibers around the “bud”
region. The lateral margins of the sheath may be
inserted on the petiole base at the same or different
levels.
The lower portion of the sheath of all four
species is covered with a dense velvety tomentum
(Figure ~ A , c ) but is usually lost on exposure. As
the young inner growth expands, the sheath fibers
pull apart permitting the tubular portion to expand
with internal growth pressures. The division
of the petiole also separates. Even in the very inimature
leaf sheath, a groove is present in the otherwise
solid portion of the sheath. 4s the “bud” expands,
the groove separates and in mature plants
the opening of the split exposes the young inflorescence
(Figure 9 ~ ) , but of course this is normally
FIGURE 10.-Leaf sheaths of the three Jamaican species of Thiinnx showing: A, B, the long
linguiform apex of T. excelso aiitl T. pnivilloin ieqpecti\ely; c, the deep V-form, nonlinguifoim
sheath of T. radiata.
NUMBER 19 21
obscured by the older, dead, outer leaf bases. As
the inflorescence develops and elongates, it continues
to separate the sides of the split petiole and
extends through the opening. At maturity the inflorescence
normally (in Jamaica) appears below
the petiole of the leaf to which it is in fact axillary.
Rarely does the inflorescence emerge above the
subtending leaf petiole.
Petiole: Petioles of Thyinax are slender, and
extremely strong and flexible. They vary considerably
in length, width, and form, even within a
species. However, the length of the petiole in relation
to the length of the longest leaf segment is an
excellent character for distinguishing between T.
radiata, wherein the petioles are shorter than the
longest segment, and T. parviflora, with petioles
longer than the longest segment of the leaf. The
range of variation of petiole length of 218 samples
of Thyinax parviflora from among 26 different localities
is (35-)40 to 155 cm. The mean lengths
for each locality varied from 40 to 135 cm. These
figures are to be compared to (36-)44 to 88 (-94)
cm (F=61.6 cm) for T. radiata (50 samples), and
125 to 200 cm (j;=165 cm) for T. excelsa (6 samples).
The range of variation in petiole width of
the same individuals of T. parviflora mentioned
above is 0.9 to 2.4 cm. The mean widths for each
locality varied from 0.9 to 2.0 cm. These figures
are also to be compared to 2.1 to 3.2 cm (X=2.5 cm)
for T. radiata and 2.5 to 3.9 cm (?=3.4 cm) for
T. excelsa.
Throughout most of their length petioles of
FIGURE 1 I.-Bud regions: A, Thrinax radiata; B, T. parviflora, showing the strongly linguiform
sheath of the latter, Xote the minute scattered scales on the petioles in “A” and the absence of
them in “B.”
22 SMITHSONIAS COKTRIBUTIONS TO BOTAKY
Thrinax are ancipitous and biconvex; however,
the solid portion within the sheath is usually concave
adaxially. At the junction between the petiole
and the sheath they may be flat or slightly concave
in T. paruiflora and T. excelsa or with a median
ridge in T. radiata. The petiole is widest at the
junction with the sheath and narrowest a short
distance below the hastula, where its margins may
be variously bent and distorted as the blade expands
(Figure 12).
The adaxial surface of the petiole is glabrous
or glabrescent. The abaxial surface of the emerging
petiole is densely floccose lepidote initially, but the
degree of persistence of the indumentum varies
among the species depending on exposure. The
petioles of T. parviflora lose their scales very soon
after exposure (Figure ~OB), while a conspicuous
amount of scattered scales persists on the petioles
FIGURE 12.-Three different iiews of the hastular legion:
A-C, Thrinax radiata; D-F. T. parvipora. (A, D = side views;
B, E = abaxial views; c, F = adaxial views.)
of T. radiata (Figure 10c) and T. morrisii. The
abaxial surface of the petioles of T. excelsa remains
velvety for a considerable time (Figure 1 0 ~ ) .
Hastula: The first mention of the hastula, or
ligule as it had been known to Beccari and others,
as a key character was by J. A. & J. H. Schultes
(1830: 1301). They stressed that the difference between
their newly described T. argenten and the
earlier T. paruiflora was that the former had a
“prominentia centrali minore.” They also distinguished
their new T. pumilio by its “defectu prominentia
in centro frondis.” All members of the genus
Thrinax exhibit a prominent adaxial hastula of
one form or another.
The hastula appears as an extension of the apex
of the petiole into the laminar portion of the leaf.
It is always in a plane with the petiole and overlaps
the region of segment insertion but may bend,
twist, or become semitubular depending on the
degree of expansion or folding of the blade. The
length of the hastular overlap was thought to be of
some importance, but as a generic key character, it
has been found of little use. The range of variation
in hastula length of 164 individuals of Thyinax
parviflora from among 26 different localities is
(0.1-)0.4 to 4.6 cm. The mean lengths for each
locality varied from 0.6 to 3.2 cm. These figures
can be compared to 1.0 to 2.0 cm (Y= 1.4 cm) for
T. radiata (20 samples) and 1.8 to 2.6 cm (y=2.1
cm) for T. excelsa (6 samples).
Although the shape of the hastula is quite variable,
the general outline and proportions hold
fairly true for each species but are extremely difhcult
to describe. It is important to point out that
the hastula is distorted by expansion of the leaf
blade and is thus best observed on the unexpanded
“bud leaf.” Occasionally the hastula may be torn
by expansion of the blade. The shape and form of
the hastula may be seen in Figure 12A,C,D,F.
Abaxially the apex of the petiole usually terminates
in a short overlapping collar of tissue which
may exhibit one or two small prominences 1 to 2
mm long. Throughout the habitats at the lower
elevations, however, Thrinax parviflora commonly
exhibits abaxial hastular projections of considerable
size (Figure 1 2 ~ ) occasionally exceeding the
adaxial hastula in length. Such prominent abaxial
hastulas have not been observed in any other species
of the Thrinax alliance. Leaves of a plant in
the population of T. parviflom at Cockpit, ClarenNUhfBER
19
FIGURE 13.--Thrinax parvipora, newly expanded leaf blade, with adhering scales: A, abaxial
surface; B, adaxial surface. c, D, Abnormal processes from the abaxial hastular region, on plants
found at Cockpit, Clarendon, Jamaica.
23
24 ShiITHSONIAS CONTRIBUTIONS TO BOTANY
don, Jamaica, were found to possess exceedingly
long, slender, rigid filaments arising from the abaxial
hastula (Figure 13c,~). The filaments appear
to be extensions of the retuse or slightly bifid
minor prominences mentioned above. No explanation
can be offered for this curious phenomenon
at present.
In bud, and upon emergence, the abaxial hastular
region of the young leaf is densely covered with
floccose lepidia, which are normally lost soon after
exposure. The hastular collar is usually fringed
with larger more persistent scales (Figure 1 3 ~ , ~ ) .
The newly exposed hastula of the T. moiiisii leaf
is always densely velutinous with silvery-white
tomentum.
Lamina: The leaf blade of Thyinax and its close
allies is completely palmate in form, having no
central costa or primary midrib. Although of basically
the same structure throughout the genus,
the leaf blade assumes a number of different and
characteristic forms. They also vary considerably
in size, texture, and indumentum.
The unexpanded leaf blade as it emerges from
the “bud region” is a compressed bundle of tightly
folded plicate segments with only the major nerves
and segment margins exposed. The exposed surface
is at first appressed lepidote, but after the blade
expands as in Figure I~A,B, the scales persist for
only a short time on the principal nerves.
The leaf blade often has been inaccurately described
as peltate because as the segments expand,
as a result of pulvinus-like regions along the largest
nerves in T. radiata and occasionally shade forms
of T. @arviflo~u and T. moirisii, a complete circle
is formed by the position of the segment apices,
and the outermost segments overlap across the
petiole (Figure 14). Leaves of mature T. paiviflora
are normally so folded, crumpled, twisted, or curled
that they present a three dimensional structure,
which is described more fully elsewhere. The fully
expanded leaves of T. excelsa are flat and nearly
circular in outline, but the outer lobes do not normally
meet nor overlap.
Segments: Segment number in members of the
Thrinax alliance has also been given importance
as a key character. Although the segment number
is fairly constant for the leaves of any given plant
of T. parviflora, varying from three to five about
the mean, the variation within a single population
may be as much as 22 segments between the greater
and smaller numbers. The measurements at locality
No. 33 in Jamaica were found to cover nearly the
entire range of segment numbers throughout the
entire population in Jamaica and considerably
overlap the numbers recorded for leaves of T.
radiata and T. excelsa. There is no direct relationship
between the size of a leaf and the segment
number. Thrinax excelsa with a leaf 3 m broad
may have 52 segments, while a leaf of T. @arviflora
only 1.5 m broad may have a similar number.
Leaves of juvenile plants normally have fewer
segments than those of mature plants.
The range of variation in segment number for
264 samples of T. parviflora leaves, from among 28
different localities, is 35 to 60. The mean number
of segments for each locality varied from 37 to 56.
These figures can be compared to 51 to 63 seg
FIGURE 14.--Thvinax paruiflora at Palmer’s Cross, Clarendon,
Jamaica. The region where the segments are joined is the
palman; note the tubular hastula.
SUllBER 19 25
ments (X=56) for T. radiata (50 samples) and
(52-)55 to 65 segments (X=59) for T. excelsa (41
samples).
Individual segments may vary in size, outline,
and degree of fusion at the margins. The segments
directly in line with the petiole are the longest
and broadest within the leaf. Those on either side
become progressively smaller away from center.
The degree of segment fusion is also greater at the
center where two or three segments may be fused
for more than three quarters of their length or the
two central segments may be fused for their entire
length.
Apart from the greater fusion of the central segments
the amount of fusion between segments elsewhere
across the blade is somewhat irregular. For
purposes of comparison a specific region of the
blade was selected to represent the leaf. Segments
to be measured were chosen by grasping the central
four segments in the hand and counting the next
three segments on each side of the group held.
The third segment on each side of the central
group of segments (thus two segments per leaf)
were selected and removed for measuring. In order
to collect a maximum amount of information, concerning
the leaf blade, with a minimum of bulk,
the most satisfactory method was to strip the selected
segments from the leaf, down to the hastula.
The remainder of the blade was then discarded,
except for the hastula, which was trimmed out and
retained if undamaged. The segments thus selected
from the same region on each leaf provided comparative
information regarding the length, breadth,
form, texture, indumentum, and length of connation.
The liastular region (Figure 12) and accompanying
segment bases provided additional information
regarding tlie petiole width, hastula form
and length, and the number of segments per leaf.
The leaf segments have been described variously
as lanceolate, ensiform, linear, etc., but these terms
are more applicable to leaves or leaflets with curved
margins. The margins of Thrinax leaf segments
are straight throughout their greater length particularly
in tlie proximal half. They may, however,
have near the broadest region a pinching in or
“shoulder” (Figure ~ ~ A , c , E ) . The chart of simple
symmetrical plane shapes published by the Systematics
Association Committee for Descriptive Riological
Terminology (Taxon 11, 1962) provides a
much more accurate terminology of palm leafsegment
outline. In the chart,? numbers 25, 54,
and 63 correspond closely to the outlines of the
leaf segments in Thrinax. These are basically narrowly
rhomboid, but vary from trullate to obtrullate
(Figure 1 5 ~ , ~ ) .
The leaf segments of T. parviflora are more often
than not conspicuously pinched in just above the
broadest part (Figure ~ ~ A , c , E ) . This character is
not peculiar to T. parviflora, however, for it is
found in T. radiata as well, but it is not so prominent
in the latter.
Segment apices may be bifid to varying depths
or not at all. In the leaf segments of T. parviflora
the apices are usually acute and often not at all
bifid. When bifid the apical points may be strongly
divergent (Figure 1 5 ~ , ~ ) . In T. radiata tlie segment
apices are long and slender (Figure 1 5 ~ , ~ ) .
usually deeply bifid and attenuated into a filament
which is readily lost. The outermost segments in
all species may exhibit long, slender, fiberlike filaments
up to twice their length. The occurrence of
filaments in the sinuses between the segments (such
as is typical of Sabal and Washingtonin) is rare;
however, several leaves of T. parviflora have been
found where similar but more delicate filaments
were produced. What may appear to be a greater
number of anomalies in T. parviflora should rather
be taken as an indication that the greater the
number of specimens examined in the field, the
greater number of anomalies will be found.
Although segment width as such is of little diagnostic
significance, the position of the broadest
part along the segment is very important. In
Thrinax parviflom the broadest part of the segment
is well above tlie sinus (Figure 1 5 ~ ) . In both
T. mdiata and T. excelsa the broadest point is at
the sinus (Figure 1 5 ~ ) , or very close to it. The
measurement of segment width was always made
at the broadest point. The range of variation of
segment width of 398 samples of T. parviflora from
among 29 different localities is 2.4 to 6.4 cm. The
mean segment widths for each locality varied from
2.4 to 5.6 cm. These figures are to be compared to
(4.6-)4.S to 6.2 (-6.4) cm (X=5.6 cm) for T. radiata
(50 samples) and (4.4-)4.6 to 7.2 (-7.4) cm (?=
6.2 cm for T. excelsa (20 samples).
The range of variation in leaf-segment length for
Reproduced in tV. T. Stearn, Botanical Latin, pages 318-
319, 1966.
26 SMITHSONI.4S COKTRIBUTIONS TO BOTANY
T. parrriflora is very great and overlaps considerably
with T. radiata, which in general exhibits
much longer segments. Those of T. excelsa on the
other hand are very long and even at the lower
limits of the range of variation measurements overlap
with those of T. radiata only at the 5 percent
level. The range of variation in segment length
for 428 samples of T. pa~uiflora from among 30 different
localities is 39 to 96 cm. The mean segment
lengths for each locality varied from 42 to 92 cm.
These figures can be compared to (73-)79 to
113 (-115) cm (z=92.9 cm) for T. radzata (50
samples) and 114 to 173 cm (X=144 cm) for T.
excelsa (16 samples).
Palman: The region of the leaf blade wherein
the segments are fused together by their margins,
measured from the hastula to the sinus between A
I
FIGURE 15.-Thrinax leaf segment outlines showing fusion
(indicated by arrows): A, T. parvipora; n, T. radiata. c, E,
Two types of constrictions on segments are depicted; D, F,
bifid apices.
the segments, is known as the palman (Figure
15A,B, between points indicated by arrows) like the
palm of the hand from the wrist to between the
fingers. The palman has been described variously
as one-third or two-thirds the length of the segments;
however, in T. paruiflora it is so variable,
depending on the age of the plant or exposure,
that such descriptions are of no practical value.
Actual measurements of the palman do, however,
have some comparative value among the taxa.
Measurements of the palman of each of the three
Jamaican species are compared as follows: the range
of variation in palman length of 749 samples of
T. paruiflora from among 29 different localities is
7 to 34(-42) cm compared to (30-)33 to 54(-55)
cm for T. radiata (50 samples) and (48-)52 to
84 (-100) cm for T. excelsa (16 samples). The mean
lengths for the palmans for each locality of T.
parzriflora varied from 9 to 28 cm compared to 44
cm for T. radiata and 71.8 cm for T. excelsa. The
length of segment fusion or palman depth was
measured using the same segments described under
length and width of leaf segments.
Because, as previously mentioned, the central
segments are fused for a greater length than the
outer ones, the palman is narrower distally and
toward the base than it is at the side. A common
characteristic of the leaves of T. radiata is the complete
fusion of the margins between the central two
segments. Elsewhere in the blade the palman outline,
formed by the sinuses, is fairly uniform. This
latter is also true for leaves of T. excelsa and T.
morrisii. In T. pa?-ui/?ora, however, apart from the
cluster of central segments with conspicuously
longer fusion, the length of the fusion between the
segments is very irregular throughout the blade.
Undoubtedly, the largest palman measurements
for the genus are those of T. excelsa. They clearly
exceed even the largest measurements of T. parviflora
and overlap the very largest of the measurements
of T. Yadiata by only a few centimeters: and
the overlap in palman depth measurements between
T. parviflora and T. radiata occurs only at
the 5 percent level. Therefore palman width or
depth can be used in part as a diagnostic character
in Jamaica, providing of course only fully mature
specimens are used and the measurements are
made in precisely the same manner as those here.
Indumentum: The adaxial lamina1 surface, except
for the major nerves, is for the most part
R’UhfBER 19 27
glabrous or glabrescent. Minute scales may be
found occasionally on the adaxial surface near the
apical portion of the segments, especially on freshly
opened leaves. Although a waxy surface is characteristic
of palms throughout the Thrinax alliance,
it appears that the leaves of T. paraiflorn and especially
T. morrisii have a particularly great amount.
One of the first peculiarities noted by the author
in regard to T. fiarvifEo?a was an abundance of
white “hairlike” material on the surface of the
leaves in a number of herbarium specimens. On
closer examination the material appeared as
minute needlelike crystals, which dissolved in absolute
alcohol. When drops of alcohol were washed
across a leaf segment and allowed to dry on a piece
of glass, a quantity of white waxlike substance resulted,
The same held true for both fresh and dried
A h
FIGURE 16.--Infiorescence bracts, the ultimate tubular bifid
bracts on each primary branch: A, B, Thrinax parvipora,
adaxial and side views; c, T. radiata, adaxial view; D, E, T.
excelsa, side and adaxial ciews respectively; F, c, prophyl in
position, on inflorescence axis, in the axil of the leaf sheath.
Seed germination: H, T. excelsa; I, T. parvipora.
leaves of T. pnrviflora. When the same experiment
was run using leaf segments of T. radiata the
amount of wax deposited was negligible by comparison.
Abaxially the laminar surface is variously lepidote,
the differences providing excellent diagnostic
characters. AIthough the abaxial surface of the
leaves of T. pawiflorn appears to be glabrous, it is
in fact glabrescent with widely scattered microscopic
scales which are readily lost with age. Unlike
the other species, the scales on some specimens of
T. pamiflora appear to be reduced to simple hyaline
hairs or “Y-shaped” scales which become detached
quite readily. A magnification of about 25
times is necessary in order to see the scales on
T. parviflora.
The scattered scales on the leaves of T. radiatn
are quite conspicuous and may be seen easily with
the unaided eye as tiny flecks over the abaxial surface.
On a fresh leaf the individual scales may be
shortly fimbriate from an oval glandlike center.
On drying or with age the fimbria may be lost
leaving the conspicuous central portion which can
be readily seen on herbarium specimens as well, especially
with a 10-power lens. The abaxial leaf surface
of T. radiatn may, in addition to the scales, be
glaucescent, but it is never silvery white.
Thrinax excelsa is readily distinguished from the
aforementioned by having the abaxial surface of
the leaf blade densely covered with interlocking
fimbriate hyaline scales. Fully mature leaves of
this species are conspicuously white to silvery white
below as a result of the dense covering of hyaline
scales. The leaves of juvenile plants of this species
appear glabrescent but nearly always exhibit
patches of interlocking scales and hairs.
Among the widely distributed populations of
Thrinax morrisii, the degree and development of
abaxial scales varies considerably. Whether this is
a reflection of the ecology or is truly genetically
determined remains a question for further study of
the living plants throughout their total range of
distribution. Specimens from Anguilla, the type
locality, appear depauperate and exhibit very
small, poorly developed scales; one specimen
nearly lacking scales altogether. Specimens from
Puerto Rico through the Bahamas and Florida
exhibit increasingly greater numbers of scales with
greater and greater degrees of fimbriation on individual
scales. In the Cuban plants the fimbria are
SMITHSOXIAS COKTRIBUTIONS TO BOTANY
greatly extended and often become interlocking
much like the scales on leaves of T. excelsa but
are not quite as densely arranged.
INFLoREscENCE.-ktxis and Bracts: The inflorescences
of the four species of Thrinax are basically
of similar structure but differ markedly in size,
puberulence of the branches, and color at anthesis.
Originating in the axil of a leaf, at a point immediately
below the petiolar groove, the inflorescence
is adnate to the leaf for about 1 cm (Figure 1 6 ~ ) .
The first inflorescence bract is inserted about 2 cm
above the base (Figure 16~). It is bicarinate, flattened
adaxially and convex or rounded abaxially,
and is completely closed at first. The apex is
pierced by elongation of the second bract.
The inflorescence of Thrinax and its close allies
consists of a short weak axis with between 7 and 22
primary branches supported mainly by large, fibrous
tubular bracts telescoped one inside another from
the apex to the base. Each bract, with the excep-
LN.
#
9
30
12
5
38
31
8
10
7
27
25
15
26
2l
2
3
18
42
34
14
17
ut
22
29
28
40
13
16
19
-
2
250
175-250
756-900
1,250-1,600
400
10-20
40
10
500
200
L,500-1,650
500
250
tion of one to several empty basal bracts, subtends
a primary branch, enclosing the basal portion of a
primary branch and the lower portion of the next
higher bract. The apical aperture of each principal
bract is oblique. The bracts are densely appressed
lepidote varying in color among the species. In
T. radiata and T. morrisii the scales are gray or
white becoming tan with age; those of T. parviflo~a
are light brown or tan at first darkening with age;
and in T. excelsa the scales on the bracts are dark
rusty brown.
Measurements of the variation in inflorescence
length within Thrinax parvifloru are recorded in
Figure 18. The inflorescence in T. paruiflora usually
equals the petiole in length but may occasionally
exceed it, extending up among the leaf
blades, but is not known to exceed the leaf in
length. Inflorescences of T. radiata and T. morrisii
are always longer than the leaves and extend conspicuously
well above the crown of leaves. The
w ............. ..... v ........... ...... .. - ....... 0..
.9 ...... ......... V .................. .... D ..... .............. n - ........................... .. ................................. ...... ... ...... .................. “ ...... ........... ...... ......... Y .................................. .. - ........ ..................... - R .. .................. .. ....................... .... w ...................... .......... - ..... ........................
NO.
TITS I
I
3
I
I
3
9
5
25
I
I
25
3
I
2
2
3
1
3
I
2
I
2
4
I
25
2
I
2
3
1
I
-
FIGURE 17.-Pedicel length, Thrinax parvipora, arranged by increasing means, top to bottom,
showing the wide range of variability of this once “important” taxonomic character. The x in
bar graphs represents mean length, solid line 90% of all measurements, and dots indicate 5%
of total off each end of range.
43
24
2h
A. 1
9.
c b 11111111
1111 -
w
4C 50 60 70 80 90 100 U O DO 130 140 19 160 170 lf0
Inflorescence length in em
37
24
31
A.
B.
C.
5 10 15 20
Number of primary branches per inflorescence
25
3 93
263
496
1.3 2.0 3-9 4.0 5.0
Pedicel length in m
A.
B.
C,
2.0 4.0 6.0 8.0 10.0 l2.0 14.0 16.0
Length of ultimate branches in ern
A. - I-*-
B.
C.
- - -1
w .-I
1-
39
32
50
35 40 45 50 55
Wmber of segnents per leaf
60
A.
B.
C.
20 40 60 80 100 120 140 160
Petiole length in em
A.
B.
C.
I. -1
=-
I.-, .
35
24
29
0.2 0.4 a6 0.8 LO u L4 1h 1.8 2.0 2.2 2.4
Petiole width at apex in em
35
21
25
A.
B.
C.
2.5 5 *O
Trunk diameter in em
7.5 1.00
FIGURE lS.-Measurements of characters from the three intensively studied populations in
Jamaica: A, Hollymount Hill; H, Mango Tree Hill; c, Cockpit, Clarendon. Sample size on right;
bar graph as in Figure 41.
30 ShlITHSONIAS COSTRIBUTIOSS TO BOTASY
inflorescence in T. excelsa rarely equals the petiole,
it is usually strongly curved and much shorter than
the petiole.
Each primary branch of the inflorescence supports
a bicarinate bifid bract (Figure 1 6 ~ - ~ ) with
ciliate margins, midway along its peduncle and
partially included within the tubular primary
bract. During development of the inflorescence and
before the flowers are exposed from the primary
bracts the bifid bract completely encloses the primary
branch and its flower buds. Early on the day
in which anthesis takes place, the primary branch
including its branchlets elongates rapidly, extending
15 to 20 cm in length during a 10-hour period.
The branched portion of the primary branch is
pushed out beyond the bifid bract and becomes
pendulous (Figure 9 ~ ) .
The color of the primary branches and branchlets
at flowering is characteristic of the species,
particularly in T. excelsa. All parts of the branches
and the flowers of T. excelsa are a bright rosy pink
or purple, while in T. paruiflora they are creamy
white to yellowish. In T. radiata and T. morrisii
they are white, only becoming cream colored or
yellow with age.
The primary branches throughout the genus are
known to be once branched; however, in T. parviflora
an occasional plant may produce a primary
branch that has the lower branches secondarily
branched. The primary branches of T. pa? vipora
are typically densely puberulent throughout most
of its range. Nevertheless at the higher elevations
puberulence becomes less pronounced and at the
highest elevations is restricted to protected areas in
the axils of the branches and in lines on the peduncle.
The primary branches of T. radiata, T.
morrisii, and T. excelsa are entirely glabrous.
Flowers: The bisexual flowers of the Thrinnx
alliance are exceptional in the Palmae in having a
single whorl of perianth, kariable stamen number
from 5 to 15 or more, and a unicarpellate unilocular
gynoecium. The flowers of Thrinax are very
like those of the rest of the alliance. Except for the
peculiar stamen character in Hemithrinax there is
little of practical diagnostic significance in the
flower. The flowers and associated inflorescence
.
. . . .
. .
FIGURE 19.-Frequency and distribution of pedicel lengths in the three areas of intensive study
in Jamaica: A, Hollymount Hill; B, Mango Tree Hill; c, Cockpit, Clarendon. Each point on the
ordinatr represents 10 measurements; pedicel length on abscissa in mm.
NUMBER 19 31
parts vary greatly in size and form even within a
single species. Those of T. paivifiora are the most
variable among Jamaican species. Thrinax radiata
flowers are very like the larger T. parviflora flowers.
Thrinax excelsa flowers are rather uniform in
size and form and differ from the aforementioned
species primarily in the pink perianth and a simple
horizontal stigmatic opening.
All Thrinax flowers are to some degree pedicelled,
although in T. morrisii the pedicels may be
very short and disklike. hluch artificial division of
the genus has resulted from the use of pedicel
length, its presence or absence, primarily because
of inaccurate conclusions based on inadequate material.
Based partly on this character specimens of
T. pal-viflora have been described as distinct taxa
and placed in t~7o different subgenera. Beccari
(1907, 1933) considered T. tessellata and T. harrisiana
to be distinct. He distinguished them from
T. pamiflora in his keys by their very short pedicels.
Bailey (1938) also considered T. tessellain
and T. harrisiana to be distinct but referred them
to Sargent’s subgenus “Sessiliflorae” with “sessile”
flowers. Both names are synonymous with T. parviflom
Sw., not of Bailey. Measurements of the variation
of pedicel length of plants in the type locality
of T. tessellata (Hollymount) are recorded in Figures
17 and 19, as locality #lo, where they can
be compared with the variation throughout the
species. From a study of a large number of specimens
it becomes evident that the flowers are more
frequently long pedicellate than “sessile.” No inflorescences
have been found where conspicuously
long pedicels are lacking, and none are truly sessile.
The perianth is a single series in cupular form
and has a slight basal depression Ivhere it fits over
the apex of the pedicel (Figure ~ O A , B ) . The upper
margin is irregularly “6-lobed” or with 6 triangular,
acute or attenuate prominences (Figure
2lc,c,P, and others). According to a study of the
anatomy of coryphoid palms, hforrow (1 965 )
stated that in Thriiiax there are 14 vascular traces
to the periantli determined at one level as compared
to 3 in Hemitlzrinax. Stamen filaments are
mostly free but are always dilated slightly at the
base where they are usually very shortly connate
The material utilized by Morrow in his anatomical studies
was unfortunately wrongly identified as T. excelsa. This was
the T. excelsa of Bailey, not Lodd. ex Griseb. It should therefore
he called T. mdiata.
beneath the ovary (Figures ~ I B , L ; 2 2 ~ ; Z~C,E;
24G,J,K); although in a key utilizing corollaandroecium
structure Morrow (1965:341) stated:
“C. Filaments free of adnation and connation” by
which he distinguished Tlzrinax from other members
of the alliance which possess, “C. Filaments in
a ring, becoming free of connation higher up.” The
stamen filaments of Thrinax may be fused in varying
lengths up to the apex (Figure 2 4 ~ ~ ) . One unit
may in fact appear as a single broad filament with
two or three anthers affixed (Figure 26). It is to be
regretted that in Morrow’s discussion of the “Evolution
in the Calyx and Corolla of Coryphoid
Palms,” he neglected any mention of the peculiar
and unique form exhibited by Thrinax and its
allies, apart from an exception to his argument
against a polyphyletic origin of the calyx of coryphoid
palms on page 335. The occurrence of a
perianth in a single whorl must have some evolutionary
significance, either as a very primitive condition’
or a highly advanced one in which adnation
and connation of calyx and corolla parts are carried
to the extreme.
From the earliest description of the genus the
androecium has been said to have 6 stamens, and
although flowers are commonly observed with 6
stamens the most frequently occurring numbers
are 7 or 8 in T. parvifiora, 7 for T. radiata, and
FIGURE 20.-Schematic drawings of developing frUit:A, immature
fruit on pedicel, longitudinal section; B, base of fruit
and perianth, nl, testa, B2, pericarp, n3, stamen filament,
154, perianth; cl, endosperm c2, 3, testa, c4, vacancy by
shrinking endosperm.
32 SMITHSOSIAN CONTRIBUTIONS TO BOTANY
8 or 9 for T. excelsa. Five stamens per flower is by 5 to 10(-15). The mean number of anthers varied
no means a rare condition and the number of sta- from 6.5 to 8.8 among the 11 collections. These
mens quite often exceeds 10 or 12. The range of figures are to be compared to 5 to 8 (-10) (y=6.5)
variation in anther number of 784 flowers of for T. radiata (127 samples) and (6-)7 to 10 (-11)
T. @rviflora from among 11 different collections is (X=7.5) for T. excelsa (50 samples).
FIGURE 2l.--Thrinax parvipora, iariahility of Rowers: A-T., Hollymount Hill; H-I, DeLils Backbone;
J-P, Cockpit; B, L, filament bases; D, 0, longitudinal section of pistil; c, G, P, perianth;
N, cross section showing perianth and filaments.
NUMBER 19 33
The anthers are basifixed, and although the sacs anthers are always slightly divided or retuse at
both apex and base. The individual sacs are commonly
apiculate and unequal in length. The range
of variation in anther length of 1107 anthers randomly
selected from among 12 different collections
are normally joined by a very narrow connective
the condition has been observed where one sac is
attached only at its apex while the other is affixed
normally along the length of the filament. The
& D
H
B I Q J
N
FICLRE ZZ.--Thrinnx finrviflora, iariabilit) of flowers: A-G, Long Mountain; H-K, Mango Tree
Hill; L-N, Palmers Cross; I, F, XI, pistil form; G, s, perianth, basally; K, filaments and perianth
from above.
34 ShIITHSONIAS CONTRIBUTIONS TO ROTzlhY
of T. parviflom is (1.1-)1.2 to 4.0 (-4.2) mm. The The gynoecium is unilocular, “bearing a single
range of means varied from 1.6 to 3.4 mm. These basal bitegmic ovule. The orientation of the microfigures
can be compared to (2.0-)2.2 to 3.4 (-3.6) pyle is slightly oblique but it is more orthotropous
mm (Xz2.8 mm) for T. ~ a d i a t a (199 samples) and than campylotropus” (Morrow, 1965:lOB). “The
(1.2-)1.4 to 2.0 (-2.3) mm (y=1.6 mm) for T. funicular aril is developed to the same extent in
excelsa (100 samples). Thrinax as in Coccothrinax and Hemithrinax
n
N F-3 U
C
H
FIGURE PJ.--Thrinax mdiata: A, (,-I, variability of flowers; B, base of the ovary in vertical
section showing ovule; c, base of filaments and perianth; D, stigma; E, filaments; F, branchlet
and pedicels.
SUXIBER 19 35
(1965:llO). The ovary is globose to slightly pyriform,
usually abruptly constricted into a short style,
which flares out into an infundibuliform stigmatic
region in flowers of T. ~awiflora and T. radiata, and
is usually quite oblique or sometimes lateral. In T.
excelsa the stigma appears as a simple conical depression
in the apex of the style. The margins of
the stigmatic region in T. paruipora and T. radiata
are frequently undulate and ciliate.
Observations over a period of time on a plant of
B H
9 I
F
B
G
C
L
E
FIGURE 24.-Thrinax excelsa: A, B, flowers at anthesis; c, D, apical portion of branchlet showing
the relationship of terminal flower and apex: E, shows an apical flower with its subtending
bract; F, perianth cupule; G, filaments surrounding the base of the ovary, some variously connate;
H, longitudinal section of pistil showing v-form stigmatic region, and ovule; I, stigmatic region
in the apex of the style; J, connation of filaments: K, stamen filaments dissected from flower;
L, base of ovary in cross section, showing ovule and funicular aril; Lr, funicular aril removed
from ovule.
36 SMITHSOiiIAN COSTRIBUTIONS TO BOTANY
T. paruiflora reveal that the stigma is tightly compressed
laterally obscuring the stigmatic surface at
anthesis. Within twelve hours all anthers had dehisced
and the released pollen was adhering to all
surfaces of the inflorescences, but the stigmas were
still closed. Twenty-four hours following anthesis
the stigmas were open in the form of a funnel.
Pollen was still plentiful on the inflorescence
branches and by shaking the inflorescence a large
quantity was dislodged. A second plant in the immediate
vicinity was at anthesis at the time the
stigmas were receptive.
The length of the style-stigma is quite variable
and although fairly constant among the flowers on
FIGURE 25.-Immature fruits of Thrinax pamiflora showing
variation of stigmatic remains; both from the same collection
a sing1e plant, it varies among the
plants in a single population. The stigmatic reon
the Devils Backbone, Jamaica. mains are most prominent during development of
FIGURE 26.--Thrinax parviflora: A, abnormal flower with two anthers on a single filament;
B, seed dissections showing intrusion of the testa (upper row longitudinal and cross sections
through the embryo, bottom row apical and basal intrusion, left to right respectively); c, D,
developing fruit exhibiting symptoms of fungal infection.
NUMBER 19 37
the fertilized ovary and in the young fruit. Figure
25 illustrates the variability of the length of the
stigmatic remains on the young fruit of plants of
T. paruipora growing side by side on the Devils
Backbone.
The ovary enlarges rapidly following anthesis
and therefore provides a very poor character for
comparative studies. The surface of the ovary in all
three Jamaican species is smooth at first but follo~ving
fertilization considerable change takes place,
especially in T. pU7Uifl07”U. As the ovary enlarges,
the surface rapidly becomes slightly papillate and is
soon completely covered by minute granulations
(Figures 25, 2 6 ~ ) .
In a discussion of the nature and evolution of
the unilocular gynoecium that Rforrow queries as
“unicarpellate or multicarpellate,” he (1 965) suggests
that
there are at least three possibilities:
1) The unilocular gynoecium is a single carpel, produced
from ontogenetic or phylogenetic abortion of two other
carpels.
2) The unilocular gynoecium is a single carpel but ha?
captured the vascular supply of two carpels which abort,
3) The unilocular gynoeciuni is composed of three carpels,
which grew together in such a manner as not to be recognizable.
He further suggested that “first . . . the unilocular
forms have skipped a developmental step; and
second, the development of the ovary occurs prior
to anthesis.”
Morrow (1965:355) stated that, “Normally unilocular
species (Thrinnx et al.) have been seen
with 2 locules,” and in this connection it seems
appropriate to describe a curious phenomenon ob-
FIGURE Z’i.-.Anomalous flowers of Thi itznx p1n1iflorn; apical
flowers exhibiting partial fusion of parts. Explanation in the
text.
served in a number of specimens of Thrinnx pmviporn.
A herbarium specimen at the Institute of
Jamaica had on it a flower that seemed to be bicarpellate,
but since soaking and dissection would
have destroyed the flower, investigation was delayed
until fresh material could be obtained. Fresh flowers
at anthesis were collected on Portland Ridge,
Jamaica, in which apparent fusion of flowers was
in process. Figure 2 7 ~ shows the apex of a racliilla
upon which the apical flower possessed two distinct
carpels and nearly double the normal number of
stamens. The perianth formed a single 12-lobed
whorl surrounding tlie base of the androecium.
The possible explanation became apparent when
the flower illustrated in Figure 27A-C was studied
and dissected. On one side the perianth, although
partly fused at the base, turned in apically and
continued partly between the two carpels (Figure
25B). On the opposite side the perianth was a
continuous “normal appearing” cupule enclosing
both halves of the “flower” (Figure 27A). Two of
the stamen filaments obviously originated between
the two carpels. The arrangement of androecia and
gynoecia diagrammed in Figure 2ic clearly indicates
that two flowers are partially fused on the
apex of the rachilla. Fusion of the perianth is onl)
partial1.y complete. Other flowers have been observed
in which the two carpels are apparently
situated side by side without stamen filaments
between them and with tlie perianth completely
enclosing the base as a single multilobed unit. A
situation of this sort could possibly explain the
2-locular ovary cited by Morrow.
Fruit: The fully developed but not fully mature
fruit of T. p a ~ i p o r a may exhibit any one of several
different characters. Fruit not infected or damaged
by fungus attack (see section on Phytomorphosis)
may be granulate, densely papillate, or even
tuberculate and they often become scurfy as the
h i i t matures. Young fruit of T. radialn and T. excelsn
are generally smooth to slightly pustulate or
with widely spaced papillae. F~illy mature fruits of
all four species are smooth and white or creamy
white.
The seed of Thrinax is subglobose with a subapical
embryo and is intruded partly or completely
by the testa. The seeds of all three Jamaican species
are completely intruded from base to apex. The
seeds of T. mowisii differ in that the seeds are not
completely intruded by the testa, rather the testa
38 SVITHSONIAS COX TRIBUTIONS I 0 BO I'ASY
intrudes little more than halfway from base to
apex. In fruit not quite fully developed (Figure
~ O B ) , the endosperm (cl) is rubbery and the intrusion
of testa is easily observed as a dark fleshy
tissue (c2) continuous with the very thin testa
(c3) surrounding the endosperm. In longitudinal
section the drying endosperm has pulled away
from the central intrusion of testa (Figure 20c4;
see also Figure 2 6 ~ ) .
GERMINATION OF SEEDs.-The mode of germination
of the seeds of Thiinax is similar to that of
Phoenix as described by Gatin (1906) and Tomlinson
(1960). The cotyledonary haustorium remains
in the endosperm and the cotyledonary petiole
(Figure 1 6 ~ ~ ) by elongation pushes the radical out
and down, burying the plumule and cotyledonary
sheath a few centimeters into the substrate. The
plumule then emerges through a split in the upper
surface of the cotyledonary petiole.
Of interest, taxonomically, is the fact that in
T. excelsa and T. moiyisii the cotyledonary sheath
is pubescent throughout but that of T. pawifloin
is pubescent for only halt its length, the lower
portion remaining naked.
The first postcotyledonary leaf is tubular, sheathlike
and bladeless, and usually pointed at the apex.
The first seedling leaf, termed an eopliyll by Tomlinson
(1960), has a narrowly lanceolate green
blade. Several successive seedling leaves are lanceolate,
but leaves with three, and eventually more,
segments are soon produced. The leaves produced
henceforth become more and more complex as they
appear with increasing numbers of segments.
P O L L F N . - T ~ Y ~ R ~ X pollen is pale cream colored
to white, elliptic to broadly elliptic in outline, and
monosulcate. The exine, when viewed by means
of the light microscope, appears to be fairly uniformly
psilate-foveolate among the four species;
however, photographic studies by means of the
scanning electron microscope clearly suggest the
possibility of distinguishing between species by
means of pollen-exine morphology. Further studies
FIGURE 28.-Portion of fruiting inflorescence of Thrinax'nwrrisii; fruit about natural size; note
the bumplike pedicels.
SUhlBER 19
of pollen from the extremes of the ranges and ecologies
of the four species and their subspecies would
be necessary in order to establish whether or not the
types illustrated here are in fact really different or
if they merely represent isolated parts of a range
of variation. Certainly the character of the exine of
T. paruiflora is somewhat distinct from that of
T. excelsn, and the exine morphology of T. radiatn
and T. ~ n o i ~ i ~ i i , while differing somewhat from the
others, appears to be simply variations on a similar
theme.
Anatomy
In his studies on the anatomy of palms, Tomlinson
(1961:3) was correct in his observation that
“specific identifications have never been very relia-
FIGURE 29.-Scanning electron micrographs of Thrinax pollen enlarged about 2600 times: A, T.
parvipora; B, T. mdinfn; c, T. excelsn; D, T. morrisii. (Photographs courtesy of the SEM
laboratory staff, from samples prepared b j - the Botany Palyology Laborator) at Smithsonian
Institution.)
40 5>51THSONIAN CON1 RIBUTIOSS TO HOTAUY
FIGURE 30.-Scanning electron micrographs of Thritiax pollen magnified about 10,000 times.
A, T. parviflora; B, T. radiata; c, T. excelsa; D, T. morrisii. Xote how different the surface
pattern appears under the much greater magnification.
SUMBER 19 41
ble.” But he states: “In many genera this is no
disadvantage since interspecific differences are often
only quantitative and are obscured by anatomical
variation within a single species.” This statement
may be true for other palms but in the genus
Thrinax, once the taxonomy is worked out, certain
anatomical characters follow the same pattern as
the key morphological characters. It is now possible
to identify accurately each species of the genus
simply by comparing the anatomical characters of
the lamina with the following key or the accompanying
illustrations.
Anatomical Key to the Species of Thrinax
Using Laminar Characters
I. Laminar cross section composed almost entirely of palisade cells: stomata deeply sunken:
abaxial outer epidermal cell walls greatly thickened, surface very irregular . . T. morrisii
1. Laminar cross section composed of either distinct layers of palisade and mesophyll, or palisade
cells few and scattered; stomata not sunken: abaxial epidermal cell walls not greatly
thickened nor is the surface highly irregular.
2. Palisade and mesophyll layers quite distinct; hails when present large and cornplex with
numerous surface cells in cross section.
3. Stomate guard cells in cross section, hooklike in appearance . . , , T. radiata
T. excelsa 3. Stomate guard cells in cross section, not at all hooklike in appearance
2. Palisade cells absent or rather uncommon and widely scattered if present; hairs when
present, small and with few surface cells; stomate guard cells, in cross section hooklike
in appearance
hIATERIAL AND h!fETHODS.-In most cases Samples
of 2 sq cm each of dry herbarium material were
selected and soaked overnight in aerosol (Ayensu,
196‘7). Thin sections were then made by hand using
a sharp razor blade with the specimen supported by
elderberry pith. The sections were then stained
using a 0.05 percent solution of toluidine blue in
0.15 R/I phosphate buffer for a few seconds and
mounted in ethylene glycol under glass cover
slips. Neither soaking in sodium hypochlorite solution
nor heating in hydrochloric acid seemed to
make the material easier to section. The material
to be sectioned was selected from the midportion
of leaf segments, about halfway bet.rveen apex and
hastula, between the major veins.
The following account of the anatomy of the
lamina of Thrinax is based in part on a treatment
of the genus by P. B. Tomlinson (1961), but it
has been modified significantly by more recent personal
observations. Tomlinson’s treatment was unfortunately
based on a mixture of both Coccothrinax
and ThTinax, closely allied but distinct genera.
He (p. 296) concluded that
from the material . . . examined, although some of it has
not had a reliable name and may have been confused Hith
Coccothrinax, there appears to be considerable \ariation in
the anatomy of the lamina, particulaily in the stiucture and
distribution of the hairs and the disttibution of the ceins
~
T. pamiflora
and fibrous strands, Since the taxonomy of Thrinax and
related genera seems to be rather confused, little reliance
can be placed on the anatomical information which has SO
far been recorded until a detailed survey can be made upon
a large number of species from accurately established sources.
The elements that do not agree with observations
on verified Thyinax materials have been removed,
and in addition to my own new observations those
that do apply to the genus Thyinax have been
retained.
LAMINA o.--Hnins frequent in abaxial costal regions,
each with sunken, elliptical, sclerotic, multicellular
bases. Ciilicle thick, Epidermis wholly cutinized,
outer wall somewhat thickened. Adaxial
epiderniis fairly uniform, costal regions tlifferentiated
only above largest veins, cells rather variable,
mostly more or less cubical or slightly longitudinally
extended, walls not sinuous. Xbaxial
epidermis differentiated into narrow costal and
wider intercostal regions. Costal cells rectangular,
longitudinally extended, narrow, shallow; intercostal
cells less regular, shorter, wider, deeper,
and slightly papillose. Stomata restricted to
abaxial intercostal regions, those of both T. ra-
’ Grateful appreciation i? extended to Mrs. Li5a Keys,
graduate Ttudent at the CniLersity of \faswchusetts, for her
freelt \olunteered help in the preparation of materials and
hand sections used in the laminar anatomy.
42 ShIITHSOSIAS CONTRIBLTIOSS TO BOTASY
FICCRE 31.-Camera lucida drawings of Thriiznx leaf sections: A, T. parviflorn (Read 1736,
Devils Backbone); B, T. escelsn (Read 193Gb, John Crow Mountains); c, T. radiata (Read 1862,
Morant Point); D, T. morrisii (Read 798, Grassy Key, Fla.).
KUMBER 19 43
diata and T. parvipora similar in cross section as
in Figure 31~,c, while T. excelsa and T. morrisii
share similar guard-cell types which differ from the
other two species (Figure 31n,~). Hypodermis 1- to
2-layered below each surface but adaxial liypodermis
mostly 2-layered; cellfiles often replaced by
fibers. Adaxial cells rectangular, longitudinally extended.
Abaxial cells more cubical, irregular in
intercostal regions. Chlorenchymn with a fairly distinct
2- to 3-layered adaxial palisade (except T.
parvipora which has only widely scattered infrequent
palisade cells, the chlorencliyma being primarily
composed of mesophyll). TIirinax morrisii
is distinctive in having the chlorencliyma composed
almost entirely of palisade cells. Xbaxial mesopliyll
cells more or less isodiametric, compact. Fibers frequent,
mostly in strands below each surface. Adaxial
strands irregular, with 1-21 (rarely more) fibers
per strand irregularly intruding the chlorenchyma,
these adjacent to the epidermis or separated from
the epidermis by a layer of small hypodermal cells.
Abaxial strands with fewer fibers, sometimes cylindrical,
mostly adjacent to adaxial epidermis. Veiny
mostly in abaxial mesophyll or equidistant from
either surface. Outer parenchymatous sheath complete
around, or interrupted below small veins,
always interrupted both above and below large
veins; sheath-cells cubical, often containing chloroplasts.
Inner sclerotic sheath completely fibrous
around large veins, usually including sclerotic
parenchyma adjacent xylem of small veins. Phloem
of large veins not subdivided, rarely slightly sclerotic.
Transverse com missures usually in abaxial
mesophyll, rather infrequent, wide. Ribs most
prominent adaxially in proximal part of lamina.
Surface layers similar to those of lamina, with few
small hypodermal fibrous strands. Central ground
parenchyma containing several vascular bundles:
those in adaxial ribs with well-developed, somewhat
confluent, fibrous sheaths; those in abaxial
ribs more or less surrounded by a common sclerotic
cylinder. Expansion cells in bands, including
fibrous strands, within each fold of the lamina.
These studies have provided the only positive
means of identifying Thrinax radiata with living
taxa.
Leaf Axis: Epidermis with more or less cubical
cells; walls thickened, cutinized. Stomata occasional,
slightly sunken. Hypodermis I-layered, inconspicuous,
sometimes slightly lignified. Peripheral chlorenchyma
indistinct. Fibrous bundles occasional in
chlorenchyma. Vascular biindles uniformly scattered,
not congested at the periphery; each with a
complete fibrous sheath, usually including 2 wide
metaxylem vessels and an undivided, not sclerotic,
phloem strand.
Cell Inclusions: Stegmata of leaf in short, often
continuous, files mostly adjacent to vascular fibers.
Silica bodies spherical, spinulose, diameter 12
microns; basal wall of silica cell slightly thickened.
Enlarged cells filled with raphid bundles are fairly
common between the smaller veins.
T’ascular Elements: Vessels: elements of leaf with
sclariform perforation plates on oblique end walls.
Sieue tzibes: elements of leaf with compound sieve
plates on more or less transverse end walls. It was
noted by Partliasarathy (1968) that the Thrinax
alliance and Licztala are exceptional among coryplioid
palms in possessing more arecoid than typically
coryphoid sieve plates. Furthermore, he stated
that Thrinax in common with other closely allied
genera is also set apart from most coryphoid palms
by having but a single phloem strand, and sclerotic
parenchyma cells in the metaphloem of many vascular
bundles.
Literntiire: The greatest criticism concerning
studies by earlier workers is that most of the material
was Lunreliablj identified, as mentioned by
Tomlinson (1961). Furthermore the material was
either obtained from botanical gardens which are
questionably reliable, unless vouchered materials
are available at the institution, or the materials
were obtained from herbarium specimens Tvitliout
citation of specimen numbers or collector. Thus
the unvouchered materials used by past workers
cannot now be identified with certainty.
Tomlinson (1961:296) cited as his material, ‘IT.
argenten Lodd. -Kew. Leaf,” which is probably a
Coccothrinax; and “T. tessellata Becc. Jamaica.
Leaf,” which name is now in synonymy of Thrinax
parviflora Sw. Drabble (1904) treated T. excelsn
Lodd. and T. pa?vipora Sw., the former name having
been applied to T. radiata in Jamaica and the
latter to T. radiata in Florida. Gillain (1900) treated
T. argentea Lodd. and T. gTaminifolia IVendl.,
the first a Coccothrinax and the second a nomen
incertum. Jost (1887) and Zawada (1890) simply
cited “Tlirinax sp.” And of the nine taxa attributed
to Thrinax in the study by Pfister (1892) only “T.
exceZsn Lodd.” could be a Thrinax but certainly
44 SMITHSONIAN COK?'RIBUTIONS TO BOTANS
not of that species. Rudolph (1911) and Solereder
and Meyer (1928) worked only with species of
Coccothrinax.
Cytology 10
The species of Thiinax in Jamaica, in addition
to many other similarities already mentioned, have
a chromosome complement similar in number antl
morphology to other members of the alliance. The
results of the present study when compared with
previous work (Read, 1963, 1964, 1965a, b) lead
to the conclusion that, as far as can be determined,
it is not pissible to distinguish between genera ot
tlie Thrinax alliance by means of their chromosome
karyotjpes. The relatively large size, 1-3.0
(-3.5) microns in length, of the cliromosomes places
Thrinax, together with its closest allies, in a group
of palm genera having chromosomes of a similar
size but which differ primarily in having a 2-
whorled perianth and a 1- to 3-carpellate ovary
(i.e,, Sclzippia, Ciyosophila, T r i t h i n a x , and Chely-
O C G ~ ~ I I S ) . Except for the genera just mentioned antl
several genera of dioecious palms, which also have
large chromosomes, all other coryphoid palms
studied thus far have mostly very small chromosomes,
0.5 to 1.5 (-2.0) microns long.
The karyotype for each species exhibits an almost
complete gradation of sizes from smallest to
largest with some recognizable differences in morphology.
With such small chromosomes, however,
and with the present technique it has proved unsatisfactory
to attempt a classification of size or
morphological groupings. Chromosome size may
vary considerably among the preparations of a
single species or individual depending on the timing
of pretreatments or fixation and especially tlie
degree of squashing. And although certain cliromosomes
are recognizable as having subterminal, submedian
or median constrictions, it is extremely
difficult to differentiate between real and apparent
differences. For example with such small cliromosomes,
as so often happens, they do not allvays lie
flat or parallel with the level of resolution. It is
then possible to misinterpret a chromosome which
1°From work related to a project supported by National
Science Foundation grants G-18770 and extension GB-13.54.
IVhile in Jamaica the Botany Department of the I:.IV.I. acquired
phase-contrast attachment5 especially for use by the
author in his research on palm cytology.
may have part of its arms bent down or under;
thus a submedian or subterminal primary constriction
may appear as a median or submedian respectively.
An ideal preparation would require squashing
to less than 0.5 micron in order to flatten the
chromosomes sufficiently. This has not been possible
because of the presence of the pollen grains
in the preparation.
MATERIAL AND METHODS.-A~~ material used in tlie
present study was collected from wild plants in
the field in Jamaica.
4 detailed description ot the technique has already
been published (Read, 1965), therefore only
a brief review need be given here. Portions of inflorescence
were collected at antliesis or as soon
before as possible. The pollen was collected by
allowing it to shed on a piece of paper in a draft
free room. It was then stored in gelatin capsules
and placed in an airtight container with dry silica
gel and a color indicator, at room temperature
(about 26°C). It was necessary at times to dissect
the unopened anthers from the flowers and permit
them to dry in the gelatin capsules and silica
gel. It was important not to store more pollen in a
capsule than would effectively coat the walls when
dry, otherwise the possibility of fungal and bacterial
contamination became serious.
Pollen was sown on a nutrient medium (100 ppm
H,BO, in water; 0.02% colchicine in distilled
water; 10% lactose or sucrose and 10% gelatin) for
tlie study of pollen-tube mitosis. After heating
medium slightly it was applied to cover glasses by
means of a small cotton swab. Tlie pollen was
dusted lightly over the surface of the dry medium
and the cover glass then overturned onto a Van
Tiegliem cell lined with moist absorbent paper. ,4n
improved method was to place the rings that form
tlie Van Tieghem cells directly on a layer of moist
absorbent paper lining the bottom of a plastic
tray. After the prepared cover glasses were in place
the container was closed to prevent drying. Tlie
tray was maintained at room temperature (about
26°C) and germination was usually found to have
proceeded sufficiently within 8 to 10 hours.
When examination of sample slides indicated
that mitosis was in process, the cover glasses were
stained with acetocarmine and made permanent by
the vapor transfer method. All counts and drawings
Tvere made on freshly prepared slides. The tapping
and differential squashing necessary to produce the
X'UJIBER 19
FIGURE 32.-Photomicrographs of the chromosomes of species of Thritiax at pollen tube mitosis:
A, R, T. pnrviflora, Read 1657; c, T. pnrr?iporn, Read 1675~; D, T. excelsn, Read 1625. (X ca. 2500.)
45
46 SMITHSONIAS CONTRIBUTIONS TO BOTASY
FIGURE 33.-Photomicrographs of the chromosomes of species of Thyinax at pollen tube mitosis:
A, B, T. radiata, Read 1671; c, D, T. movjsii, Read 143ia. (X ca. 2500.) Note the chromosomes
obviously stiH within the developing pollen tube in Figures B, C.
KUJIBER 19 47
best figures was not possible after a few hours.
Drawings were made with the aid of a camera
lucida, and the photographs were made on 35 mm
film either in black and white or color (Kodachrome
11). The use of phase-contrast aided immeasurably
in distinguishing between closely situated
chromosomes and also produced photographs
of better quality.
Before the present study was undertaken, reports
of chromosome numbers in Thrinax were already
in the literature; however, as a result of the present
revision of the nomenclature among the Jamaican
taxa, it is necessary to revise the names cited for
the species for which chromosome numbers have
been reported. Thrinax excelsa as reported by
Venkatasubban (1945) and Read (1963) should be
named T. iadiata. Thrinax parvifloia and T. floridana
as reported by Read (1964, 1965a respectively)
should likewise be named T. mdiata. Thrinax
miciocarpa as reported by Read (1963, 1964, 1965a)
should be named T. morrisii.
Thrinax radiata has been reexamined using
native material from its natural habitat. Authentic
T. excelsa also lias been studied from material
collected in the field. It was particularly interesting,
early in the research, to compare the chromosomes
of T. parviflora from plants growing in the extreme
climates in which the two biotypes occurs. The
chromosome number in each case lias been determined
to be n=18, agreeing with all other Cory
phoid palms studied to the present. Idiograms of
each species in Jamaica are illustrated in Figure 34.
Thrinax pnrviflora from a lowland habitat at
Cockpit, Clarendon, is represented by a, b, and the
high elevation phenotype from plants on the Devils
Backbone is represented by c, d, e. Thrinax ?-adinfa
is represented by idiograms f, g, and T. excelsa by
idiograms h, i.
FIGURE 34.-Idiograms of Thrinax showing their cornparati\ e karyot) pes. a, b, T. parvipora,
Read 167%; c-e, T. patvzflota, Read 1736; f, g, T. radmta, Read 1671; h, 2, T. eucelsa, Read
1625.
48 SMITHSONIAS CONTRIBUTIONS TO BOTAKY
Phytomorphosis and Zoomorphosis in
Thrinax parviflora
The specific epithet “tessellata” was applied by
Beccari (1907:Zil) to a specimen from Hollymount
at the upper elevations on Mt. Diablo. The name
was given in allusion to the fruit which Beccari
described as having the “epicarpio distincte suberosotessellato,”
a condition common in the cool
moist environments where a fungus infection radically
alters the exocarp of the fruit producing the
condition shown in Figure 2 6 ~ . The fruit can become
infected at almost any stage of development.
Figure 2 6 ~ shows two stages in the development of
the fungiis on young fruits. The fruit on the right
has a small area where the epidermis has only
begun to erupt as a result of the fungal activity.
The fruit on the left has no trace of epidermis
remaining as in Figure 26c. Fruits exhibiting both
partial and complete alteration of the epidermis
are part of the isotype of T. tessellata, herein treated
as synonymous with T. parviporcr. The fungus,
although observed in situ, has not yet been isolated
and identified.
In the same environment where fungal damage
is common on tlie fruit, the branches of tlie inflorescence
also undergo a dramatic change. Extensive
grazing by caterpillars causes considerable
damage to the epidermis and outer tissues of the
branches (Figure 35). So extensive is the damage in
some areas, it is seldom possible to find healthy or
uninutilated epidermal tissue anywhere on the inflorescence
branches. The grazing creates a distinctive
pattern of trenches, exposing the fibrovascular
bundles, intersected with narrow ridges of ungrazed
tissue in Figure 35B. At the lower elevations only
occasional damage by larval grazing is encountered
on specimens and this is usually of minor importance
(Figure 35A). At the higher elevations, however,
it is seldom possible to determine the nature
of the epidermis (e.g., degree of puberulence), or
for that matter even to find ungrazed epidermis at
all on fruiting material.
The apex of tlie petiole in the region of the
abaxial hastula and infrequently even the adaxial
hastula are often grazed extensively. Often the
abaxial liastular region of the youngest leaf on a
plant is completely grazed of epidermal tissue exposing
a network of fibrovascular bundles.
Breeding System
The breeding system is very difficult to understand
completely, primarily as a result of the difficulties
of observing anthesis in the field. Individual
plants may bear as many as 6 to 10 inflorescences
each year and at one time. The most conspicuous
feature of an inflorescence at anthesis is the predominance
of anthers. Lacking conspicuous and
enclosing perianth parts, tlie anthers are exposed
immediately as the primary branches are extended
beyond the primary bracts. The anthers in turn
obscure the pistil. Each inflorescence may have
between 7 and 21 primary branches. These decrease
in size and the number of secondary branches
successively toward the apex. Those at the apex
may have only 6 to 10 secondary branches. The
primary branches below the apex may possess from
I7 to 47 secondary branches and each of these may
possible for each plant to prOdLlce between 100,000
and 300,000 flowers over a 24-hour period. Consid-
FIGURE 35-Grazing by caterpillars has pioduced vai)ing have fro111 about 40 to 115 flowers. Therefore it is
amounts of damage to the infloiescence branches of Tflit?lax
parzupoia. Vasculai bundles ale exposed in ’‘13.’’
SUXIBER 19 49
ering also that each flower may possess froin 5 to 15
stamens, each of which produces an abundance of
light dry pollen, it can easily be seen that the production
of pollen by a single plant is phenomenal
indeed. The anthers dehisce almost immediately
as the sun rises and atmospheric moisture decreases.
Normally all parts of the inflorescence are covered
with scattered pollen very shortly after anthesis.
When an inflorescence was permitted to reach
anthesis overnight in a closed office, all surhces,
books, and objects in tlie room were visibly coated
with pollen within an hour.
The light, dry pollen of Thi-inax is conceivably
carried over very long distances by the wind and
has been determined to remain viable for periods
up to 3 or 4 months at room temperature (about
26°C). Pollen of T. fiarvifloia collected on 12 January
1967 was maintained dry in gelatin capsules
that were kept in containers of silica gel with a
color indicator at room temperature. Samples of
this pollen were sown and germinated on the medium
used in the cytological studies at intervals
over a 4-month period. Pollen viability was excellent
at first (between 70 and 80 percent) on 15
January. It dropped to 41 percent by 9 February,
and on 27 February had further dropped to 37
percent. On 1 April viability was only 1 percent.
It is conceivable then, considering the great quantity
of pollen produced by a single plant at one
time, that crossing need not be a rare occurrence.
4 n attempt was made to test inhibition to selfing.
Pollen froin two different plants was to be placed
on the stigmas of emasculated flowers of one 01 the
plants. A difference in the rate of pollen germination
(determined by fixing and staining an hour
or so after pollination) favoring the outcrossed
flowers would indicate a resistance to selfing. If
this were the case, pollen from nearby plants
would conceivably be able to germinate and coinpete
rvith the mass of pollen from tlie flower’s oivn
anthers.
Attempts at emasculation of flowers on fully
developed inflorescences in the field proved futile.
Within seconds following removal of the enclosing
bracts the anthers clehisced and scattered pollen
over the pistils. Further attempts were made on
inflorescences removed and placed in containers of
water in the laboratory. It was possible to remove
the unopened anthers by spraying thein with water.
It was observed, however, that pollination was effectively
prevented by the lateral compression of
the stigmatic surfaces. Before pollinaton could be
achieved with certainty tlie flowers fell without
having exposed their stigmatic surfaces.
Other attempts were made using inflorescences
still attached to the crown of a plant which was
cut at the apex of the caudex and removed to the
laboratory. hIost of the leaves were removed and
the cut portion was placed in a container of water.
Several fully developed inflorescences proceeded to
reach anthesis during the following night in the
laboratory, but again the flolvers fell without having
exposed their stigmatic surfaces.
Further observations on plants in the field confirmed
that tlie stigmas are tightly compressed at
anthesis and follow roughly the following schedule:
FIGURE 3G.-Flower of Thrinnx pari1i//orn. Photographed with
a Contaflex camera, hand held on a binocular-dissecting
microscope and using Kodachrome I1 film exposed for 10
seconds. Kote the form of the stigmatic region.
50 S\fITHSOhIIAN COh 1 RIBUTIOUS TO BOTANY
21 Sept. 7:30 p.hi.
22 Sept. 8:OO A.M.
A few rachillae and flowers exposed.
All rachillae and flowers exposed; flowers
creamy-white including ovary; anthers
dehisced; stigmas tightly compressed.
22 Sept. 8:OO P.M. Pollen covering all infloyescence parts;
stigmas still compressed.
23 Sept. 9:30 A.M. Ovary yellow, conspicuous; quantities of
pollen still present; stigmas open,
funnellike.
That this is not always the situation is borne out
by pickled material in which the anthers ale not
fully tlehisced while the stigma is fully open and
by the photograph (Figure 36) showing a flower at
anthesis with its stigma fully open and apparently
receptive.
The fact that the stigmatic surtaces are often
closed during anthesis, effectively preventing selfing,
permits a period during lvhich pollen from surrounding
plants can mix either in the air or on
the inflorescence before receptivity of the stigma
occurs. Thus although the chances for selfing are
far greater, outcrossing is not ruled out altogether.
Success of the species seems, however, to depend on
the fact that, lacking pollen from another population
or another plant, seed production is assured
in quantity by selfing, as evidenced by the excellent
fruit production of solitary individuals in the field
or in cultivation.
Flowering in Jamaica: In January 1966, several
plants of T. parvifloi-a on Mango Tree Hill, in
Trelawny, were observed to have inflorescences
exposed in bud for about 20 cm. Subsequent and
frequent trips to the area found the inflorescence
unchanged until 14 May when the same inflorescences
were at a stage one or two days past anthesis.
At Cockpit and Palmers Cross in Clarendon the
entire population was at full anthesis on 9 June,
while all the plants on Long Mountain, St. Andrew,
had flowered a month earlier on 8 May.
It was observed, in 1966, that most inflorescences
of a single plant reached anthesis within a 48-hour
period. This is notable considering that the 4 to 6
or more inflorescences reaching anthesis at the same
time had developed in the axils of successively produced
leaves over a 12-month period two seasons
earlier. Apparently each solitary inflorescence produced
within each leaf axil remains at a particular
stage until several reach the same stage. They then
proceed development together and reach anthesis
at the same time. Each plant presumably has 3
seasons of inflorescence production at any one time.
The 3 or more older lea\es subtend 3 or more spent
inflorescences; the 3 or more younger leaves subtend
inflorescences at or past anthesis, or in fruit;
and, apically, 3 or more youngest leaves subtend
resting or developing immature inflorescences.
Each season a select number reaches anthesis at
once, during which time the next season’s inflorescences
will be reaching a certain stage of equal
development.
In Table 1 the localities, within each subspecies
grouping, are arranged in descending elevation from
the top toward the bottom. It can be Seen that in
any one year, widely separated populations flo~v-
TABLF 1.-Flowering and fririting dates for Thrinax
parviflora arranged by phenotypic expression (localities
27-30 represent subspecies parvipora; 33-29
represent subspecies pubertila)
I.ocality
nti m her
27
10
8
I
5
32
31
30
33
2
14b
38
34
3
36
17
40
26
42
21
4
22
24
29
An thesis
Jan 1967
Mar 1967
Jan 1967
Jan 1967
Jan 1960
Jan 1967
Nov 1964
May 1966
Jul 1967
May 1966
Jul 1967
Apr 1963
May 1966
Jun 1967
Sep 1967
May 1966
May 1966
May 1966
Jun 1966
Jim 1966
Jun 1966
Jim 1967
Immature
fruit
Feb 1955
Mar 1967
Jan 1958
Jan 1956
Apr 1955
Mar 1960
Aug 1955
Jill 1966
Feb 1967
Aug 1966
Mature
fruit
Jun 1966
4pr 1966
May 1966
Nov 1966
Nov 1965
Nov 1955
Oct 1955
Aug 1955
Nov 1965
Nov 1955
Nov 1965
SUhIBER 19 51
ered several weeks to a month apart. This was especially
true in 1966.
A plant removed from a hill near Silver Spring,
Hanover, and planted in the palm collection at
hlona, produced an inflorescence, antl reached anthesis
on 26 May, but the pollen was sterile. Additional
plants were collected at Cliarlton, TVintlsor
Cave, and Negril Hill on 4, 5, and 6 June respectively.
These Jvere a week or more past anthesis,
when collected, providing some overlap in flowering
among widely spaced populations. Nearly all
populations of tlie warm, dry, lower elevations
leached anthesis between 8 hlay antl 9 June 1966.
During the first half of 1966, no plants \\‘ere found
at antliesis or even postanthesis anywhere in the
higher elevations. The presence of mature fruit on
29 June at Hollymount Hill suggested that the
plants may have bloomed during tlie winter.
In 1967 several individuals from throughout the
higher elevations were collected at or near antliesis
between January and March. Flowering and seed
production must be very irregular and rarely plentiful
among tlie plants at the higher elevations,
for after having collected tlie fev available inflorescences
for herbarium specimens antl stud), no fruiting
specimens have since been found among tlie
thousands of plants observed. In January on tlie
Devils Backbone one of the plants Tvitli fully developed
inflorescences was cut below the crown
and removed to the laboratory. At 9:00 P.M. on
12 January not one flower was exposed, but by
8:OO the following morning all piiinai) branches
were fully exposed and all flou ers were shedding
pollen.
Although most of tlie lowland plants have been
observed to flower during the summer months, a
single plant of T. paivipoin near Portland Point
was found at postanthesis in Febiuary 1967. It is
not known if fruit set as a result, because no fruit
was found later in the year.
TVidespread drought and widely scattered, irregular,
late, and below normal rainfall in 1967 no
doubt had a strong effect on the flowering of
Thyinax paiviporn, because the pattern of flowering
at the lower elevations in 1967 was markedly
different from that described for 1966. A trip to
hfango Tree Hill in late May, where all plants
hatl finished bloom by the same (late in 1966,
yielded not a single inflorescence. Kone were exposed
for more than a few centimeters. Two plants
on Long hlountain, which flowered in early May
in 1966, flowered on 22 and 23 June 1967. Several
other plants on Long Mountain hatl inflorescences
exposed, but they did not reach anthesis until late
September, several weeks following a period of
heavy rain which broke tlie 8-niontli drought.
At Cockpit and Palmers Cross, Clarentlon, not
one inflorescence was observed on 24 June 196i
when nearly every plant in the Portland Ridge
area, a few miles to the south, was at full antliesis.
Genetic exchange between distant populations of
T. pavuifloia by means ol pollen exchange is probably
very low. Since nearly all plants in a popiilation
shed their pollen during a 48-liour period, and
since populations widely removed from one another
may flower a week or more a p t , it is unlikely that
pollen from one population can be cari ied through
wind and rain and still be able to compete a -tveek
later with the greater mass of pollen already available
on an inflorescence in another population.
Except for this unlikely but possible exchange of
pollen, genetic exchange between local populations
niost likely results from the indiscriminate distribution
of seed by birds and the resulting plants
crossing with their neighbors in tlie new locality.
Plants of T. Iadiatn have been observed in flower
or fruit tliroiighout the year. Most of the plants in
a given population are usually at the same stage of
bloom or fruit, but there seems to be no definite
season. Thrinax excelsa has been observed in flower
from .4pril to September, and fruit has been collected
from July through September.
Seed Divt~ibzctioiz: Doves and pigeons are common
throughout the range of Thyinax in Tainaica,
antl have been reported by the country people as
niost plentiful when the “thatch” is in fruit. It
was said of tlie hills near Silver Spring, Hanovel,
that, “Bald Plates” [sic] are seen in flocks feeding
on the thatch “peas.” At the time of my visit to tlie
‘ires, the palms were not in fruit and no “Bald
Plates” were in evidence. There are eleven species
of dokes or pigeons indigenous to Jamaica and the
Baldpate or TYliite Crowned Pigeon is coinnion
throughout the distribution range of T. Tadiota
in the northern Caribbean. Several of the others
are common to Jamaica and neighboring islands
but the Baldpate is said to move freely among the
islands (Bond, 1961).
52 SSIITHSOXIAN COSTRIBUTIONS TO BOT’ASY
Variation of Thrinax parvijlora plants, the true extent of genetic differentiation
Thyinax parviflora is a highly variable and polymorphic
species and occurs over an equally variable
and complex system of environments from near sea
level to about 914 in (3000 f t ) elevation. Considerable
effort has been expended in attempting to
correlate variability of phenotype with environment.
Lacking experimental evidence, holvever, of
any genetic bases for the variation, it is not possible
to do more than describe the range of variation in
the various environmental zones. Until such a time
as plants from the different areas are grown under
identical conditions and/or with reciprocal transcannot
be known.
The extent of variability can be and has been
measured by field studies on plants in 42 localities
throughout the entire range of distribution of the
species, and in nearly every ecological situation
over a 2-year period.
In the t\vo extremes of climatic conditions there
are what appear to be two distinct biotypes which,
although lacking really significant qualitative diflerences,
look quite different as a result of the combined
effect of a number of apparently minor
quantitative differences. Two localities, each hav-
FIGURE 37.-Tlirinux parviflora: A, Subspecies parvipora
growing at “R”-ridge above Charlton, Mt. Diablo (the
late Mr. L. XVynter assisted in the collection of palms while
I I\ as i l l Jamaica); 13, subspecies finbe~rt2n growing near
Beverly Hills, Long hlountain, Kingston.
KIJlIBER 19 53
ing a large number of individuals in a nearly uniform
environment and representing as nearly as
possible (1) the hottest, driest, lowest elevation and
( 2 ) the coolest, wettest, highest elevation, were selected
for intensive study of the widest range of
variation in the species. Each population was selected
primarily for its ecological situation and for
having an adequate sample of individuals in a
relatively accessible area. See the section on Morphology
for an explanation of the method of
selection of plants and parts for measurement and
study.
It will be demonstrated under the section on the
typification of Thrinax parzrifZo~a that the characters
of Thrinax, at the higher elevations of Mt.
Diablo, most closely correspond with characters
drawn from the type specimen and original diagnosis
of the species. These plants, typified by a
large sample taken at collection locality #lo on
Hollymount Hill at 792.4 ni (2600 ft) elevation,
represent one extreme in the overall variation of
the species. Figure 37A illustrates the typical “pinwheel”
form of the leaf and the open crown on tall
slender trunks. In addition the very short, blunt or
acutely pointed hastula and the relative absence of
an abaxial hastula, glabrous to glabrate (Figure
~ O C , D , E ) inflorescence branches, generally short pedicellate
flowers or fruits, and short anthers are characteristic
of the plants in this locality. Inflorescences
are shorter than the petioles, erect except in fruit,
and the lowermost branches of the lower primary
branches are occasionally branched more than once.
Flowering is very irregular but occurs between
December and March.
Also characteristic of all plants at the higher
elevations are certain biotic changes of the inflorescence
and fruit (described under Phytomorphosis
and Zoomorphosis in Thyinax paruiflom). These
changes have led to considerable confusion and
misinterpretation. In fact one specific epithet, now
in synonymy, is descriptive of the change in the
character of the fruit, and the plants to which it
applied have been suggested as having some significance
in evolutionary theory. In a discussion of
the durian theory in relation to palms, Corner
(1966: 180) wrote: “Perhaps in this durian complex,
there should be added Thrinax tesselatus [sic] of
Jamaica, the fruits of which appear to have small
flattened warts as in Pholidocaipus.”
In contrast with the Hollymount Hill “biotype”
the plants of the lower elevations, below 365.76 m
(1200 f t ) from the base of Mt. Diablo south to
Portland Ridge, and typified by collection locality
#22 at Cockpit, Clarendon, have quite a different
aspect somewhat resembling the plant in Figure
3 7 ~ . The heavy crown, with relatively broad leaves
. A
F
FIGURE 38.-~-c, Frequency distribution of anther lengths; D-F, range of pedicel lengths. A, D
are Thrinnx pnruifiora from Devils Backbone; B, E are from the Swartz t)pe specimen; and c, F
are from T. rndintn hlorant Point, Jamaica. (Histograms are based on 100 measurements; bar
graphs are based on 29 measurements. Measurements in mm.)
54 SMITHSONIAN CONTRIBUTIONS TO BOTANY
and slightly drooping segments, supported on a
short trunk is typical. The adaxial hastula is prominent,
long pointed, and often tubular when the leaf
is fully expanded. The abaxial hastula is also
conspicuous, frequently forming a triangular projection
up to 1.6 cm long. The arching inflorescences
equal or exceed the petiole in length, and
all branching parts including the elongate pedicels
are densely puberulous (Figure ~ O A , B ) .
The length of the pedicels, in flower or fruit, and
the length of the anthers have frequently been
used as key characters for distinguishing between
taxa. These characteristics have been given some
emphasis in this study to test their usefulness and
to emphasize the danger inherent in the use of
small samples such as are available on herbarium
specimens of palms. It can readily be seen in the
histograms in Figures 19, 38 that pedicel length is
certainly not a good character.
Figure 38~,c illustrates the variability of anther
length in Thrinux parviflora and T. m d i a t a . If only
two localities are considered, the character might
appear useful; however, plants with very short or
very long anthers occur in the same population.
This situation occurs in nearly all the characters
tested. Given only two extreme localities, which
were the basis of Beccari’s work, because the): represent
classical collecting areas, two quite distinct
taxa are the result. Additional collections however,
lrom many localities over the entire range of the
species throughout the island fill in the gaps and
blend the overlapping measurements, resulting in a
kind of clinal variation of nearly all characters.
Accumulation of mutations expressed as certain of
the extreme variations, with subsequent introgression,
might be achieved as a result of the fact that
the highland biotype flowers from November to
March, while the lowland plants flowered from
April through September.
Throughout the range of T. pa~vif7o~-cr, definite
trends in leaf form and trunk size exist which appear
to follow certain gradual changes in climatic conditions.
The most significant variations in the physical
environment that appear to influence the phenotype
are changes of moisture availability in the
areas of different rainfall and of temperature with
differences in elevation. Consistent east-west trends
may also be due to combinations of these factors.
Several features such as branching and puberulence
of the inflorescence rachillae and hastula form,
although different at the two principal environmental
zones, seem to vary little throughout the
range of the species and appear to be characteristic
of particular ecotypes.
Caudex: The caudex diameter and height of
mature individuals are extremely variable even
within a small local population, but the largest
mean diameters and shortest trunks occur in the
hot, dry, exposed areas of St. Andrew, St. Catherine,
and Clarendon. Plants with trunks having
smaller mean diameters are found in regions of
greater rainfall toward the west and in the cool
moist gorges (i.e., Cane River, Rio Cobre), or at
the median elevations of the hills or Cockpit Country.
The very tall plants of smallest trunk diameters
are found only at the cool, moist, high elevations
or in exceptional areas of unusual climatic
FIGURE 39.-Polymorphic leaves of Thrinnx pnrviporo: G. R.
Proctor, of the Institute of Jamaica, holds a leaf from a
juvenile plant in hi5 right hand, and a leaf from a flowering
sire plant in his left hand; near Ramgoat Cave, in the Cockpit
country of Jamaica.
SUhIBEK 19 55
conditions such as that described under Donkey
Trail and “R”-ridge.
Leaf Polymorphism: Among the plants of Thrinax
pawifiora growing at the higher elevations, there is
a striking change in leaf morphology from seedling
to maturity and eventual exposure of the crown at
tlie canopy. The seedling stages up to the formation
of the trunk are typical for the genus. The change
in leaf form, however, subsequent to the first appearance
of a trunk is possibly unique among fanleaved
palms (Figure 39).
The leaves produced at tlie time of trunk formation
are circular in outline and rather flat or
slightly funnel shaped with tlie free part of the
segments rigid. As the trunk elongates the leaves
attain their greatest diameter, have a broad palman,
and the segments arch slightly. With increased
age oE the plant the palman is slightly
smaller and the longer segments tend to droop.
When the plant reaches maturity the leaves appear
to be much smaller as a result of the much reduced
palman and the fact that the drooping segments
begin to curl.
Fdly mature plants in tlie full dense shade below
the canopy may have leaves with a fairly broad flat
palman and drooping curled segments such as are
illustrated in Figure Sin; however, when the plants
are exposed either on a cliff face or above the canopy
they assume a strikingly different form. The
basal portion of the blade folds up and forward,
the palman is very small and uneven, and the
segments are twisted and tightly coiled (Figures
8, 39). At this latter stage the blade often appears
as a “pinwheel” on an excessively long slender
petiole. The petiole in fact is shorter on older
plants, but the much contorted blade creates an
illusion of a longer petiole.
Vm-iabilify of the Leaf Elsewhew: The form of
the leaf blade throughout the range of tlie species
is the most highly variable characteristic of the
entire complex of variation, reflecting in effect different
phases of the polymorphism exhibited tluring
the development of plants at the highest
elevations. The change in shape or form is in fact
simply the result of quantitative changes in the
component parts of the leaf. The form ol the leaf,
which must not be confused with outline, is three
dimensional resulting from varying degrees of folding,
twisting, and curling of the palman and segments.
The overall trend appears to be a decrease
in siLe of parts with maturity, aging, increased
height and exposure and with change from hot and
dry to wet and cooler climates.
The change in leaf form during maturation of
individuals, throughout the greater portion of the
population, is simply the production of larger
leaves accompanied by increases in the number,
width, and length of segments, and the breadth of
the palman. Mature plants, not yet having reached
the canopy, produce larger, more heavily textured
leaves, juvenile types showing a tendency for the
lower portion of the blade to be folded upward.
In the hot dry plains of Clarentlon near Palmers
Cross the mature leaf blade (Figure 7) has a broad
uneven palman and wide very leathery segments.
Elsewhere in the dry evergreen thicket formations
of St. Andrew, St. Catherine, and Clarendon, the
leaf blade is not quite so leathery, tlie broad palman
is relatively flatter but with the basal part of the
blade folded up and forward (Figure 37n), and the
segments arching or drooping. The leaves of plants
exposed at the canopy or in a clearing exhibit a
narrower palman and increasingly lax segments.
Tall, very old plants have leakes in which the changes
just described are even more strongly expressed. In
certain areas of lower temperatures and/or higher
moisture availability (i.e., Cane River Gorge, Rio
Cobre Gorge, Fern Gully, and westward from
Clarentlon) mature plants have leaves with proportionally
smaller palman and narrower segments
which twist or curl slightly (Figure 37A).
In regions of higher rainfall in tlie west and
northwest of the island such as at Negril Hill,
Silver Spring, Lances Ray, Dolphin Head, and
TYindsor Cave, the species exhibits a markedly reduced
leaf with a narrow uneven palman, and
narrow twisted segments. This form of leaf blade
is expressed to different degrees depending on maturity
or exposure but appears almost intermediate
between the two extremes discussed earlier. During
the early part of the study, plants having the latter
leaf form were classified along with plants of the
highest elevations until it was realized that the
dense puberulence on the inflorescence branches
placed the plants more consistently among those
of the lower elevations. In Cockpit Country, plants
of median elevations (hut on exposed drier locations)
resemble the lowland forms most closely.
In regions of still greater moisture availability
and cooler temperature?, as on the upper levels of
56 SXIITHSONIAN CONTRIBUTIOSS TO BOTASY
Dolphin Head, Somerset, Top Hill, Albion Mountain,
and Mt. Diablo, the leaf blade of mature
plants is smaller and more tightly curled and
twisted in proportion to increased elevation and
exposure at the canopy. The greatest extreme in
reduction in size of leaf parts accompanied by a
great amount of folding, twisting, and curling occurs
on the highest levels of Mt. Diablo where the
plants (Figure 8) exhibiting these characteristics
are very tall and exposed above the canopy.
Hastula: Although quite variable throughout the
range of the species, at the higher elevations the
hastula is generally short and blunt adaxially and
lacking or vestigial on the abaxial surface. This is
one of the few characters of which measurements
do not appreciably overlap.
With the exception of the plants at the highest
elevations, the adaxial hastula is quite prominent,
long pointed, and frequently tubular. The abaxial
hastula is also conspicuous whether as a flap of
tissue projecting from the apex of the petiole or
as a well-defined triangular extension with a retuse
apex. In the sample studied at Cockpit in Clarendon,
a specimen was collected in which the abaxial
hastula was developed into two slender processes
projected up along the undersurface of the blade
for 20 to 30 cm (Figure 13c,~). Although not noticed
at the time the specimens were collected, two
leaves taken from the same plant exhibited the
same anomalous condition, suggesting perhaps that
all leaves produced by the plant possessed similar
processes.
Inflorescence: The usually arching inflorescence
equals or exceeds the petioles among the plants of
lower elevations but varies considerably in length
and aspect at the higher elevations where it may
be erect and much shorter than the petiole.
Of particular interest is the occurrence of p i -
mary branches which are twice branched, a unique
condition so far as is known among members of the
Thrinax alliance. As a rule the individual primary
branches are once branched. At the lower elevations
the ultimate branches are commonly forked but
never as elaborately branched as those at the highest
elevations. Several specimens have been collected
on Aft. Diablo in which the lowermost primary
branches are twice branched producing a secondary
condition resembling the primary branch from
which it originates. These highly branched secondary
branches, however, are subtended by the same
triangular nonsheathing bracteoles as are the simple
unbranched ultimate branches.
Pedicels: One of the key characters used to distinguish
between the “species” of the Thrinax
pU7aifiol-U complex, and among species of other
genera as well, is the length of the floral and fruit
pedicels. As was mentioned earlier the importance
given to the length of the pedicels by the various
workers in the genus has necessitated particular
attention to testing its reliability. The results presented
in graph form in Figures 17 and 19 illustrate
the uselessness of the character in Jamaica
and emphasize the need for adequate samples irom
throughout the entire range of the species complex.
It also points out the danger of basing taxonomic
conclusions on too few characters and in certain
cases upon the wholly inadequate presently available
herbarium specimens.
Samples from 144 plants in 32 different localities
totaling some 12,000 measurements are tabulated
in Figures l i and 19. It can be seen readily that any
attempt to correlate pedicel length with a particular
phenotype or locality is futile. The frequency
histograms (Figure 19), representing only three
localities, alone show so much overlap at the higher
frequencies that one can conclude the shortest
pedicels (lo occur at Hollymount Hill and the longest
occur at Cockpit, but these extreme areas share
the identical lengths with such great frequency as
to be of hardly any taxonomic significance. TVhen
all 32 localities are considered together as in Figure
l i the ranges and means of variation overlap considerably.
They show a general tendency to increase
gradually from high to low elevations.
Pubel ulence: The entire population of Thi innx
pawiflora is divided into two fairly distinct expressions
of puberulence on inflorescence branches.
This character, never having been mentioned in
any study of the genus, provides the only fairly
consistent means, apart from the hastula, for separating
the species into two biotypes. Throughout
the range of distribution of the species, with the
exceptions to be mentioned below, all parts of the
inflorescence branches are densely puberulous. This,
as far as can be ascertained, is another unique character
in the Thrinux alliance; certainly no other
members of the genus have been reported to have
puberulent rachillae. Although the presence of
puberulence can sometimes be determined by the
naked eye, it is usually necessary to use a lens of
SUMBER 19 57
at least X 8. The degree of puberulence does not
vary much at the lower elevations (Figure ~ O A , B ) .
On the upper peaks and ridges where cool mists
and clouds prevail (e.g., Hollymount Hill, Union
Hill, Grier Mount, Devils Backbone, Dolphin
Head, Somerset, Stirling Castle, Ramgoat Cave,
etc.), there is a striking change in the puberulence.
The plants here produce inflorescences with apparently
glabrous (Figure 4 0 ~ ) or glabrate (Figure
4 0 ~ ) branches. Frequently the puberulence is limited
to widely scattered hairs in lines or patches
usually on the rachis or peduncle. It is doubtful
that any plants produce inflorescences which have
truly glabrous branches, but it is difficult to see the
hairs at anthesis and pollen grains can easily be
mistaken for puberulence. Occasionally inflores-
FIGURE IO.-Different degrees of puberulence exhibited by inflorescences of Thrinax parvifiora,
ultimate branches: A, Cockpit, Clarendon; 13, Spring Garden, St. Catherine; c, Devils Backbone;
D, E, Hollymount Hill.
58
0
LOCALITY #
4
5.0
52
5k
.
22
4 1
,.A. .. i . . * = low
4 1
SMITHSONIAN CONTRIBUTIONS TO BOTANY
15 14 13 12109 8 7 5
. . . . . . . . . . . . . . . . . . . . . . * .
. . . . . .
. b . 1 . . . . . . . ' . . . . . . . . . . . .
' I
I
\
'%I.-()-, -
457 't
..... $. 304 -
914 -
762 -
*. . *. .-it; 152 - \ir...-. Ir ........... * ..
.and phenotype 0 = high elevation phenotype
1- . May Pen
w p - 1 *.e-
I n ' I 9 8 \ a
W
3 :402,000
Suanish Town
*- A* ST. ANN
'.
FIGURE 41.-Transect across central Jamaica, showing pedicel length and phenotype, from
lowest to highest elevations where Thrinnx parviporn occurs. The large dot in the bar graphs
represents the mean length of the pedicels, the solid line represents 90% of all measurements,
and the broken line indicates 5% of the total taken off each end of the range,
SUXIBER 19
cences of plants in exposed localities produce
branches with considerable puberulence (Figure
40c). Insect grazing on the branches of plants at
the higher elevations often entirely denudes the
epidermis preventing accurate determination.
In certain localities such as at Dolphin Head and
Windsor Cave phenotypically similar plants growing
in the same general area may produce inflorescences
with dense puberulence or almost glabrous
branches. Transplant experiments would be most
helpful to test the effect of climate on the degree of
puberulence.
Transplant Expel iment: An attempt was made
to transplant mature individuals from the highest
elevations into a garden at the University of the
West Indies Botany Department to be grown with
individuals removed from low elevations. The elevation
at Rlona is approximately 600 feet, and the
area frequently suffers from prolonged dry periods.
The primary purpose was to obtain flowering material
for cytological studies dependent on the production
of inflorescences within the year or two
following transplanting. Only one plant produced
inflorescences and this was very soon after having
been transplanted. The flowers produced nonviable
pollen.
Another possible and valuable result of the experiment
would have been the observation of the
degree of puberulousness on the inflorescence
branches of plants removed from the high elevations.
It was strongly suspected that there would be
an increase in puberulousness if the high-elevation
plants formed inflorescences in the hot dry climate
of Mona. Leaf form and pedicel lengths produced
at Mona would also have provided valuable information
about possible genetical differences between
the two apparent biotypes. This experiment has
been terminated as a result of an abnormally long
drought and a severe water shortage from January
to September of 1967. Most of the plants remained
green and in fact produced new leaves during the
growing season of 1966; however, lvatering was
stopped in the spring of 1967, and during two
weeks of particularly hot drying winds, without
rain or water, nearly all leaves rapidly turned
brown and the trunks began to Tvrinkle from slirinkage.
All plants have since died due to the continued
drought and lack of water.
Seeds from many remote areas have been germinated
and the resulting plants ivill be transplanted
in a test garden. The comparative results
of seedling transplants will take many years because
of the extremely slow growth rate characteristic
of palms in this genus. Seedling palms grown
in containers will become established more easily
if root damage is minimized and water is available.
It is the earnest desire of the author to bring together
in one test garden seedlings of all taxa in the
genus for comparative studies at some future date.
TRANSECT FROM PORTLAND RIDGE TO HOLLYMOUNT,
JAMAICA.-A line drawn from Portland Ridge, in
the south, to the northern slopes of Mt. Diablo
(Figure 41) passes through regions of low to high
elevations from sea level to about 914 m (3000 ft),
and from hot, dry nearly xerophytic conditions to
cool, moist mesophytic woodlands and high pinnacles
frequently bathed with mist. The rainfall
varies from less than 126 cm (50 inches) per year at
Portland Ridge to more than I78 cm (70 inches)
per year on the heights of Rlt. Diablo. Along the
line, changes in temperature and moisture availability
are mostly gradual up to the middle elevations
on Mt. Diablo where, as a result of the
combined effect of exposure, elevation, and the
very steep escarpment, the changes are more abrupt.
At the elevation where cloud formation most frequently
occurs, noticeable differences in the climate
and flora are evident over a very short distance.
Thrinax parvipora grows on practically every natural
outcrop of limestone along the transect line.
During the present study numerous collections
were made in nearly every accessible area along an
irregular line froin south to north, and on Rlt.
Diablo, in an attempt to obtain adequate material
over a range of variable ecology. The collections
were also part of an investigation to determine the
distinctions for maintaining two taxa in the region,
one in the lower elevations (T. parvifiola) and one
at the highest elevations (T. tessellata). The principal
means of distinguishing between these two
species had been the supposed absence of pedicels
and the appearance of the fruit in the latter. The
tessellated appearance of the fruit at the higher
elevations, resulting from fungal infection described
earlier, and the length of the fruit pedicels are
poor characters which were based on Tvholly inadequate
material.
The transect runs through two of the three intensive
study areas. Samples of inflorescences were
taken from every available plant at both Cockpit
60 SMITHSONIAK CONTRIBUTIONS TO BOTANY
and Hollymount Hill. In addition pedicel measurements
were taken from every collection locality
along the route of the transect. These measurements
are all compared in Figure 41, and it is obvious
(as has already been shown) from the results
that the length of the pedicels cannot be used for
separating taxa. Although the high-elevation plants
have a predominance of short pedicels, those of
lower elevations exhibit an abundance of pedicels
in the same length ranges, and there is considerable
overlap at the 90 percent level.
Only two apparently reliable characters can be
used to characterize the high- and low-elevation
phenotypes along the transect: the hastula and the
degree of puberulence on the primary branches of
the inflorescence. The designation in Figure 41 for
lowland phenotype represents plants with long,
pointed hastulas and densely puberulous primary
branches, accompanied usually by large leaves and
relatively short and stout caudices. The designation
for high-elevation phenotype represents plants
with short, acute or blunt hastulas and glabrate
(Figure 4 0 ~ ) to lightly puberulent (Figure ~ O C , D )
primary branches.
Additional measurements are included in Figure
18 comparing variability at the three intensive
study areas. Bar graphs giving the range of variation
and means for the various organs of plants
along the transect route may be found in Read
(1968).
Blue Mountain Trail: Along the southern face
of Mt. Diablo an exceedingly steep escarpment
rises nearly 609 m (2000 f t ) above the valley. Along
most of the escarpment, which faces south, climatic
conditions up to about 502.92 m (1650 ft) elevation
are very similar to those at lower elevations. Full
exposure to the sun, particularly in the winter
months, perfect drainage, and average rainfall permit
a dry evergreen, thicket-type vegetation in
which the palms are almost indistinguishable from
plants growing near sea level on Portland Ridge.
At approximately 609 m elevation the vegetation
changes to open woodland in protected valleys and
gorges, and tall, slender plants of the typical highelevation
palm dominate the exposed ridges and
rocky pinnacles. There does not appear to be a
transition zone for the palms along the trail, because
suitable rocky outcrop was not encountered
between 502.92 and 762 m elevation. It is important
to note that the point between 518.16 and 609 m
(1700 and 2000 ft) elevation is the altitude at which
clouds generally form and obscure the peaks and
ridges, commonly throughout the year, but more
frequently in the late summer and winter months.
The clouds and mists common above 609 m elevation
on Mt. Diablo result partly from and contribute
to the higher moisture availability and
cooler temperatures reflected in a different flora.
The greatest extremes in phenotypic expression
are coincidentally found growing on the southfacing
escarpment of Mt. Diablo above an area
known as Charlton. At locality #13, between 457.20
and 502.92 m (1500 and 1650 ft) elevation, several
palms exhibit among themselves the largest trunk
and leaf dimensions, and although two of the
plants in the group had pedicel lengths averaging
within the normal range of means for lowland
lorms, the third plant exhibited exceptionally long
pedicels which caused the mean for the group to be
among the longest pedicel lengths (Figures 17, 41)
of any population on the island.
Near the upper limit of the escarpment, overlooking
the group of exceptional plants just mentioned,
typical high-elevation plants occur between
609 and 807.72 m (2000 and 2650 ft), in the cloud
zone. The very tall, slender-stemmed palms with
very small tightly folded and curled leaf blades
grow among somewhat dwarfed thicket wherein
occur numerous epiphytes, ferns, and herbaceous
plants.
“R”-)idge: ’4 short distance to the west of the
Blue Mountain trail, a spur extends southward
from Mt. Diablo and descends to 457.2 m (1500 ft)
elevation at a pass over which the Ewarton-’IVorthy
Park road crosses. Apparently the spur and ridge
provide a barrier to moisture-laden prevailing
winds from the east, for the appearance of the
vegetation seems to indicate a rapid increase in
moisture availability as one ascends a short distance
along the road to the pass. At approximately 365.76
m (1200 f t ) the road passes from an area of exceedingly
dry, nearly xerophytic rocky outcrop to an
area of mesophytic woodland on an east-facing
slope. Thiinnx parvif?ora exhibiting the lowland
phenotype are common on the slope. X short distance
uphill, but beyond a bend in the road, the
vegetation changes to dense woodland where tall,
slender palms of the high-elevation phenotype are
found. Patches of mesoph) tic woodland alternate
with xerophytic scrub between 396.24 and 457.20 m
NUMBER 19 61
(1300 and 1500 ft) elevation along the road, and it
is possible to observe both phenotypic expressions
of T. parviflora growing a short distance apart but
still apparently in different habitats. Near the pass
at 457.2 m (1500 f t ) the vegetation mostly resembles
moist woodland, which is usually limited to
much higher elevations on Mt. Diablo. The whole
of the ridge area is one of strong ecological contrasts.
On a small limestone and marl ridge, at the
highest elevation along the road, there is a strange
mixture of both high mountain and lowland species
of plants other than palms (e.g., Sloanea jamaicensis
and Bursera simaruba) growing side by side. Here
too was found the unique situation where both
phenotypes of T. parviflora are growing very near
one another. Retaining their peculiar individuality,
the two phenotypes are distinguished by the tightly
congested leaf blade with twisted segments and
more slender trunk of the high-elevation form (Figure
42~,c) as compared with the larger, flat or
simply folded leaf blade with straight or drooping
segments and much heavier trunk of tlie lowland
form (Figure 4 2 ~ ) . In attempting to verify these
apparent differences it was found that the old inflorescences,
taken from plants exhibiting the characteristics
of the high-elevation phenotype, were so
completely damaged by insect grazing that they
were useless for comparison with the densely puberulous
inflorescences taken (collected in young fruit
in July) from plants exhibiting characteristics of
the lowland phenotype. The occurrence of two
different phenotypes (differing it must be stressed
only in quantitative characters which overlap considerably
with plants elsewhere on tlie island) in
the same population, while proving nothing and
raising numerous unanswerable questions, certainly
suggests the existence of two genotypes, at least in
that particular locality.
Among a group of tall, slender plants growing
on a steep slope in a depression behind the ridge,
where even the lowland form attains considerable
height (up to 9 or 10 m), two plants were observed
which might represent intermediates between what
may now be considered two genetic biotypes. Both
FIGURE 42.-Phenotypes of Thrinax paruiflora found growing within a short distance of one
another on “R”-ridge, at about 457 m ele\ation, along the Ewarton-\\’orthy Park road. The
two plants at the left are tjpical of the two extreme forms; the plant on the far right could
possibly be a hybrid between the two biotypes.
62 S.MITHSONIAN CONTRIBUTIONS TO BOTANY
FIGLRE 43.-Region of the Cockpit country showing extent of deep continuous box canyon running
from Barbecue Bottom to Burnt Hill. Adapted from the topographic maps of the
Directorate of Colonial Surveys, D.C.S. 1st edition 1954. Jamaica l/50,000. Further explanation
will be found in the text.
SUXIBER 19 63
of the intermediates exhibited flatter leaf blades,
intermediate in size, with rigid segments (Figure
4 1 ~ ) , and the trunks were intermediate between
plants exhibiting the clearly recognizable phenotypes
of high or low elevations. It should be mentioned
that apart from the area just described it is
almost impossible to be certain about the position
of certain plants growing in exposed situations at
middle altitudes without resort to examination of
the puberulence on the inflorescence branches. In
the population described above both forms appear
most robust and the differences between them appear
somewhat exaggerated.
The area here in question was not found until
very late in the study and, although of great significance,
it is regretted that time and circumstances
did not permit additional field xvork. It remains an
exceedingly interesting area for future concentrated
study when plants of both forms can be compared
by study of inflorescences at anthesis and in fruit.
DONKEY TRAIL TRAKSECT.-In the northeastern
corner of the “Cockpit Country” in Jamaica, a
series of large interconnected cockpits extends four
miles southward along the road from Kinlosr to
within half a mile of Burnt Hill intersection. The
cockpits (indicated by solid shading in Figure 43),
including the one at Barbecue Bottom, vary from
nearly a mile wide in the vicinity of Campbells, to
only 100 or 200 m wide at the southern end. The
formation resembles a “box canyon” with walls
100 to 200 m high throughout its length and opens
to the north where it is exposed to the prevailing
northeasteily winds. Figure 43 illustrates the major
topographic features of the region.
In mountainous regions, increase in elevation is
normally accompanied by a gradual decrease in
temperature, and cloud formation is usually restricted
to the higher peaks and ridges. An anomalous
situation occurs in these cockpits where the
usual climatic conditions are inverted. High ridges
surrounding the southernmost cockpit (Figure 44),
although rising above 548.64 m (1800 ft), support
an open, dry evergreen thicket which resembles
that on the lower slopes of hlt. Diablo. Likeivise,
plants of Thyinax pawipoin growing on the ridge
very closely resemble those of much loTver elevations
elsewhere. As one descends into the cockpit
the air temperature decreases noticeably in a very
short distance and the vegetation changes from dry
evergreen woodland on the upper slopes to wet
forest on the lower slopes and in the bottom
Thrinax pawipora growing inside the cockpit (Figure
44n) exhibits the phenotype usually associated
with the highest and wettest elevations of Mt.
Diablo.
Because extremely different phenotypes were
found growing a short distance from one another,
a transect study was made in order to compile
additional information in support of ecoclinal variation
in Thrinax parviflora. During a 48-hour
period from 25 to 27 May 1967, a climatological
and morphological study was made along a transect
projected through a population of T. paruiflora on
the south side of the cockpit at “R” in Figure 44.
FIGURIC 44.-Donkey Trail special area, enlatged from section
indicated in Figurc 43: A, start of the donkey trail; B, C,
points at which climatological readings were made, Further
explanation in the text.
64 ShIITHSONIAX CONTRIBUTIONS TO BOTANY
e l e v a t i o n i n f e e t
1825
LOW Elevation
1 1740
Phenotype
May
tF
RH
tF
RtI
High Elevation
Phenotype \
1707
1700
1695
1690
1685
RH "? 1500
1400 tF
RH
25th
P.M.
o m
0 3
IA o> .. ..
83 7;
78 7C
87 9E
76 70
89 95
76
89
75
96
26th
, . . . . . . .
N cu m c -
4
81 79 77 76
72 82
80 74
76 89
75 74
91 91
27th
A.M.
I n 0 3 m
L n c - .. ..
71 73
69 72
88 86
59
38
FIGURE 45.-Donkey Trail Cockpit profile (tF = temperature in degrees F; RH = the ielatiie
humidity; o = matuie plants; u = juienile plants; Xertical scale, 1 cm = 10 ft; angle of slope
55-60 degrees).
SUXIBER 19
I,ocnlity
65
T?mfieratzires
i i i a y i - minimum
muin
Flowering
season
The transect originated on the highest ridge at
556.26 m (1825 f t ) elevation ("E" in Figure 44)
and extended down into the cockpit for 47.24 ni
(200 f t ) to 509.02 m (1670 f t j elevation along a
belt 3.05 m on each side of a string stretched between
the two points. The upper 15 'm of the
transect consisted of nearly vertical exposed cliffs
above the donkey trail. Below the pass a scree slope
at an angle of 55" comprised the remaining 32 m.
The donkey trail originates at point "A" in Figure
44, passes over the ridge at 539.49 m (1770 ft), and
continues diagonally across the path of the transect
providing access to tlie stations where climatological
information was gathered (Figure ~ ~ B , c , D , E ) .
A constant recording thermometer i ~ a s pIacet1 at
the highest (most inaccessible) point at 556.26 m
(1825 ft) elevation, near the base of a large palm.
A rnaxiinum-minimum recording thermometer was
placed at the 509.02 m (1670 f t j level. Periodic
recordings of temperature and relative humidity
were made using a sling psychrometer and the percentage
relatiye humidity was computed with the
aid of the Psychorometric Tables compiled by C. F.
1Iarvin.l' The information obtained is recorded in
Figure 45 and is compared with the distribution of
phenotypes on the side of the cockpit.
Mango Tree Hill, one of the original areas of
intensive study, is about a mile to the southwest of
the Donkey Trail study area. The palms mostly
exhibit the lowland phenotype antl closely resembIe
those on the uppermost ridge at the Donkey
Trail area. A maximum-minimum recorded thermometer
was also maintained on Mango Tree Hill
throughout the 48-hour period for comparison rvi th
temperatures recorded on the Donkey Trail ridge
(Table 2).
Although the sample studied at the Donkey Trail
site is small and the period covered by the climatological
data is short, certain significant trends antl
features are thought to be worth noting. A marked
difference in climate between the ridge and slope
of the cockpit is not only evident in the differences
in temperature and percentage of relative humidity,
but is also reflected in the distribution of species
of certain plants. Hohenbeyio xmota, H. polycephalo,
and Tillnndsin sp. aff. fnsciciilatn are plentiful
on the ridge but entirely absent on the slope
'I I'uhlished by the C.S. TVeather Bureau, TVnFhington, D.C.
(1V.n. No. 235, U.S. Department of Commerce, 1041).
Vango Tree Hill
Tiansect 1825'
rianrect 1670'
below. Gzizrnclnia lingulata, ferns, and herbaceous
plants occur on the slopes but are entirely absent
on the ridge. IYere it not for the fact that the sky
was overcast during part of the study, the disparity
in the temperature and relative humidity readings
between the ridge and the slope might have been
even greater. At all times a strong cool breeze blew
through the Donkey Trail pass at 539.49 m (1770
f t ) elevation where the temperature was as much
as 1.68"C lower on the northern side, a distance of
only twenty feet. Dew and fog began to form in the
bottom of tlie cockpit soon after 7:OO p.v. The fog
persisted up to the 487.68 m (1600 f t j level until
after 8:00 A.M., when the sun's ra)s reached tlie
lower levels of the cockpit. During the Ivinter
months when the angle of the siin would be to the
south, longer periods of shade and fog, along with
a greater difference in temperature between the
ridge and the north-facing slope of the cockpit
would be expected. The conditions just described
for the inside of the cockpit resemble conditions
pielailing above 609 m (2000 r t ) elevation on the
>It. Diablo range of mountains.
It can be seen from a study of Table 9 that in
each organ studied size differences occur, but they
are not great enough to be considered important;
however, the more congested and twisted leaf hlatles
and taller, more slender cautlices of the plants in
the cockpit create a distinct impression of two
distinct biotypes. In Table 3 each elevation recorded
represents a single plant along the transect.
Each mearurement, ivith the exception of caudex
height antl diameter, represents the average of four
measurements for each plant. ,411 characters <tudietl,
with the exception of the caudex height, tend to
decrease gradually in size with decrease in elel ation
within the cockpit and could probably be correlated
with the lower temperature and increased humidity.
Although numeroils seedling., of TIii inq1c pn).uifloro
84" F 66" F Vay through August
83" F 50" F \Ia\ through 4ugust
5Cjo I' Gi" F Januai\ thiough March
66
Selected characters
SMITHSONIAN COKTRIBUTIONS TO BOTANY
1825 1815 1810 1750 1740 1707 1700 1695 1690 1686
Petiole length in cm . . . . . . . . . . . . . . . . .
Petiole width at the
sheath in cm . . . . . . . . . . . . . . .
Petiole width at the
a p e x i n c m . . . . . . . . . . . . . . . . . . . . . . .
Hastula (adaxial)
length in cm . . . . . . . . . . . . . . . . .
Hastula (abaxial)
l e n g t h i n c n i . . . . . . . . . .
Leaf segment number . . . . . . . . .
Leaf segment length
in cm . . . . . . . . . . . . . . . . . . . . . . . .
Segment width at the
widest point in cm . . . . . . . . . . . . .
Palman length in cm . . . . . . .
Caudex diameter in cm . . .
Pedicels of fruit,
Degree of puherulence
Caudex height in m . . .
mean length in mm . . . . . .
relates to Figure 40 . .
72 79 143 68 68 75 67 78 72 85
2.1 1.8 2.5 1.3 1.1 1.5 1.2 1.6 1.6 1.7
1.8 1.5 2.0 1.0 1.0 1.1 1.1 1.2 1.2 1.4
2.5 3.4 2.1 0.9 1.3 1.0 0.6 - 0.9 0.8
0.5 0.4 0.3 0.1 0.1 0.1 0.1 0.1 0.1 0.1
54 43 55 41 38 45 44 42 41 44
74.0 62.0 85.0 50.5 51.5 53.0 46.5 50.0 54.0 63.0
4.2 5.3 6.6 3.5 3.2 4.6 4.3 3.8 4.3 3.8
11.2 21.2 32.5 14.5 14.5 14.7 19.5 9.5 14.5 21.2
10.0 8.0 11.0 7.7 6.0 6.5 6.2 6.2 6.1 5.5
5.0 1.4 1.0 3.3 3.6 4.0 4.0 4.0 5.0 5.0
1.3 1.7 - 1.0 0.6 1.0 0.6 1.1 1.1 0.6
B B - - C - D - E E
occur below 502.92 m (1650 ft) elevation, inside
the cockpit mature plants are absent, and below
487.68 m (1600 ft) even seedlings were exceedingly
rare.
Thrinax Swartz
Thrinnx Sw., Prodr. pp. 4, 57. 1788.-Linn., Gen. PI. ed. 8
(Schreb. ed.) 2:772. li9l.-Slv., FI. Ind. Occ. 1:613, t. 13.
1797.-Linn., Gen. PI. ed. 9 (Spreng. ed.) 1:255. 1830.-
Mart., Hist. Nat. Palm. 3:254, pro partc. 1838.-Drndc in
Engler & Plant], Sat. Pflanzenf. 2 (3):34. 1887.-I%aillon,
Hist. PI. 13:317. 1894.--Sarg., Silva 10:49-50, pro partc.
1896; Bot. Gar. 27:83-84. 1899; Silva 14:79-80. 19OP.-Cook,
Bull. Torr. Bot. Club 28:534-535. 1901 .-Becc., TTehhia
2:9-10, 247-250. 1907.-Britt., Torreva 8:240-241. 1908,-
Becc., Annal. Roy. Bot. Gard., Calcutta (hfartelli ed.) 13:
327-336. 1933 [“1931”].-Dalgren, Index Amer. Palms 272-
278. 1936.-Bailey, Gent. Herb. 4: 129-149. 1938; 8:94-99.
1949.
Simpsonin 0. F. Cook, Science, n. s. 85:332-333. 1937 [name
invalid (sine descr. 4at.); type-species: Thrinax microcarpa
Sargent].
Section Porothritiax Drude ex Sargent, Silva 10:49. 1896
[type-species: Thrinax microcarpa Sargent].
TypE-SpEcIEs.-Thrinax pawiflorn Swartz.
Small to medium, solitary, erect, palmate-leaved,
unarmed, hermaphroditic, coryphoid palms; caudex
columnar, tan to gray, smooth or fibrous, obscurel)
ringed with leaf scars, base usually a mass of roots
forming a mound on rock substrate.
Leaves palmate; the sheath fibrous, unarmed, at
first tubular, velutinous, the petiole with a predetermined
abaxial groove below the insertion of the
sheath which splits on expansion of the bud permitting
egress of the inflorescence (at least in
Jamaican taxa), the petiole above the sheath mostl)
flat atlaxiallj, conspicuous, various in outline, abaxially
a ridge, flap, or small projection; blade flabellate,
entirely lacking a central costa, adaxial surface
glabrous, abaxial surface variously lepidote; segments
narrowly rhomboid, trullate, or obtrullate,
fused variously toward the base forming a plicate
palman, mostly bifid at the apex.
Inflorescences interfoliar, elongate, erect to arching,
divided into numerous, pendulous, primar)
branches, these mostly simply branched; lowermost
primary bract bicarinate, inserted about 1 cni above
the base of the peduncle, others tubular with an
oblique aperture, all primary bracts densely allpressed
lepidote, each enclosing the base of the
next higher bract and the peduncle of a primary
SUhIBER 19 67
branch with its bract, the latter bicarinate and
bifid, and partially enclosing the flowers prior to
anthesis, inserted midway along the flattened peduncle
of the primary branch; each ultimate branch
subtended by a narrow triangular bract.
Flowers bisexual, mostly protandrous, solitary
on prominent or very short bracteolate pedicels;
perianth a single series, forming a 6-lobed or dentate
cupule, not enclosing reproductive parts; stamens
(5-)6-12 (-15); filaments usually straight,
slender, mostly free; anthers auriculate basally,
retuse apically, basifixed, locules linear to oblong,
dehiscent introrsely by lateral slits; pistil unilocular,
uniovulate, with a narrow very short style
flaring upward to a laterally compressed infundibuliform
stigma; ovule with a well-developed funicular
aril, erect at the base of the locule.
Fruit small, depressed-globose to globose, white
at maturity, often with persistent apical stigmatic
remains and persistent basal perianth; epidermis
smooth at maturity when fresh; mesocarp thin,
mealy; seed depressed-globose, smooth, tan when
fresh, depressed basally at the hilum; endosperm
homogeneous but intruded partially or completely
through the center from base to apex by the testa;
embryo lateral to subapical.
DISCUSSIOK OF THE GENUS Thrinax.-The name
Thrinax, derived from the Greek word for “trident,’’
was conceived by Linnaeus the younger in a
manuscript before his death in 1783. Olof Swartz
(1788) published the name, accompanied by a 10-
word diagnosis, omitting any reference to Linnaeus.
Considering the brevity of Swartz’s original description
it is noteworthy that the description in the
eighth edition of Linnaeus’ Genera Plantmum published
in 1791 comprises a quarter of a page. TVhen
Swartz published his Flora in 1797, he ackno.ivledget1
that Linnaeus filiits had provided the generic name.
It is evident from a comparison of the Swartz description
of 1797 with the Linnaean manuscript that
much of the former’s information is in fact based
on his own observations in Jamaica. In the ninth
edition of Linnaeus’ Genera Plantarum, editor
Sprengel cited, “L. fil. (1781)” as authority for the
genus. The reason for citing “1781” is not clear
except as it may refer to the unpublished manuscript.
12 Swartz clearly indicated that the flowers
The Linnaeus filius unpublished manuscript was cited
by TV. H. Aiton in the Preface (p vi) of H o ~ t n s Kewensis
in 1789.
were pedicellate, but Sprengel wrote that they were
sessile.
For 42 years the genus comprised but a single
named species. In 1830, three additional species
were published by J. A. & J. H. Schultes: T. argentea,
T. pumilio, and T. radiata. Of these three
species only one remains in the genus, for T. argenlea
has been transferred to the genus Coccothrinax
and T. pumilio became the type for a new genus,
Porothrinnx, a name which cannot be applied (see
Appendix 11). Eight years later hlartius described
two more species, T. multiflora and T. barbadensis,
both of which were later referred to Coccothrinax.
If the numerous nomina nuda are included, about
23 binomials had been published in the genus by
1866. The first attempt to remove discordant elements
from the genus was made in 1866 when
Grisebach listed: “5. Porothrinax piimilio TVendl.
ined. (Thrinax Lodd.); Wr. 3219e:” without a
generic diagnosis but by implication based on the
description of T. pumilio Lodd. ex. J. A. & J. H.
Schultes. Since T. pumilio cannot with any certainty
be applied to any known taxon, it is considered
a nomen incertum. Porotlii inax, based upon
this taxon, was not validly published because it
lacked a separate generic diagnosis at the time of
publication.
In 1887, Otto Drude created three subgenera: Ezithrinax,
Hemithrinax, and Porothi inax. Eu-thrinax
was characterized as having long-pedicellate flowers
with introrsely dehiscing anthers and a vertically
sulcate ruminate albumen in the seed, and comprising
T. radiata, T. parvipora, and T. multifEora.
Subgenus Hemithrinax was characterized by a single
species, H. compncta, which has nearly sessile
or short-pedicelled flowers, sessile anthers with extrorse
dehiscence and seeds with the uniform albumen
penetrated by a broad deep basal cavity.
Porothrinax was typified by P. pztmilio of uncertain
identity.
Although Drude’s treatment was accepted in the
Silva in 1896, Sargent soon recognized that certain
discordant elements could be removed by utilizing
the character of the seed. Thus in 1899:82 he divided
the genus into Coccothrinax, with seeds “vertically
sulcate by the infolding of the testa into the
ruminate albumen,” and Thi innx, the seeds of
which have a uniform albumen more or less deeply
penetrated by a broad basal cavity. At that time no
other sati5tactoi-y characters were known for dis68
SMITHSOXIAN CONTRIBUTIONS TO BOTANY
tinguishing the two genera when seed was unavailable.
For currently reliable characters, see Table 4.
Bailey (1939) did not believe the color of the ripe
fruit to be reliable, but, except for one unconfirmed
report of a white-fruited species of Coccothrinax
and the fact that the fruits of Thrinax
“harrisinna” were erroneously reported to be purplish,
the character holds quite well. Le6n (1956)
wrote that it is possible
to distinguish without hesitation a species of Thrinax from
one of Coccothrinax: in the latter case, the sheath is clearly
cylindrical, commonly formed of more or less rigid fibers
which are strongly interwoven and in every case persistent
on the trunk for many years. . . . In Thrinax, the sheath is
more irregular and not long persistent on the trunk.
Unfortunately Le6n missed the true character of
the Thrinax sheath that provides the most useful
field character: the petiole of Thrinax, as indicated
in Figure 9c, is split in the sheath, which is not so
in Coccothrinax.
At the same time he described Coccothrinax,
Sargent further divided the remaining elements of
Thrinax into two sections: the Pedicellatne, including
the type for the genus and recognized by having
long pedicels, and the Sessiliporae, containing species
with the flowers having the appearance of being
Key to the Species of Thrinax
1. Inflorescence branches exhibiting numerous conspicuous pedicel remains, exceeding 1 mm
in length following fruit set (except occasional specimens of subspecies parvipora of 7.
pnrvipora); adaxial hastula glabrous or with closely appressed inconspicuous scales on
emergence from the bud region; abaxial leaf surface lacking rows of minute, waxy dots
associated with sunken stomata.
2. Leaves of mature plants densely covered abaxially with white or silvery, overlapping and
interlocking, hyaline, fimbriate scales (leaves of young plants have similar scales but
may appear green beneath because the scales are less densely disposed); petioles of adult
plants tan velutinous, abaxially; inflorrscence branches and flowers pink to purple at
anthesis, primary bracts densely covered with rufous appressed scales; (leaf sheath long
linguiform, densely, tan, velutinous, Figure 1 0 ~ ) grows naturally only in the interior
2. Leaves of mature plants concolorous or very slightly lighter green beneath, scales when
present widely spaced, not touching; petioles of adult plants very soon glabrescent
abaxially; inflorescence branches and flowers white, ivory, or yellowish at anthesis, scales
on primary bracts not rufous colored.
3. Leaves of mature plants concolorous, glabrescent; scales when present minute (difficult
to see with the aid of an 8 power lens); segments broadest well beyond the palman;
sheath long lingniform (Figure log), (observed best in the region of newly emerging
leaves (Figt:re 1 In); inflorescence branches and flowers ivory to yellowish at anthesis,
the axes densely granulose-puberulose (except subspecies parviporn in which the
inflorescence branches are usually glabrous-glabrate but then the pedicels toward the
apex of the branches are very short); grows naturally only southwest and west of the
3. Leaves of mature plants lighter green beneath, with scattered but conspicuous fimbriate
scales which have a conspicuous translucent central region; segments broadest at the
point of fusion; sheath not lingniform, rather deeply V-form (Figures lOC, 11.4):
inflorescence and flowers white at anthesis, the axes glabrous; common, but grow’s
naturally only in close proximity to the sea, below 250 ft elevation on islands throughout
the northern Caribbean and the Bahamas, and islands off the coasts of Florida,
Mexico, and British Honduras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. T. radiata
1. Inflorescence branches exhibiting only small inconspicuous, nartlike, or disklike, pedicel
remains, less than 1 mm (averaging about 0.4 mm) in length, following fruit set; adaxial
hastula densely velutinons with white, silky, scales on emergence from the bud region;
abaxial leaf surface usually exhibiting rows of numerous, minute, white, waxy, dots
associated with sunken stomata. (Inflorescence arching, exceeding the leaves, branches
white to yellowish with age; leaves often blue green to whitish abaxially and often exhibiting
closely arranged, pale, hyaline, fimbriate scales abixially.) 4. T. morrisit
of eastern Jamaica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. T. excelsa
Blue Mountains in Jamaica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. T. paruiflora
. . . . . . .
SUhIBER 19 69
sessile, Shortly thereafter Beccari (1 907) published
his comprehensive study of American Coiypheae in
which he also divided the genus Thrinax into two
sections adopting the groupings described by Sargent
but under different names. Beccari’s division,
however, was based on the degree of intrusion of
the testa into the endosperm: Euthrinax, having
the seed completely perforated, and Typhlothrinax,
having the testa intruding only part way into the
endosperm. Since Sargent’s interest was only with
palms of North America it remained for Beccari to
raise Hemithrinax to full generic status again,
which, with Coccothrinax and Thrinax, comprised
the three genera of the subtribe Thrinaceae. In a
supplement, in English, Beccari (1933) dealt with
Thrinax in much the same manner as earlier but
the addition of floral dissections made it a valuable
adjunct to his earlier work.
In 1937, Cook of the Bureau of Plant Industry,
United States Department of Agriculture, proposed
a new genus, Simpsonia, which was based on Sargent’s
Thrinax microcarpa, but since the proposal
does not conform to Article 36 of the International
Code of Botanical Nomenclature it is not validly
published. Bailey (1938: 131) revived the use of
Sargent’s Pedicellatae and Sessiliporae because “the
species may vary within themselves in the extent
of the centralium in the seed.”
Common Names: As might be expected the common
names applied to the members of the genus,
as well as to the genus as a whole, are almost as
difficult to straighten out as are the scientific
names. The first common name to appear in the
literature was that applied by Sloane (1696). He
called the palms “thatch,” a name also applied to
any other palm used in the roofing of shelters, but
in many cases the distinction is made by the addition
of a prefix or qualifying word. In 1937, Cook
TABLE 4.-Charactel-s known to be reliable in
separating Coccothrinax from Thrinax
Characters I Coccothrinax Thrinax
Petiole ..............
Fruit ..... ..........
Seed , . ..............
Endosperm ....
solid in the sheath
fleshy, purple-black
cerebriform smooth
infolded ruminate nonruminate,
split in the sheath
pulpy, white or ivory
partially or completely
intruded
by the testa
referred to a group of sturdy wind-resistant palms,
including a Thrinax, as hurricane palms. Bailey
(1938) referred to members of the genus as peaberry
palms, and in 1949 he included members of
Thrinax along with Coccothrinax under the name
broom palms.
1. Thrinax parviflora Swartz
Thrinax parvipora Sw., Prodr. 57. 1788.-Ait., Hort. Kew.
33473. l789.-Sw., F1. Ind. Occ. 1:614, t.13. 1797.-Linn.,
Sp. P1. sec. (Flrilld. ed.) 2:202. 1799.-Pers., Synops. P1.
1:383. 1805.-Lam., Encycl. Mkth. 7:635. 1806.-Ait., Hort.
Kew. 2nd ed. 2:307. 1811.-Titford, Hort. Bot. Amer. p.
112. 1811.-Lunan, Hort. Jam. 2:28. 1814.-Pers., Synops.
P1. (Diet. ed.) 2:1091. 1830.-Linn., Syst. Veg. sec. (J. A.
& J. H. Schult. ed.) 7(2):1300. 1830.--Mart., Hist. Nat.
Palm. 3:255, t.103, pro parte. 1838.-Kunth, Enum. P1.
33253. 1841.-Griseb., F1. Brit. \t’. Ind. p. 515. 1864.-Cook,
Bull. Torr. Bot. Club 28:535, pro parte. 1903.-Becc.,
Webbia 2251, pro parte. 1907; Annal. Roy. Bot. Gard.
Calcutta (Martelli ed.) 13:327, pl. 28: fig. 3, pro parte.
1933 [“1931”].-Read i?i Adams, F1. P1. Jam. pp. 75-76.
1972.
TYPE.--SWUT~Z s.n. Jamaica (a single sheet with
juvenile leaf and packet of flowers).
Small to medium palm, 1-10(-13) m high, caudex
columnar, often very slender, 5-14(-15) cm in
diameter, internodes 1.0-6.5 cm apart on lower
trunk, reduced to lines above.
Leaf sheath 45-70 cm long, apex long linguiforni
7-22 cm long, netlike, entire, becoming ragged and
fibrous with age, tan, velutinous, glabrescent with
age on exposed surfaces; petiole (35-)50-145 (-160)
cm long, abaxial surface densely tan lepidote, soon
glabrescent on exposure, (1.0-) 1.2-2.2 (-2.4) cm
wide at the sheath, (0.8-) 1.0-2.2 (-2.4) cm wide at
apex; hastula triangular or rounded acute (0.1-)
0.5-3.4 (-4.6) cm long, may become tubular with
expansion and folding of blade, abaxially with a
distinct ridge 0.1-0.5 (-1.0) cm long or a triangular
projection 1.0-1.7 cm long; leaf in bud densely
tan-brown lepidote on all exposed surfaces, often
persistent on principal nerves and exposed portion
of lower segments.
Leaf blade when expanded very variable, circular
and flat on young plants and in the shade, the basal
lobes folded up away from the petiole when in the
sun, and on older plants, or variously folded and
contorted on tall slender plants, abaxial surface
with microscopic widely scattered scales; palman
70 ShIITHSOSIAPi CONTRIBUTIOSS TO BOTANY
variously folded and very irregular in outline,
(7-)8-34 (-42) cm broad; segments rhomboid (35-)
37-57 (-60) in number, (38-)40-90 (-96) cm long,
(2.3-) 2.8-6.0 (-6.4) cm wide, Jvidest beyond point ot
fusion, the middle two or three segments a1tva)s
fused for a greater length but rarely totally.
Inflorescence erect to arching, 40-175 cm long,
bracts greenish with brown appressed scales, each
inflorescence with 7-22 primary branches, those in
the lower half with 17-45 ultimate branches, these
ivory to yellowish at antliesis, densely puberulous
to glabrescent or glabrous, several of the loivermost
branches occasionally forked or highly branched
but subtended by a narrow triangular bract; flowers
ivory to yellowish, fragrant, (lj-) 32-60 (-84)
per branch; stamens 5-10 (-15) per flower, anthers
linear to oblong (1.1-)1.2-4.0 (-4.2) mm long; infundibuliform
stigma oblique to lateral, ciliate;
developing ovary papillate to granulate.
Fruit papillate to scurfy granulate, tan before
maturation, become white and smooth or slightly
punctate, except when infected by a fungus, then
tessellated, 6.4-7.3 (-7.5) cm in diameter; pedicels
bracteolate, (0.1-)0.5-2.8 (-5.2) mm long; primary
branches yellow to orange at maturity of fruit
(green or insect grazed at high elevations); rachis
6-25 cm long; ultimate branches 4.5-17.0 cm long,
1.1-1.5 mm thick at point of insertion on rachis;
seed tan to dark brown 5.4-6.2 (-6.5) mm in diam.,
normally completely perforated by the testa. Haploid
chromosome number, n= 18.
TYPIFICATION OF Thrinax parviflora
The original publication of Thyinax parvipora
by Olof Swartz (1788) consisted of a very short
diagnosis, which characterized the genus, and a
reference to Patrick Browne’s account of the species
published 32 years earlier. There was no reference
to his own collections or to the younger Linnaeus’
unpublished manuscript. Patrick Browne’s (1756)
description of the palm growing and thriving on
all the rocky hills of Jamaica clearly indicated the
palm intended. Swartz (1797) reproduced the
Browne diagnosis and elaborated on it considerably
with his own observations made while in Jamaica
from 1784 to 1786. The description of such characters
as long, glabrous, compressed-terete petioles
which were longer than the 1-2-ft-long leaf blades,
inflorescences shorter than the leaves and pale yellowish
anthers leaves no doubt which Jamaican
taxon was being described. Only one taxon in
Jamaica exhibits these characters. An illustration
in Sivartz (1797: pl. 13) shows dissections of the
flower and fruit, but they are very schematic. The
drawing unquestionably portrays Thyinax which, if
interpreted on the basis of the scurfy-looking fruit,
provides additional support for the present determination
of the species.
In the herbarium at the Naturhistoriska Riksmuseet
in Stockholm, where many of Swartz’s collections
are maintained, three herbarium specimens
attributed to him have caused considerable confusion
over the past 186 years. Through the courtesy
of Professor Dr. Tycho Norlindh of the Riksmuseum,
I was able to make a careful study of the
specimens while they were on loan to Dr. H. E.
Moore at the Liberty Hyde Bailey Hortorium, Cornell
University, in New York. The three specimens
represent two distinct taxa, both Jamaican, but it
is obvious that only one of the sheets can be the
type for Swartz’s Th7 ii2ax pnrviflora. The three
specimens each have inscriptions in the hand of
J. E. Wikstram, the director of the Botanical Nuseum
and Botanical Garden of the Royal Academy
of Sciences from 1818 to 1856. These inscriptions,
according to correspondence betn.een Dr. E. Asplund
of the Riksmuseum and Dr. L. H. Bailey in
1937, “often giving valuable information regarding
the provenience [sic] of the specimen, but sometimes
being only a repetition of what is already
present on the label (in old time mostly on the
back of the sheet).” Each sheet has a quite different
history and so will be discussed separately.
The first specimen (Figure 4 6 ~ ) containing a
short section of a Thrinnx inflorescence also contained
a portion of a Bnctris inflorescence; the latter
has been removed to another sheet. This sheet
retaining the Thrinax specimen and labeled “Sabal
umbraculifera illart.” could not have been collected
by Swarti, for as Professor Norlindh has explained
in correspondence:
It once belonged to the Linnean Herbarium and can impossiblj
ha\e been collected by Swaitz, who made his collection$
in the \Vest Indies in 1781-86. Di. Dahl obtained it
fiom Linnaeus (patei), thus before 1778, and then it was
incorpoiated in the herbarium of Baron Clas Alstromer and
later left to the Ro!al Academy oE Science, the supeniser of
Riksmuseum.
SUMBER 19 71
According to the letter from Asplund to Bailey,
“Sabal umbraculifera Mart.” below the right specimen is
written by TVikstrom. On the back of the sheet TVikstrom
has repeated the same determination and also written “Herb.
Alstroemeri”. In another hand is written “Palmeto Dahl.”,
“Xorypha umbraculifera”, “Dahl a Linni- P.”, and “Jamaica”.
Those words may be written by Clas Alstromer or Andreas
Dahl (or perhaps by a copyist, since they are obviously written
in text hand). After “umbraculifera” is added “Jacq.”,
probably by W’ikstrom. The specimens on this sheet were
given by Linnaeus to Dahl and came from him to Alstromer.
Alstromer and Dahl were pupils of Linnaeus. . . . The former
was a member of the Swedish aristocracy and had a botanical
garden and museum in Gothenburg, of which he had employed
Dahl as a keeper. His herbarium came to this museum
in 1848. The father of Alstromer and Linnaeus were among
the founders of the Royal Academy of Sciences.-There is
nothing indicating that Swartz did ever see this specimen.
The second specimen (Figure 4 6 ~ ) containing
three pieces of a Thl-inax inflorescence, including a
terminal portion with six primary branches attached,
is labeled “No. 4 Sabal umbraculifera Mart.
Jamaica Swartz” and “Herb. Swartzii” in Wikstrom’s
hand. This and the preceding specimen
belong to the same taxon, although probably not
from the same source. It seems unlikely that, had
either of these specimens been seen or utilized by
Swartz, they would have been incorrectly labeled.
The annotations by Wikstrom rvere made 40 to 60
years after the publication of T h ~ i n n x by Swartz,
FIGURE ‘%.-Two herbarium sheets often erroneously considered as part of the Swartz type for
Thrinnx pnwi/lo~.n. Discussion in the text. (Supplied by courtesy of Prof. Korlindh, Riksmuseum,
Stockholm.)
ShfITHSONI.4S COiXTRIBUTIONS TO BOTAKY
yet there is no indication that they ever were associated
with the genus until Beccari annotated the
sheets in 1907 and identified them with Thrinax.
The third specimen (Figure 47) has a small flabellate
leaf and a packet containing fragmentary
rachillae and flowers. This specimen is labeled by
Wikstrom as “Jamaica Swartz.” and “Thrinax parviflora
Swz.” thus as early as the 1800s this specimen
was recognized as one of Swartz’s own. In addition
the packet contains flowers wrapped in paper labeled
“Thrinax parviflora,” below which is written
in a different hand, “Swartz scripsit.” The latter is
no doubt Wikstrom’s handwriting for it is the same
as on the other sheets. Thus Wikstrom provided us
with the earliest indication of a specimen seen ancl
perhaps even labeled by Swartz.
In 1907, Beccari annotated all three sheets as
“Spec. typ.!,” apparently not recognizing that two
taxa were involved, also ignoring that only one of
the sheets had been labeled Thrinax by Swartz. Dr.
L. H. Bailey (1938:138), when writing about these
same specimens, concluded, as has the present author,
that the last sheet described (Figure 47) “undoubtedly
may be considered authentic Thrinnx
FIGURE 47.-Type specimen of Thrinax pamiflora
Sw. preserved in the Riksmuseum, Stockholm,
Sweden. Inset is a photomicrograph of one of the
flowers in the packet on the type sheet (X ca 6).
(Courtesy of Riksmuseum.)
KUMBER 19 73
TABLE 5.--Comparison of Swartz’s type and description of Thrinax parviflora
with two of the Jamaican taxa
Swartz type and
Characters description
Petiole length ................
Blade width ....................
Pedicel length at
an thesis ........................
Flower color ...................
Anther length ..................
’ longer than blade
1-2 feet
(0.5)0.7-1.5 (1.7)
mm long
yellowish
(1.4)1.5-2.0 (2.2)
Thrinax parvifiora
subsp. paruipora
longer than blade
2-3 ft or more
(0.3)O.j-1.5 (1.7)
mm long
ivory to pale yellow
(1.1)1.2-2.0 (2.2)
Thrinax radiata
shorter than blade
3-6 ft
(1.4)1.6-2.4 (2.7)
mm long
white
(1.7)l.a-2.3 (2.5)
parvipora of Swartz.” Bailey had corresponded with
both the Linnean Society of London and the herbarium
of the Royal Botanic Gardens at Kew, in
1937, but failed to turn up Swartzian Thvinax in
either herbarium.
An undated sheet containing an inflorescence of
Thrinax located in the Banks herbarium of the
British Museum was examined, at my request, by
Dr. William T. Stearn, who made the following
comments (in correspondence ) concerning Banks
and Swartz:
There is one sheet from Bank’s herbarium which Swartz
almost certainly saw during his long period of work in London
on his TVest Indian collections. If Banks already possessed
a specimen from the FVest Indies, obviously, according
to the attitude of the time, he did not need another, and the
specimen he possessed was simply annotated with Swartz’s
determination. If, on the other hand, Banks lacked a representative
of the species, then Swartz gave him one and his
specimen was mounted immediately and annotated according
to Swartz, thus obviating error. Swartz did not mount his
own specimens. He was a very generous man and evidently
gave away many specimens. After his death his specimens
were mounted at Stockholm and written up for the most
part by Wikstrom; presumably Swartz had written the name
on the covers, as was then usual with loose specimens, and
presumably, as was also usual, alas! the covers were then
thrown away. The sheet here collected in Jamaica by Robins
was photographed by Moore in 1950 (this photograph B. hi.
2577).18 In the capsule are two fruits presumably given to
Banks by Swartz, as the sheet is annotated as having on it
Jamaican material from “Dr. Swartz” and these fruits are
the only ones to which this note could apply.
l3 The specimen in the herbarium of the British Museum
(Herbarium Banks) “Jamaica. Robins” has been chosen as
the lectotype of Thrinax martii, as it is the only known
specimen of those cited by Martius, upon which T. rnartii
is based.
There is no doubt now that the only specimen
considered authentic Swartz material, and therefore
the holotype, is the single sheet, illustrated in Fig
ure 47, in the Riksrnuseurn in Stockholm, which
has a small but complete leaf and a packet of
flowers. The other two sheets, until now considered
identical, are removed from the species and are
treated under T. radiata. The Robins specimen in
the Banksian herbarium also represents T. radiata,
with the exception of the fruits which cannot at
this time be placed with any certainty.
In Jamaica there are three distinct and welldefined
taxa in the genus Thrinax that have been
recognized as a result of extensive field work and
numerous collections. One of these must of course
be associated with Thi-inax parvipora Sw. All three
taxa have a 6-dentate perianth, an emarginate
stigma, and a 1-seeded fruit. Apart from the information
in the Flora, the type specimen provides
additional evidence in support of the present determination
of the species. The juvenile leaf on the
sheet is unfortunately of little significance, although
it has certain indefinable characteristics such a5
texture and general appearance which suggest a
similarity to the seedlings collected in the mountains.
The leaf lacks lepitlia of any sort (T. 7-crdiata
seedling-leaves usually exhibit scattered lepidia on
the abaxial surface). The attached packet contains
fragments of flowers and small pieces or rachillae,
pedicels, and numerous anthers. Measurements of
the anthers and pedicels are compared with similar
measurements taken from plants at Devils Backbone
and illorant Point in Figure 38. Although Swartz
placed the genus in the Hexandria, flowers in the
packet had as many as 7 stamen filaments still
attached. Many flowers of Thrinax in Jamaica have
74 ShlITHSONIAS CONTKIBUTIOSS TO ROTASY
been found with 6 stamens; however, it is more
common to find any number up to 12 or 15.
Comparing the Swartz type and description with
the taxa in Jamaica as presently understood, one
taxon is immediately eliminated from consideration
because two of the characters mentioned by Swartz
in the Flora certainly do not apply. T1i)anax excelsa
of the John Crow RIountains in eastern
Jamaica has leaves no less than 6 to 12 feet across
and very striking pink flowers. Another taxon,
Thyinax 7 adiata, of littoral areas, has leaves ineasuring
more than 3 feet across and the petioles are
much shorter than the blade, the flowers are white
and the floral pedicels are more than 1.6 niin long
even at anthesis. Thus Thiinau Tadiatn, which has
erroneously been knolvn as T. excelsa in Jamaica,
T. ruendlandiana in Cuba, T. pnivipo7a in the Bahamas,
and T. fioridnna in Florida, can also be
ruled out. Recent anatomical findings agree.
Although Thrinax paiuifloia Sw. is an extremely
variable species (see section on variability) the
population above 609 m (2000 ft) elevation on Mt.
Diablo has been found to resemble S~vart7’s t j pe
and description most closely. It is this population,
seen on all the cliffs and rocky hills along the main
route across Jainaica (now wbspecies p a i ~ i f l o ~ u),
which was chosen to represent one of the two
Tamaican taxa in Table 5 comparing them with the
Swartz material.
In 1838, LIartius published an amplified account
of T. pal-vifiora consisting of information from several
sources. The description, including elements of
several taxa, in fact appears to contain characteristics
of his newly described T. muIt;floia. Grisebach’s
description (1864) only contributed to the
growing confusion concerning the application of
T. parvifiora.
For the first time in 199 years, identification of
the species with a taxon in Jamaica came close
when Beccari (1 907 ) recognized the Harris collection
from Ferry River as true T. pa7-vifEo~a. Unfortunately,
elements of a second taxon from the littoral
of the north coast were also included in the
description of the species. The illustrations of flowers
and fruits that appeared in the 1933 supplement
do not accurately depict authentic T. pni vipoia.
On p1. 28: 111, figure 1 (taken from the Swart7
type) and figures 2 and 4 (taken from a Harris
specimen collected along Ferry River) represent
true T. pnrvifloia, but the remaining figures most
likely represent T. radiata. As a result, rather than
straightening out the problem, Beccari further obscured
the true nature of the species. In 1907, he
also described T. tessellata based on a phytomorphic
form from Mt. Diablo. Thrinnx haiyisinna
described by Beccari (1908) was based on a nontessellated
and immature-fruited specimen collected
on the heights of Dolphin Head.
Ilihen L. H. Bailey (1938) undertook the challenge
to revise the genus, he did not fully understand
T. pa~uifloia in Jamaica. Although Dr. Bailey
collected and correctly identified material of authentic
T. parvifiol-a from Salt River, Clarendon,
and the Port Royal Mountains, in St. Andrew, he
nevertheless chose to typify the species by using the
Florida taxon. Although Florida and Jamaican
plants of T. radiata differ only in minor quantitative
characters such as leaf segment and fruit size,
Diagnostic Key to the Subspecies of Thrinax parviflora
1. Adasial hastular projcction 0.3-1.2 cni longx, lacking or vestigial abaxially: branches of the
inflorescence glabrous, glabratc, or lightly; pubei-ulose; inflorescence crcct and shorter than
the petioles at anthcsic; fresh anthers short (1.2-2.0 mm long): mature plants with leaves
forming an open crown of conspicuously twisted and curled leaf segments . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . wbspecies parvifZora
2. Adaxial hastnlar projection \cry prominent or even long pointed, 1.5-4.4 cm long, prominent
and up to 1.6 cm long abaxially; branches of the inflorescence densely granulose puberulose
throughout; inflorescence arching, equaling or longer than the petioles at full anthesis;
fresh anthers elongatc (ca. 2-4 mm long): mature plants generally with leaves forming a
dense crown of broad blades with more or less flat to curved or slightly twisted leaf
segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . subspecies puberula
’* Measured from beneath the hastula, against the lamina.
hUMBER 19 75
it is curious that Bailey should have chosen those
from Florida as representative of T. pnruipo.1-a while
at the same time referring the Jamaican plants to
T. excelsa. Specimens of true T. pawipom, which
Bailey collected in Jamaica, were not considered
in his final analysis.
Following Bailey’s revision of the genus in 1938
nearly all references to T. parvifiora that appear in
the taxonomic literature actually refer to T. mdiata.
Common Names: Sloane’s early common name
for the species in 1696 was simply “thatch,” but in
Patrick Browne’s description in 1756 the names
“palmeto-royal” or ”palmeto-thatch” were cited
L. H. Bailey cited “iron-thatch” as the common
name on a specimen collected at Alligator Pond, in
the Parish of Manchester, in 1931. Presently the
species is called thatch, thatch-pole, or broom palm.
Thrinax parviflora Swartz subsp. parviflora
Thrinax earviflora Sw. subsp. paruiporn, I’rodr. 57. li88.
[Type: Olof Su’artz, Jamaica (S).]
Thrinax tessellata Becc., TVebbia 2:27l. 1907. [Type: W. X.
Maxon & G. A‘. Collins 5, 26 May 1904. Hollymount, Jamaica,
2600 ft (792.4 m) elev. (B, specimen destroyed!;
US, neotype, W. R. Maxon 2353).]
Thrinax harrisiana Becc. in Fedde, Repert. 6:94. 1908. [Type:
Harris 9273. Dolphin Head, Jamaica, 1500-1600 ft. (455.2-
487.7) elev. (B, specimen destroyed?: UCIVI, isotype).]
SPECIMENS EXAMINED AND COLLECTING LOCALITIES
(Parish and locality number in Jamaica).-Abbreviations
for herbaria are those of J. Lanjouw and
F. A. StaAeu in Index Herbariorum, Reg. Veg. 31(1)
ed. 5, 1964, with the addition of FTG, for the Fairchild
Tropical Garden, Coral Gables, Florida, and
DA, for L’gcole National d’hgriculture, Damien,
Port-au-Prince, Haiti.
Locality numbers correspond with those given
on the map of Jamaica (Figure 48).
ST. ANN
5, Stirling Castle Forest Reserve: 609 m, Jan. 1958, R. A.
Horvard 2 G. R. Proctor 15055 (GH, I J ) ; Jan. 1960,
R. A. Hozc’nrd et al. 20513 (IJ).
6. Lydford Post Office: 457.2 m, Dec. 1953, R. A. Horcard
C- G. R. Proctor 13506 (BH).
7 . Deiils Backbone, 616.62 m, ridge along parish boundary
east of Aft. Diablo: Jan. 1967, R. TI’. Read 1736, 1737,
1738 (’3, CCT\’I, US).
8. Union Hill: 669.96 m, Apr. 1908, A‘. L. Britton &
A. Hollick 2753 (KY, US); Jan. 1967, R. 1%’. Rend 1939
(BH, IJ, US).
9. Grier Mount, 838.2-914 m: Dec. 1953, R. A . Howard 2
G. R. Proctor I3611 (BH); Jan. 1966, X . Ti’, Rend 1573
(US).
ST. CATHERINE
10. Hollymount Hill, 807.72 m, the first pinnacle on right
side of Hollymount road after reaching 762 m elevation:
--
ST. THOMAS
J AMAlCA
FIGURE 48.-Distribution of Thrinax parvipora in Jamaica. The locality numbers refer to list
of Specimens Examined under treatment of the species.
SMITHSONIAS CONTRIBUTIOSS TO BOTANY 76
11.
12.
13.
Apr. 1963, R. W. Read L- G. R. Proctor 1146 (FTG); Feb.
1966, R. W. Read 1595,1596 (US); June 1966, R. TV. Read
1681 (BH, IJ, S , UCWI, US); Jan. 1967, R. W. Read 1761
(US); Mar. 1967, R. TV. Read 1835 L- 1836 (US).
Hollymount: May 1903, TI’. R. Maxon 1961, 1962 (US);
838.2 m, May 1904, LIT. R. Maxon & G. N. Collins no. 5
(B, holotype of T. tessellata, not seen); May 1904, TV. R.
Alaxon 2353 (UCM’I; US, neotype); Feb. 1905, TV. Harris
no. 6 (UCIYI); Sept. 1906, A’. L. Britton 734 (NY); Aug.
1931, L. H. Bailey 710 (BH).
Blue Mountain, hft. Diablo, top of south escarpment:
807.72 m, May 1967, R. W. Read 1895 (GH).
Mt. Diablo: Mar. 1916, E. P. Kiilif, 420 (US); south slope;
457.2-502.92 m, May 1967, R. TI’. Read 1896 (FTG);
502.92-533.40 m, May 1967, R. IT. Read 1897 (FTG).
14b. Charlton to IVorthy Park road: 27-mile marker, 487.68
m, May 1967, R. TV. Read 1898 (US).
27.
30.
31.
32.
MANCHESTER
Somerset: 762-914 m, Feb. 1955, G. R. Proctor 11604 (IJ).
TRELAWNY
Ramgoat Cave: 457.2 m, Apr. 1963, R. TV. Read L- G. R.
Proctor I148 (FTG); Nov. 1964, G. R. Proctor 25675 (IJ).
Donkey Trail, ?,$ mile north of Burnt Hill intersection:
518.16-556.26 m, May 1966, R. TI’. Read 1663 (US): Jane
1967, R. W. Read 1757 ( S ) , I758 (FTG).
Burnt Hill: Jan. 1956, R. A. Howard et al. 14678 (IJ);
1/4 mile north of intersection, 533.4 m, Apr. 1966, R. TI’.
Read I640 (US).
36b. TVeather Station (south of Windsor Cave): over 400 m
elev., Apr. 1967, C. D. Adams s. n. (Read 1877) (US).
ST. JAMFS
37. Fray TVoods: 502.92-548.64 m, Mar. 1908, TV. Harris 10341
(B, not seen; IJCIVI, US).
HANOVFR
38. Dolphin Head: above Askenish, 457.2-548.64 m, Mav
1906, W. Harris 9273 (B, holotjpe of 7‘. harrisiana, not
seen; UCWI, isotlpe); Mar. 1908, TI’. Harris 10264 (B,
not seen; UCITI, US); Mar. 1908, A’. L. Britton L- A.
Hollick 2217 (US); Feb. 1931, L. H. Bailey 15053 (BH);
Sept. 1951, R. G. Robbins 2878 (UCM’I); Apr. 1955, G. R.
Proctor 10054 (IJ); Mar. 1960, G. R. Proctor 2068? (IJ);
Sept. 1960, C. D. Adams 7995 (M, UCIVI); Apr. 1963,
R. TI’. Read L- G. R. Proctor I150 (FTG).
39. Bubby Hill: 441.96 m, Aug. 1965, G. R. Proctor 26675
41b. Hills near Kemshot: Mar. 1908, A’. L. Britton 2426 (US).
(IJ).
SOURCE UNKNOWN
Olof Swartz, Jamaica (S, holotype of T. parviflora Sw.).
Thrinax parviflora Swartz puberula, new subspecies
A subspecie parviflora differt hastula adaxialiter
1.5-4.4 cm longa et aliquantum prominente, abaxialiter
usque ad 1.6 cm longa, inflorescentiae ramulis
dense granuloso-puberulis.
TYPE: R. W. Read 1565 Palmers Cross, Parish of
Clarendon, Jamaica, Nov. 1965 (US).
SPECIMENS EXAMINED AND COLLECTING LOCALITIES
(Parish and locality number in Jamaica).-
ST. ANDREW
Cane River Gorge, 76.2-182.88 m: Mar. 1931, L. H. Bailey
15089 (BH); Apr. 1963, R. W. Read 3 G. R. Proctor 1139,
I140 (FTG); Mar. 1966, R. TV. Read 1603 (US).
Dallas Mountain: Apr. 1963, R. TV. Read cb G. R. Proctor
1141 (FTG); 542.54 m, Nov. 1965, R. TV. Read .b C. D.
Adams 1559a (US).
Long Motintain: Mona Hill, Sept. 1906, N. L. Britton 373
(XY, US); vicinity of water tank, Montclair Dr., Beverly
Hills, 213.36-243.84 m, May 1966, R. TY. Read 1657,
1658c (BH, US); June 1967, R. W. Read 1950 (US); Mona
Mt., May 1904, TV. R. Maxon 2147 (US).
Ferry Hill: along river side, 30.48 m, May 1904, W.
Harris 4 (B, not seen); Ferry River, G. A’. Collins s. n.
(US): May 1904, TV. R. Mavon 2190 j. 2197 (US).
ST. CATHERINE
14a. Charlton to ‘/%‘orthy Park road (26-mile marker): 381
m, June 1966, R. TV. Read 1672 (FTG).
14b. Charlton to W‘orthy Park road (26.5-mile marker):
487.68 m, June 1967, R. TV. Read 1941, 1944 (US).
15. Mendez Hill to Spring Vale: 152.4-304.8 m, May 1967,
R. W. Read 1899 (FTG), 1900 (NY).
16. Cudjoe Hill: 152.4 m, Mar. 1966, R. W. Read 1602 (BH).
17a. Johnston Pen vicinity, near rock quarry, 2-4 miles
south from Spanish Town, off the Friendship Rd.: 121.92
m, Nov. 1969, R. TI’. Read 1655a, 1655b (US).
17b. Guanaboa Vale, 9 miles NLT of Spanish Town: Nov.
1965, R. TI’. Read 1551 (US).
18. Spiing Garden: 76.2 m, Mar. 1966, R. II’. Read 1600,
I601 (BH).
19. Hellshire (Healthshire) Hills: near Salt Island, 76.2 m,
Sept. 1908, I T . Harris ~9 S. L. Biitton 10524 (UCWI, US).
20a. Rio Cobre Gorge: near Bog Walk, ca 182.88 m, Feb.
1931, L. H. Bailey 15065 (BH).
20b. Great Goat Island: Apr. 1920, TT’. R. Maxon 3 E. P.
Killip 1593 (US).
CLARENDO~
21. Palmeis Cross: in a sisal field, 45.72 m, Apr. 1963, R. 14’.
Read 3 G. R. Pioctor 1147 (FTG); Sov. 1965, R. W.
Read 1565 (BH, S, CTCFI’I, US, holotjpe of 7’. parviflora
subspecies ~ i u b e i t ~ l a ) ; June 1966, R. TI’. Read 1676b (BH,
s, UCWI, US).
NUMBER 19 77
22.
23.
24.
25.
26.
28.
29.
31.
33.
34.
35.
36.
40.
41a
42.
Cockpit: 4.75-12.19 m, June 1966, R. W. Read 1675a,
16756 (US); Feb. 1967, R. W. Read 1827b (US).
Salt River: ca 15.24 m, Nov. 1935, I.. H. Bailey 202 (BH).
Portland Point: 3.05-6.1 m, Feb. 1967, R. W. Read 1825
Clarendon Park: 83.82 m, Feb. 1966, R. IV Read 1583
(US).
(BH).
MANCHESTER
Round Hill: 60.96 m, July 1966, K. 14’. Read 1689 (US).
Alligator Pond: Feb. 1931, L. H. Bailey 15080 (BH).
ST. ELIZABETH
Bull Savanna: 152.4 m, July 1962,
(UCWI).
TRELAWNY
White Bay: 3.05 m, Aug. 1966, R.
(US).
E . T. Robertson 9326
14’. Read 1702a, 170211
Donkey Trail, y2 mile north of Burnt Hill intersection:
518.16-556.26 m, May 1967, R. Ti’. Read 1946 (US).
Mango Tree Hill (Carambi Hill): 542.54-563.88 m, Jan.
1966, R. W. Read 1574 (US); May 1966, R. TI’. Rend 1662
Sherwood Content: 228.6-247.32 m, Oct. 1955, G. R.
Proctor 11065 (IJ).
Mt. Pantrepant: ca. 91.44 m, Nov. 1935, L. H. Bailey 215
IVindsor Cave: 152.4 m, Aug. 1955, G. R. Proctor 10540,
15631 (IJ).
(US).
(BH).
HANOVER
Silrer Spring: near Fish Riler, 60.96 m, Mar. 1966, R. TI’.
Read k Iwor Corninan 1614 (K).
Lances Bay: 15.24 m, June 1966, R. 1%’. Read I669 ( S ) .
WESTMORELAKD
h’egril Hill: New Hope Estate, 53.34-76.2 m, Nov. 1955,
G. R. Proctor 11206 (IJ); inland from Revival, June 1966,
R. W. Read I670 (US).
SOURCE UNKNOWN
1849 TVulZschZageZ 8573 “Borassus flabelliformis L.?” (M);
1878 W. 1512, 2 sheets (UCWI).
2. Thrinax radiata Lodd. ex J. A. & J. H. Schult.
Thiinax radiata Lodd. ex J. A. & J. H. Schult. in Linn.
Sjst. Veg. sec. 7(2):1301. 1830.-Mart., Hist. Nat. Palm.
3:257. 1838.-Kunth, Enum. P1. 3:253. 1841.-Becc., Annal.
Roy. Bot. Gard, Calcutta (Martelli ed.) 13:335. 1933 [“1931”].
[Type: Specimen in Herbarium Regium Monacense “e horto
paris” (M).]
Coccothrinax radiata (Lodd. ex J. .4. & J. H. Schult.) Sarg.
ex K. Schum. in Just’s Bot Jahresb. 27 (1):469. 1901. [4s to
the type only, not subsequent usage and associations.]
Thrinax floridana Sarg., Bot. Gaz. 27:84. 1899; Silva 14:81,
t. 735. 1902.-Small, F1. S.E. U.S. p. 221. 1903.-Becc.,
TZ‘ebbia 2:262. 1907.-Small, F1. Fla. Keys p. 26. 1913.-
Sarg., Manual Trees N. .4mer. p. 97. 1926.-Becc., Annal.
Roy. Bot. Gard., Calcutta (Martelli ed.) 13:328. 1933
[“1931”].-Read, Amer. Hort. Mag. (Ledin ed.) 40:28.
1961.-Read in Moore, Principes 7:168. 1963. [Type:
Sargent, Thrinax no. 2, 14 R’ov. 1898 (“cultivated at Miami
from Long Key”), Florida (AAH).]
T. zoeizdlandiana Becc., W‘ebbia 2:265. 1907. nom. illeg.?
Pomona Col. Journ. Econ. Bot. 3:403, f. 168. 1913.-Sarg.,
Manual Trees N. .4mer. p. 97. 1926.-Burret, Sv. Vet-akad.
Handl. ser. 3, S(7):lO. 1929.-Becc., Annal. Roy. Bot. Gard.,
Calcutta (Martelli ed.) 13:332. 1933 [“1931”].-Lebn, F1.
Cuba. 1:260. 1946. [Type: Tl’riglit no. 3219 (1860-1864),
Cuba (G).]
T. martii Griseb. et H. I\’endl. ex Griseb., Cat. P1. Cub. p.
221, 1866 [excluding elements of Coccotlirinax]. [Type:
Jamaica. Robins (Inflor.) Herb. Banks (BM).]
T. putnilio sensu Mart., Hist. h’at. Palm. 3:256, as to tab.
103, f. IV, 1-4. 1838 [et auct. non Lodd. ex J. A. pi J. H.
Schult. in Linn. Syst. Veg. sec. 7(2):1301. 18301.
Porothrinax puiiiilio H. TVentll. ex Griseb., Cat. P1. Cub. p.
221. 1866 [pro parte, as to the specimen cited, not the type].
T. parni/?ora auct. non Sw.: Mart., Hist. Sat. Palm. 332.55,
pro parte. 1838.--Sanv., Anal. Acad. Ci. Habana 7:563.
1871.-Vasey, Cat. For. Trees US., USDA Rept. no. 11.
1873:186. 1876.-Chapm., Bot. Gaz. 3:12. 1878; F1. South.
U.S. p. 651. 1883.-Sarg., Gartl. & For. 9:162. 1896.-Becc.,
ITebhia 2:25.5, pro parte. 1907,- Small, Journ. N. Y. Bot.
Gard. 26:49. 1925; 31:58. 1930; Sci. Monthly 32:242. 1931.-
Becc., Annal. Roy. Bot. Gard., Calcutta (Martelli ed.) 13:
327, pro parte. 1933 ~’1931”].-Small, Manual S. E. Flora
p. 241. 1933.-Bailey, Gent. Herb. 43135. 1938.--Buswell,
U. of Miami Bull. 19(6):49. 1945.-Bailey, Gent. Herb. 8:
991. 1949.-Bomhard, U. S. Dept. Agric. Inf. Bull. no. 22,
p. 21. 1950.-Asprey & Lor-eless, Journ. Ecology 46:547-570.
1958.-Alain, Principes 5:68. 1961.-Long pi Lakela, F1.
Trop. Florida p. 245. 1971.
T. excelsa sensii Bailey, Gent. Herb. 4:134. 1938 [non Lodd.
ex Griseb., F1. Brit. I V . Ind. p. 515. 18641.
T. nzultiflora sensu Read in Adams, F1. P1. Jam. p. 76. 1972
[non Mart., Hist. Nat. Palm. 33255, t. 103, f.l.a., t. 163. 18381.
Medium palm, (1.5-)2.0-10.0 (-12.0) m tall,
caudex columnar, 8.1-1 1.6 (-13.0) ern in diameter,
tapering out to an enlarged base and mound of
roots, internodes 4-7 ern apart on lower trunk.
Leaf sheath 58.4-62.2 ern long, apex oblique
down from petiole, tearing away from petiole and
soon becoming coarsely fibrous, indumentum persistent
on fibers; petiole 36-94 ern long, adaxially
flat to slightly convex, 2.2-3.1 cm wide at sheath,
rounded abaxially, 1.6-2.3 cm wide at narrowest
and 2.1-3.2 ern wide at apex, abaxially with scattered
scales, soon glabrescent or with few persistent
78 SMITHSONIAN COSTRIBUTIONS TO BOTANY
scales; hastula acuminate or sharply triangular, 1 .O-
1.9 cm long, often tending to decurve or split with
the expansion of the blade, abaxially a short (0.1-
0.4 cm long, flap or ridge, often with one or two
rounded or toothlike prominences; unopened blade
glabrescent on the adaxial nerves, scattered floccose
scales on abaxial nerves and exposed surfaces.
Leaf blade, when expanded, circular in outline,
1.2-1.6 m in diameter, abaxial surface lighter in
color than adaxial surface, abaxially dull, often
glaucescent, flecked with conspicuous gray lepidia,
zone around hastula yellow green, translucent when
viewed from below against the light; principal
nerves yellow green; palman 30-55 cm broad, relatively
flat or slightly folded; segments 51-63 in
number, narrowly trullate, often acuminate, with
long slender apical filaments which are readily
lost, (73-)79-113 (-1 15) cm long, (4.6-)5.1-6.2
(-6.4) cm wide, widest at sinus or rarely slightly
above; fusion 30-55 cm long in palman, relatively
uniform throughout blade, but gradually longer
toward upper middle; center tJvo segments near
apex of blade often fused throughout their length,
basal segments commonly overlapping.
Inflorescence erect, or arching slightly at anthesis,
1.2-2.2 m long, bracts green, with white or gray
scales, 13-2 1 glabrous primary branches, the lowermost
primary branches with (24-)27-37 ultimate
branches, these white at anthesis, tip to 12 cm long;
pedicels at anthesis (1.4-) 1.6-2.3 (-2.7) mm long;
flowers white, fragrant, 46-56 per branch; stamens
(5-)6-8 (-10) in number, anthers linear (2.0-)2.2-
8.5 (-3.6) nim long when fresh, (1.7-)1.8-2.3 (-2.5)
mm long when dry, filaments very narrowly connate
basally, angle of connation acute or rounded.
Fruit 7.0-8.2 mm in diameter smooth, ivhite at
fresh maturity, green and slightly papillate shortly
before; seed light brown 6.0-6.8 mm in diameter
completely perforated; primary branches yello.izish
green at fruit maturity, pedicels (1.2-)2.0-5.0 (-5.2)
mm long. Haploid chromosome number, n= 18.
SPECIMENS Ex.z-\rrsFD.-J.ARi.L\IC..I. ST. THOW~~. RfOmnt
Point along the road to the lighthouse, on limestone near
It‘hite Bay: Mar. 19.54, G. R. Proctor 8393 (11); Aug. 1955,
G. R. Prtoctor 10413 (IJ); Aug. 19.54, G. Tl’ehster & K. Wilson
5236 (IJ); Apr. 1963, R. TT’. Read L- G. R. Proctor 1142, 1143,
I144 (FTG); Feh. 1967, R. IT’. Rend 1791 (BH, US); July
1967, R. TI‘. Read 1959, 1960, 1961, 1962 (IJ, UCTVI, US).
Holland Bay: 1952, G. I;. Asprey 2879 (UCTVI), ST. ~ I . \ R Y .
Galina Point on dry coastal limestone, forming a woodland:
date unknown, G. R. Proctor 15546 (IJ); Apr. 1956, TV. T.
Stearn 775 (BH, UCTt’I); .4pr. 1956, A. R. Loveless 2879A
(LCT\’I): without date, Collins 153 (US). Crab Tt’oods, west
of Galina: Xov. 1965, R. 15’. Read 1553 3 1554 (US). ST.
.ANs. Queen’s Highway between Discovery Bay and Bengal
Bridge: Aug. 1935, G. R. Proctor 10568 (IJ); Feb. 1957, G. R.
Proctor I6152 (IJ). St. Ann’s na!: Feb. 1931, L. H. Bailey
15058 (BH). TRELAWKY. Near Salt Marsh on sea side of
road: Apr. 1963, R. W. Read 1149 (FTG). HANOVER. \Test of
Lucea near the sea: Apr. 1963, R. 14’. Read 1151 (FTG).
Rutlands Pen, near Bloody Bay: Mar. 1966, R. Ti.. Read 1611
(us). ~t’rST3fOREI.AND. Near Negril Lighthouse: Nov. 1953,
G. R. Proctor 11132 (IJ). Negril, on rocks near the sea:
Mar. 1962, C. D. Adoms 1097s (CCTt’I). ST. ELIZABI’TH. Long
Acre Point, west of Black River: Sept. 1907, TI’. Harris 9957
(ITCII’I, US). Loczrnu U~~KSOIYK. “,7nninicn. Robitis” s. n.
@%hi, lectotype of T. innrtii Griseb. ct H. Td’endl. ex Griseb.).
H.AITI. LA GoNxvr. Pointe Latanier: Aug. 1927, E. L.
Emrm H. 8841 (D.4, IJ, S ) ; “E troite”: Aug. 1927, TI’. J.
Eye~dnn7 292 (:IS), DEPT. DU SCD. Massif de la Hotte: western
gronp, Les Roseans, Oct. 1928, E. I.. Ekninn H. 10782 ( S ) .
CUB.4. PROV. PINAR DFL Rro. Sear Sahalo: Soy. 1920, E. 1,.
Eknizln 11443 ( S ) . Las Pozas: Apr. 1934, G. Saleson for Ledn
16102 (BH). PROV. HARANA. Vedado: 1929. Brother Ledn
14191 (BH). Bataban6: Nov. 1904, J’nn Herniann 3928 (NY,
ITS), Playa dc Marianao: Feb. 1910, S. L. Britto77 2 P. Ti’ilson
4557 (US); Mar, 1929, I-. H. Bailey I2529 (BH). RIanano (?):
scattered on high sand dunes, July 190.5, J’nn Hermaniz 898
(NY, LS). PROV. CAMAGCEY. Cayo Gnajaha: Nov. 1909, J . A.
.Shnfer 284 6 (US). PROY. LAS Vr1.1.~~. Punta Diahlo, Cienfuegos
Bay: Mar. 1910, S. I.. Britton L- P. TI’ilson 6044 (US). PROV.
ORIESTE. Punta tle hlaisi: Mar. 1931, L. H. Baiiey 15145
(BH); Oct. 1934, Herinnno Ledn 16192 (BH). Manati: June
1932, Herniniio Ledx 15670 (BH). North of “Raimon,” Nipe
Bay: May 1909, I. A . Shafer 1785 (US). “In Cuba Orientali,
1859, 1860”: C. Wright 2329 (AAH, GH, NY, US, syntypes
of T. rclendlondinna); 1860-1861, C. Ti’right 3219 (G, lectoty’e
of T. zi’enrlla~zdiona, F, CH, K, isotypes). Vicinity of
P[B]aracoa: Feb. 1902, C. I.. Pollard, E. Poliner 3 It’. Palmer
192 (US). ISIA DE PINOS. Boqueron, Ensenada de Siguanea:
Feb. 1916, S. I.. Bt-itton, P. T4’ilson &- A. D. Selby 14491
(US). Playa Bibijagua: Feb. 1956, E. P. Killip 45604 (US).
Downs of Progreso: Dec. 186.5, Schott
293 (US); (without date), G. F. Gnu,,ier 1100 (NY, U‘3); May
19,52, H. E. Moore 6366 (BH). Santa Clara: 1959, S. Kiem 329
(BH). Izamal: hiar. 1916, G. I;. Gntciner & So77s 23317 (NY,
US); Rlar. 1916, G. F. Gaictner L- Sons 23265 (NY). QUISTANA
Roo. Puerto Juarez, sand dunes beside sea opposite Isla
Mujer: Oct. 1959, H. E. dloore 8086 (BH).
CESTR.AL AMERICA. COUNTRY UNKSOWS. 1886, G. F.
Gaunier s. n. (US). BRITISH HONDURAS. All pines: Aug. 1930,
TI’. A. Schipp 661 (NY), Belize District: hiile 29 Cayo Road,
May 1960, S. Kieni 605 (BH). HONDI‘RU. Forest near Bolet’s
plantation: Feb. 1903. P. Wilson 360 (NY).
FLORIDA. DADE COUNTY. “Cultivated at Miami from Long
Key”: SOY. 1898, C. S. Snrgent no. 2 (.-IAH, holotype of 7.
floridnnn); So\-. 1898, C. .S. Sargent 202 (NY). Miami, cult.
KLMBER 19 79
from Keys: 0ct.-Nov. 1901, J . K. Small 5 G. V. S a s h 347
(NY, US). COLLIER COUNTY. Cape Romano: May 1898, C. S.
Sargent s. n. (AAH). Caximbas Island: Apr. 1892, J. H. Simpson
180 (NY). Cape Sable: Oct. 1858, A . P. Gnrber 43 (NY,
US). Northwest Cape: Apr. 1916, J . K. Suiall 7699 (NY, CS).
MONROE Couvru. Sands Key: Xiar. 1904, N. L. Britton 344
(NY). Elliott Key: (cult. at Little River, Miami), Apr. 1932,
0. F. Cook s. n. (US). Key Largo: Jan. 1909, J . K. Small AJ
. J . Carter 2953 (NY). Lower Rlatccutnbe Key: .ipr. 1932,
0. F. Cooke 3 J. T. Presley s. n. (US). Long Key: Jan. 1963,
R. TV. Rend 799 (US). Grassy Key: Feb. 1937, E. P. Killil,
32109 (US). Vacca Key: Jan. 1909, J . K. Smnll L J . J . Carter
2959 (91’). Boot Key: Apr. 1909, S. L. Britton 542 (NY).
Big Pinc Key: Jan. 1952, E. P. Kiilil, f l S f 2 (US). Torch Key:
May 1890, A . H. Curtiss s. n. (AAH). TVater Key: .ipr. 1932,
0. F. Cook L- J . T. Presley s. n. (SSC, SY, US). Racoon Key:
Mar. 1898, C. L. Pollard, G. S. Collins L- E. I-. Morris 101
(SY, US). Stock Island: Nov. 1955, E. P. Killip 45097 (US).
Area Unknown: Hammock, May 1917, F. T V . Pelinell 9617
(US); Sykes Hammock, Sept. 1929, H. O’Seill s. 1 1 . (US).
BAH.4hf.G: Cat Cay, May 1905, L. J . K. Bruce 3753 (US).
SOURCE UKKNOIVK: Specimen in Herbarium Regiunl
Alonncetzse “E Iiorto . . , kuris” (RI, holotype of T. radiata).
DIscussIox.-The name Thrinax mdiatn is the
second oldest binomial published in the genus
Thrinax. Having first been published by Renato
Desfontaines (1829) ivithout a description (but
with a reference to Martius’ unpublished work), the
name therefore had no botanical standing. The following
year, however, the father and son team of
J. A. and J. H. Schultes took up the name but
cited Loddiges, the English nurseryman, instead of
Desfontaines as the earlier source of the name.
Their short diagnosis, based solely on a small leaf,
also included a reference to “Martius in litt.” and
Desfontaines’ “Cat. h. paris.” Among the specimens
of the Herbarium Regium Monacense at Munich is
a specimen of a juvenile leaf labeled “Thiinax
mdiata Lodd. E. liorto . . . paris” (Figure 50), .Izhich
is undoubtedly the one described and indicated in
the Schultes publication as Thrinax radintn. No
specimens were cited by either Martius or the
Schultes team, but the reference to Rlartius and the
original diagnosis pertaining to a juvenile leaf
clearly indicate the specimen at Munich. A specimen
of an inflorescence studied by Lehn in 1939 at
Paris (Herbario de Houllet) could not have been
seen at the time of the original publication, since
no mention of an inflorescence was made in the
original diagnosis. The specimen at hIunicli, which
I am designating as liolotype, measuring 25 cm
a Coccothrinax on sight; however, it was not possible
until this past year to ascertain exactly its specific
identity. Beccari had annotated the specimen
as “TIi1iiinx ponceana 0. F. Cook?’ but did no
more. FolloJving a thorough study of the laminar
anatomy of all four species of Tliiinax, their various
biotypes and juvenile plants, it is now possible
with known materials to identify unquestionably
the true specific identity of the t)pe specimen of
T. uidiata.
Eight years following the publication by the
Schultes team, Martius’ publication (1838) ap-
FILVRF 49.-.i primary branch of the fruiting cluster with
nearly mature fruits of Thrinax radiatn. across, appears certainly to be either a Thiinrtx or
80 SMITHSONIAX CONTRIBUTIONS TO BOTANY
peared with but little more to aid in the disposition
of the name. The name Thrinax radiata appeared
in Grisebach (1866) and again in Sauvalle (1871)
but with no indication concerning its application
in Cuba. Drude (1887) treated Thrinax mdiata in
his subgenus ‘‘I Eu-Thrinax” along with T. paiuiflora
and T. multiflora. The subgenus was erected
on the basis of seeds, but since no seeds of the
original collection are known and no mention of
seeds or fruits appeared in the original description,
it is surprising to find T. radiata so treated.
The next significant publication concerning
Thrinax radiata appeared when Sargent (1899)
FIGURE 50.-The type specimen of Thrinnx yadiata from the
Herbarium Regium Rlonacense, Munich.
created the new segregate genus Coccothrinax. For
no indicated reason Sargent (p. 89) stated simply
that, “Thrinax argentea R. & S.; Thrinax radiata
R. & S.;” also “belong in Coccothrinax.” Perhaps
he was aware of another specimen of Coccothrinax
flowers which has apparently been erroneously associated
with the T. radiata type specimen. It remained
for K. Schumann (1901) to make the
correct combination in that genus. In 1933, Beccari
treated the name as an “Imperfectly known or
doubtful species.” In 1939, Brother L e h , apparently
unaware of the earlier transfer, published the
“new combination” and elaborated on the original
diagnosis with information from the specimen he
had studied in Paris.
It is obvious that the description of Thyinax
pal uiflol-a, which appeared in Martius (1838), contains
several elements not applicable to the Swartz
concept. Grisebach (1866) recognized this when he
published T. maytii based on “T. pawifioia Mart.
non Sw.” Martius’ literature citations apply to T.
paruiflora Swartz, but references to the leaves being
“subtus parce floccosocanescentibus” (p. 225) and
“Indumentum in pagina laminae inferiore parcum,
e floccis albis decumbentibus constans” (p. 256)
quite likely refer to Coccothrinax. The major portion
of the description, however, applies to Thrinax,
probably T. mdiata. The locality citation: “Crecit
in insula Jamaica, praesertim orientali, colles saxosos
praediligens, nec tamen a locis depressis maritimis
exulans,” clearly indicates T. radiata, as does
pl. 103 (the inflorescence to the left). This illustration
corresponds amazingly well with the Robins’
Jamaica specimen “spadices floridos in herbariis . . .
et Banksiano asservatos” (p. 255 ) from which
much of Martius’ description was clerived.
Since Thyinax martii (Griseb. & Tliendl. was
based primarily on the description and illustration
by Martius (1838), and the “Jamaica Robins” specimen
is the only one in the Banks herbarium which
could represent the material used by Martius, it
seems logical and in the best interest of nomenclatural
stability to select that specimen as lectotype.
In 1907, Beccari founded a new species, Thyinax
wendlandiana, partly based on collections made by
Charles Wright in Cuba. A specimen in the Geneva
Herbarium labeled “Plantae Cubensis Wrightianae
KO. 3219, coll. C. Wright, 1860-1864” was annotated
by Beccari in 1907 with a note stating that
NUhlBEK 19 81
iliright’s number 3219 was used in support of
“Porothrinax pumilio (fructus) , , . Thrinax Rlartii
(flores) . . . Thrinax argentea Mart. (folium) Griseb.
Cat. Cub. p. 221.” This sheet has an annotation
label identifying it as “Thrinax wendlandiana
Becc.” and has also been labeled “T~~LIs” but I am
not certain by whom. Th~is he further contributed
to an already confusing nomenclatural mess.
Although Beccari cited “Thrinax Martii Gris. et
Wendl. in Gris . . , (pro parte?)” for Coccothrinax
martii Becc., he seems to have treated “Th. Martii
Gris. et Wendl.” in its entirety under T. wendlandiana.
Therefore under the rules of priority Beccari
should have taken up the epithet “martii” instead
14The Wright specimens are a source of considerable confusion,
partly because some of the specimens have the same
number, with a series of letters. These specimens must be
examined at length and independently in order to determine
their correct identity.
of using T. wendlandiana. All the Wright material
cited under number 3219 has so far represented but
a single taxon of Thrinax. It appears that Beccari
(pp. 268 and 308) was separating out what he considered
Thrinax elements under T. wendlandiana
and referring the epithet to the genus Coccothrinax.
It is obvious, however, from a study of the specimens
cited, and lectotypified by Wright No. 3219,
that T. wendlandiana is conspecific with T. radiata.
Vasey (1876) reported a palm from Florida, which
he called Tlarinax parviflora. A. W. Chapman
(1878) confirmed this report when he reported the
occurrence of the palm on the “Keys along the
Florida Reef, extending up the west coast as far as
Cape Romano.” When Sargent (1896) published
his first palm study he too used the name T. parviflora.
The illustration (pl. 510) published at the
same time depicted the inflorescence and fruit of
what is presently known as Coccothrinnx argentata,
FIGURE 5l.-Thrinnx radiatn at Morant Point, Jamaica: A, leaf blade showing the broad flat
palman, hastular region, and overlapping lower segments; B, plant with inflorescences in fruit
with seberal new ones emerging from the crown.
82 SSIITHSONIAN CONTRIBUTIONS TO BOTANY
FIGURE 52.--Thvinnx rndiata in natural habitat on thc Florida Kejs.
NUMBER 19 83
with a leaf of Thiinax in the background. Sargent
(1899) later became convinced that the palm in
Florida was not T. parvipoia Sw., “but distinct
from that species, as collected by Charles iVright
in Cuba (no. 2329) and determined by H. TVendland.”
He therefore published a new binomial, T.
floridam, based on its having “longer and stouter
fruiting pedicels, smaller fruit and deeper seminal
cavity.” This new name again appeared in his
revised palm study (1902), along with an illustration
(fig. 735) clearly depicting tlie palm natile to
Florida.
Beccari accepted Sargent’s new species and included
it in his 1907 publication. It is important
to note here that two names, which in the present
study are considered synonjnious, appeared in Keccari’s
key separated from one another and from
T. pa~uifiot-a by the use of such characters as
perianth-lobe form, anther length, fruit, and seed
size. Several noncomparative quantitative chaiacters
were also mentioned. As Small (1925) pointed
out, “the diagnostic characters on record for the
distinguishing of the two supposed species-Tlzi 2nax
pot idan(i and T. T.T’eiidlnndiann-in Florida, may
be found on the same plant.” Beccari (1933) elen
went so far as to state that his T. iuendlandznna
also occurred in southern Florida at hladeira Hummock
and at Pumpkin Key.
In 1925, Small wrote that, “judging from recent
information and study,” the Florida taxon “proves
to be the same as Tlwinax paivifloin of the TVest
Indies.” Thus coming full circle to the original
incorrect determination by Vasey half a century
earlier. Although Small had not arrived at the correct
conclusion concerning the name of the taxon,
he had, however, concluded that the reason for the
confusion lay in the fact that “decisions were based
on too scanty material, where ample specimens
should haxe been or could have been secured”; and,
“secondl), the main studies were made in the closet
instead of in the field.”
When Bailey (1938) published his study of
Thyinax, he followed Small’s application of T.
pclJ7lipOJN. Although Baile) used the incorrect
name, he was the first to recogniie that the palms
collected and described under various names from
the littoral of Haiti, Jamaica, Cuba, Florida, The
Bahamas, LIexico, and Bri tish Honduras were members
of a single taxon. TVhen Bailey examined the
three specimens of “ T h i i n a u pnivipmn” in the collection
of tlie Stockholm herbarium, he believed
them to represent a single taxon; however, he selected
one sheet about which he stated: “That
sheet undoubtedly may be considered authentic
Thrinax pa~uzfloi a of Swartz.” It was unfortunate
that he chose to draw his information and characterization
of that species from an altogether different
sheet, for the sheet Bailey chose (1938: fig. 81)
contained a large portion of an inflorescence (figure
4 5 ~ ) which was labeled as “Sabal uinbraculifera”
as recently as 1907. (See discussion of S~varti’
type and typification under T. paivipoi a.
Bailey’s key (1938) to the species of Thiznnx
utili7etl quantitative characters such as siie of fruit
and leaf segments, and differences in the diied epidermis
of the fully developed but not nccessarily
“mature” fruit. The so-called differences in the
abaxial surface of the leaf vary on a single plant
depending on the age of the leaf selected. The
hastula was mentioned in support of “T. paivzpoia”
and the “spathelets” were mentioned in support of
“T. excelsa,” but they are not reciprocally compared.
Thyinax indintn, although distributed over
1000 miles across the northern Caribbean and on
many different islands, varies slightly in quantitati\
e characters but maintains its peculiar individuality
as a single species. 41though Baile) found no
appreciable differences between the Florida plants
and those of Cuba, Haiti, and Central America, he
nevertheless selected a plant collected by Britton
in coastal thickets of Jamaica as typical of T. excelsn.
At the turn of the century misconceptions and
confusion regarding T h ~ i n a x pervaded every attempt
to revise the genus. Britton collected Thi inax
indzatn in both the Bahamas and Jamaica, and
identified the plants as T. parviporn, although he
(1908) stated that Tl71217nx excelsn “from Jamaica,
is abundant on that island and distinct from T.
pa?v~floin, to nliich it is doubtfully ieferred b)
Beccaii.” It is difficult to know what plants were
considered T. exrplcn b) Britton, for his specimens
from Morant Point, Jamaica, were identified as
T. pnivipoia, and the specimens from the John
Crow hIountains formed the t) pe of his new species
lsVarying reports of the color of the abasial leaf surface
seem to stem from different impressions and the use of vague
descriptivc terminology such as “indifferently lighter colored,”
“whitish or silvery,” “glaucous or whitish,” “silvery-lvhite,”
all of which havc at one time or another been applied to
the Florida taxon.
84 SMITHSONIAN COSTRIBUTIONS TO BOTANY
T. rex. Bailey accepted Britton’s identification of
the Bahamian material but referred the Jamaican
collections to T. excelsn. Bailey was unfamiliar with
“T. excelsa” in Jamaica, for he wrote in 1938 that
the taxon was
known to me as a living plant only in cultilation in Jamaica,
Cuba, Trinidad, Ciudad Bolivar on the Orinoco, and British
Guiana. It is conspicuous in its ample bright green or
)ellowish-green leakes lighter coloied underneath, full broad
segments, and white or nearly white soft-fleshed bead-like
large berries; petiole broad, obtuse at maigins and compressed.
It is a handsome palm. I have seen it in cultivation
under the names T. parviflota and T. poridana.
While working on the genus in Jamaica the
present author decided on the basis of available
information that the taxon under consideration
was actually conspecific with Martius’ Thrinax
mzilfiflora. It was clear then, as now, that the text
and plates published in support of T. mziltiflol-a
contained several taxa. Since no type specimen was
known at the time, an attempt to sort out the recognizable
taxa resulted in the designation of the only
known species of Thrinax from Haiti (excluding
Navassa) as T. muZfiflorn and relegating the miscellaneous
described or figured taxa to other names.
Alas a type specimen does exist, with a label in
what appears certainly to be Martius’ own hand, in
the Herbarium Regium llonacense, Munich. There
is now no doubt about its being a Coccothiinax,
but not the same as the one identified as Thiinnx
aigentea by Martius.
COMMON NAJfEs.-This palm has been known in
Florida under a variety of names such as Florida
thatch, Jamaican thatch, or just plain thatch. Small
(1925) called the palm “silk-top thatch.” In Jamaica
the palm may be called sea thatch or simply thatch,
pronounced “tatch.” In Cuba the names “guano tle
costa,” “miraguano de lana,” and “guano campeche”
were reported by Le6n (1946). The species
may also be known as “latanier-la-mer” or “guanillo”
in Haiti and the Dominican Republic. X
note on a specimen collected in 1959, by Dr. H. E.
Moore indicates that the palm was called “chit” at
the Downs of Progreso, Mexico, where it was
collected.
3. Thrinax excelsa Lodd. ex Grisebach
Thrinax excelsa Lodd. ex Griseb., F1. Brit. IV. Ind. p. 515.
1864.-Hook., Bot. Mag. t. 7088. 1889.-Read in Adams,
F1. P1. Jam. pp. 75-76. 1972. [Type: Specimen from the
plant illustrated in Bot. Mag. t. 7088 (K).]
T. rex Britt. & Harris, Bull. Torr. Bot. Club 373352. 1910.-
Howard, R. .4., Principes 4:133. 1960. [Type: Britton 4151.
“Eastern slopes of John Crow Mountains at 450 to 600
meters altitude,” Jamaica (NY).]
Large palm, 3-11 m tall, caudex columnar 12.5-
16.0 (-20.0) cm in diameter; leaf sheath 70-90 cm
long, apex long linguiform, soon tearing away from
petiole, not netlike, densely tan velvety throughout;
petiole (1.2-) 1.6-2.2 m long, densely persistent
white to tan velvety lepidote abaxially, 4.0-
4.5 cm wide at sheath, 2.5-3.9 cm wide at apex;
hastula irregularly truncate to broadly acute 1.8-
2.5 cm long, 3.4-3.9 cm wide, abaxially often with
a retuse prominence to 4 mm long near the middle;
unexpanded blade portion densely white to tan
lepidote on all surfaces.
Leaf blade when expanded (2.0-)2.3-3.5 in in
diameter circular in outline but with the outer
segments not overlapping, free portion of segments
drooping slightly; palman relatively flat 45-100 cm
broad; segments (52-) 55-65 in number, narrowly
obtrullate, 114-172 cm long, 4.4-7.4 cm wide, widest
at the sinus, fused for 45-100 cm in the palman,
rather uniformly throughout the blade, slightly
longer toward the upper center, bifid at apex, 3.10
cm deep, abaxial surface white to silvery densely
lepidote tvitli interlocking fimbriate hyaline scales.
Inflorescences 140-1 70 cm long, always arching
at anthesis, primary bracts brown with rufous appressed
scales, primary branches (10-)15-17 in
number, about 27-30 cm long at anthesis, with 37-
47 rachillae each, glabrous, pink to purple at anthesis;
flowers rosy pink, fragrant (odor of rum and
spice), pedicelled, 90-1 14 per rachilla, bracteoles
shorter than the pedicels; stamens (6-)7-10 (-1 1)
in number, anthers (1.2-)1.4-1.9 (-2.1) mm long,
exceeding pistil in length, stigmatic area a simple
slit in the apex of the style.
Fruit 8.0-10.1 (-11.0) mm in diameter, pedicels
0.4-3.5 mm long; rachillae LIP to 17.7 cm long,
lolvermost 3.7-4.5 mm thick at the point of insertion
on rachis; seed 5.9-6.0 in diameter, completely
intruded from base to apex by testa. Haploid chromosome
number, n= 18.
SPECIMFNS Ex~~~II\.ED.-J.~X.I.;\TC.J. ST. THOSWS. “It’oodlands,
eastern slopes of south end on John Crow Mountains,”
“520 m alt.”: hlar. 1909, S. L. nritton 4151 (NY, lectotype:
Thri~7a.v rex); Mar. 1909, Ti‘m. Harris 2 A’. L. Britton I0759
NUhIBER 19 85
(NY, UCWI, US; syntypes: Thiinau reu). Southwestern slopes
of Winchester Peak, between IVheeleifield and Johnson
Mountain: Feb. 1966, R. W. Read 1581 (CS); Mar. 1966,
R. W. Read 1625 (BH, K, UCIVI); Mar. 1966, R. HI. Read
1621 (US); July 1966, R. TV. Read 1682 (BH, IJ, K. UCIVI);
Feb. 1967, R. TY. Read Ii92 (US); 29 Aug. 1967, R. TV. Rend
.5 H. E. Moore 1963 (BH, US). PORTLA~D. Uncommon Hill
abote Fruitful Vale: Apr. 1934. G. R. Proctor 8557 (IJ).
Eastern slopes of John Crow Mountains aboie Ecclesdown:
Aug. 1955, G. R. Proctor lOf83 (IJ); hlar. 1961, C. D. Adanis
9387 (UCIVI); Apr. 1963, R. TV. Read 3 G. R. Proctor 1145
(FTG).
Ct LTIV4TION: Kew Gardens, May 1889 (K, holot.ipe: specimen
fiom plant illustrated in Dot. Mag. pl. 7088).
DIscussIox.-The specific epithet excelsa first appeared,
without a description, in a catalog of palms
published in 1845 by Conrad Loddiges & Sons,
nurserymen. This very old firm, at Hackney near
London, introduced and gave names to a number
of palms up to that time. Many ol the names that
appeared in their earlier catalogs had been taken
up by several botanists before 1845; however, as
far as can be determined the palms were all juvenile
plants, and it was seldom possible to identify
the names with living plants. The Loddiges catalog
attributed T. excelsa to Cayenne.
August H. Grisebach (1864) published the name
Thrinax excelsa Lodd. with a description based on
a plant grown at Kew and notes supplied by Dr.
Joseph Hooker. Presumably the plant at Kew was
already labeled with the Loddiges name. Thrinax
excelsa was described by Grisebach as being 7 feet
7 inches tall and had a trunk 8 inches thick at the
FIGURE 53.--Thrinax excelsa: A, in its natural habitat on the John Crow Mountains of eastern
Jamaica; B, an alerage-size leaf held by IYinston Barrett, an assistant to George Pioctor of the
Institute of Jamaica.
86 SMITHSONIAN CONTRIBUTIONS TO BOTANY
time it flotvered at Kew. The Kew Report for 1882
recorded that the plant of T. excelsa had grown to
rusty-tomentose; berry globose, 3”’ diam.-Hab. Jamaica!,
Hart. Kew. -
20 feet tall, an increase of 12 feet 5 inches in about
18 years. Grisebach (p. 515) further described the
palm as having
Hooker (1889: pl. 7088) depicted Thl-inax excelsa
as a plant with a broad circular leaf, white on
the undersurface and havinc a Dink inflorescence. “ I leaves pale-green above, hoary-glaucous beneath by minute, A betlveen plants growing in
appressed down, divisions united to about one-third: “ligule
bluntly deltoid:” sheath densely coated with buff-coloured
cotton. . . . leaves 4’-5’ long, about 50-lid: divisions about
2‘ long, ~!-1f, broad; “1igule 15”) long;” spadix branches
spreading-recurved, its axis 1’ long, naked below; spathe
John Crow Mountains in eastern Jamaica, with
Grisebach’s description and pl. 7088, leaves no
doubt that they are true T. exceZsa. Unfortunately
the Color Of the inflorescence was not mentioned by
FIGURE 54.-Ripe fruit and flowers post anthesis of Thrinax excelsa. Notice the silvery undersurface
of the leaf blade, lower left.
SUhlBER 19 87
Grisebach and little notice was taken of the color
illustration in Botanical Magazine.
It is not known how the nurserymen in the United
States first came to apply the name T. excelsa erroneously
to a form of T. radiata being cultivated
in Florida, but the name has been common among
the nursery trade for many years, up to the present.
IYhen discussing his newly described T. poridana,
Sargent (1899) stated that, “This is the Thrinax
excelsa of Florida nurserymen but not of Grisebach
in Jamaica, . . . the number of stamens is said to be
nine although this fact does not appear in the
recent description of the species in the Botanical
Mngazzne (115:pl. 7088).”
In 1907, Beccari apparently became confused by
the name T. excelsa as it appeared in the literature.
He listed T. excelsa, of Botanical Magazine t. 7088,
as equivalent to Coccothrinax argentea with a
question mark, no doubt because of the white
undersurface of the leaf. In the same list of names
excluded from the genus he stated that the T. excelsa
listed by Wendland in Kerchove de Denterghem
(1878) was said to have originated in French
Guiana, although nothing in that work has been
found to confirm this. Then as a final gesture to
what is without a doubt the most magnificent palm
in the genus, he relegated “T. excelsa (Lodd.?)
Grisebach” to synonymy with Thrinax parviflora.
A year after the publication of Beccari’s work,
Britton (1908) wrote in a review of the work that
“Thrinax excelsa Lodd., as described by Grisebach,
from Jamaica, is abundant on that island and distinct
from T. fiarviflora, to which it is doubtfully
referred by Beccari.” Britton’s criticism of Beccari’s
work for the most part was just, but in regard to
Thrinax, which he referred to as “probably the
most difficult of the American genera,” he did not
understand the complexities of the genus. Beccari
did not “doubtfully” refer the species to T. p a m -
flora, he only questioned the Loddiges authorship.
Britton was correct that the species is distinct from
T. fiarviflol-a, although present indications are that
he was not thinking of T. parviflora Sw. but of other
authors, and what he thought to be true T. excelsa
of Jamaica is in fact the same T. excelsa of Florida
nurserymen (i.e., T. mdiata). Britton and IYm.
Harris (1910) jointly published a “Hitherto Undescribed
Species of Jamaica,” Thrinax rex, discovered
on the “Eastern slopes of the John Crow
Xlountains at 450 to 600 meters altitude.”
In the report on the Jamaican collecting trip,
Britton wrote (1909: 102) that
the most noteworthy tree is a magnificent fan-leaped palm
of the genus Thrinnx, with leaves of young plants over ten
feet in diameter, the old trees with trunks up to 60 feet
high or perhaps even higher and nearly a foot in diameter,
t i d y the monarch of the genus, and a splendid addition to
our knowledge of it.
In the 1933 posthumous publication of Beccari’s
Corypheae, the only alteration in the disposition of
“Imperfectly known species” was the citation
“Thrinax excelsa Lodd. ex. Grisebach F1. Brit. W.
Ind. 515. = Thrinax Harrisiana. Becc.?’ If indeed
this were so then the converse would be more correct,
Thrinax harrisiana Becc. would be a synonym
of T. excelsa because it is a later name. Thus it
remained in the confusion that has persisted up to
the present.
Bailey (1938) accepted Thrinax rex without question,
and the name T. excelsa was used for an altogether
different taxon in Jamaica. The taxon to
which Bailey applied T. excelsa is the same as that
mentioned earlier by Sargent as the T. excelsa of
the nursery trade and if not identical to the Florida
taxon at least of the same species. (For a discussion
of Bailey’s use of T. parvipora and T. excelsa see
the discussion of T. radiata). The specific epithet
excelsa is still commonly misapplied to the common
Thatch Palm in landscaping, nursery trade, and
botanical gardens around the world.
The palm described so enthusiastically by Britton
(1909) and Britton and Harris (1910) was not
collected again until 1951 when George R. Proctor
of the Institute of Jamaica “rediscovered” it while
on a journey into the mountains. It was collected
again in 1959 when Dr. R. 4. Howard OF the Arnold
Arboretum climbed the John Crow Mountains in
search of living material for introduction into cultivation.
He wrote of his experiences in 1960. Although
he felt that ‘IT. tessellata” (T. parviflora)
was “quite common and generally undistinguished,”
his impression concerning T. excelsa was quite different
for he wrote that, “Thrinax rex [sic], however
lives up to its name and even the most
hardened botanical travellers pay quiet homage in
looking up to this giant of the genus” which, he
further commented, “exists in isolated splendor in
one of the most humanly difficult environments of
Jamaica.”
The latest extension of the known range of T.
88 SI\IITHSONIAS COSTRIBUTIONS TO BOTAKY
excelsa was discovered in July of 1973 when the
author and Dr. Ornduff of the University of California
were on a raft trip down the Rio Grande in
Portland, Jamaica. h’umerous large and small plants
were observed growing on the cliffs of limestone
called “Two Rocks” overhanging a section of the
river. The most unusual aspect of this discovery is
that the species is found well below its normal
elevation at about 60 m above sea level and only a
couple miles from the mouth of the Rio Grande
near Port Antonio.
COMMON NAhfEs.-Thyinax excelsa grotvs in a
region of tumbled, jagged crags of honeycomb limestone
where few people manage to enter. Because
of the inaccessibility of the region, the palm is not
well known even among the native people and
therefore lacks a definite common name. Upon questioning
some workmen in the area only one person
knew of a palm on the ridge, which he called
“broad thatch.”
4. Thrinax morrisii H. Wendland
Thrinax morrisii H. IVendl., Gard. Chron. (ser. 3) 11:104, f.
20, 21. 1892.-Britt. 8; Wilson, Sci. Sur. Porto Rico . . .
5 : 117. 1923.-Becc., .4nnal. Roy. Bot. Gard., Calcutta
(Martelli ed.) 13:334. 1933 [“1931”].-Bailey, Gent. Herb.
4:143. 1938. [Type: H. A . Alford NichoIls, Nov. 1891. Blowing
Point, Anguilla, West Indies (K).]
T. microcarpa Sarg., Gard. For. 9:162. 1896; Silva 10:53, t. ~511.
1896; Bot. Gaz. 27:87. 1899; Silva 14:80. 1902.-Small, F1.
S.E. U.S. p. 222. 1903.-Sarg., Manual Trees N. Amer. p.
105. 1905.-Becc., 1t7ebbia 23277. 1907.-Britt., Torreya
8:240. 1908.-Britt. & IVilson, Sci. Stir. Porto Rico. 5:117
(pro parte). 1923.-Becc., Annal. Roy. Bot. Gard., Calcutta
(Martelli ed.) 13:333. 1933 [“1931”].-Bailey, Gent. Herb.
4:141. 1938. [Type: A. H. Curtiss 2679***, June 1879. No
Name Key, Florida (NY).]
,Simpsonin microcarpa (Sarg.) 0. F. Cook, Science (n. s.) 8.5:
333. 1937. [Kom. illeg. Type: Thrinnx microcarfin C. S.
Sargent.]
T. keyensis Sarg., Rot. Gaz. 27:86. 1899; Silw 14:83. 1902.-
Small, F1. S.E. U.S. p. 222. 1903.--Sarg., Manual Trees N.
Amer. p. 104. 1905.-Becc., IVebbia 2:274. 1907.-Britt.,
Torreya 8:240. 1908.-Becc., .4nnal. Roy. Bot. Gard., Calcutta
(Martelli ed.) 13:333. 1933 [“1931”]. [Type: C. S.
Sargent 1886. “North shore of the largest of the Marquesas
group of islands . . . ,” Florida (AAH).]
T. ponceana 0. F. Cook, Bull. Torr. Bot. Club. 28:536. f. 4.5.
19Ol.-Becc., lfebbia 2:282. 1907.-Britt., Torreya 8:240.
1908.-Urban, Synib. Antill. 4:128. 1911.-Becc., Annal.
Roy. Bot. Gard., Calcutta (Martelli ed.) 13:334. 1933
[“1931”]. [Type: 0. F. Cook 1005, “dry limestone hills
which skirt the southern coast . . , to the west of Ponce.”
Puerto Rico-the specimen in the USNM does not have
a number attached, hut the parts are labeled, presumably
by 0. F. Cook, as “Thrinax ponceana CK (Type),” (US).]
T. praeceps 0. F. Cook, Bull. Tom. Bot. Club. 28:536. 1901.-
Becc., Webbia 2:287. 1907.-LJrban, Symb. Antill. 4: 128.
191 1.-Becc., Annal. Roy. Bot. Gard., Calcutta (Martelli
ed.) 13:335. 1933 [“1931”]. [Type: L. M. Underwood r3 R. F.
Griggs 850, 13 July 1901, “precipitous mountain-side , , .
between Utuado and .4recibo . . . ,” Puerto Rico (US).]
T. bahainensis 0. F. Cook in Northrop, Mem. Torr. Bot.
Club. 12:20. 1902.-Britt., Torreya 8:240. 1908. [Type: “Big
Cabbage Creek, Andros Island,” Bahamas, June. KO specimen
was cited hut there is material in the USNM with
a label that reads “No. 668 Thrinax bahamensis CK type,”
collected by Dr. & Mrs. John D. A’orthrop.]
T. drudei Becc., Webbia 2:269. 1907; Pomona Coll. Journ.
Econ. Bot. 3:404. 1913; Annal. Roy. Bot. Gard., Calcutta
(Martelli ed.) 13:332. 1933 [“1931”].-Bailey, Gent. Herb.
4: 148. 1938. [Type: “Cuba. Plantae Cubenses Wrightianae
n. 3963 col. nome di Thrinax mtiltipora nell’ Erbario di
Berlino ed in quello della Harvard University” (B, destroyed,
GH, lectotype).]
T. punctulata Becc., Webbia 2:280. 1907; Pomona Coll. Journ.
Econ. Bot. 3:405. 1913; Annal. Roy. Bot. Gard., Calcutta
(Martelli ed.) 13:334. 1933 [“1931”].-Bailey, Gent. Herb.
4:146. 1938. [Type, Van Hermann no. 4245, 10 December
1904. Prov. Pinar del Rio, Monte Guanajay, Cuba (identified
as Thrinav mtrkiflora) (B, destroyed, NY lectotype).]
T. ekinanii Burret, Sv. Vet. akad. Handl. ser. 3, 6(7):27.
1929.-Bailey, Gent. Herb. 4:147. 1938. [Type: E. L. Ekman
H. 10839, 23 Oct. 1928. Navassa Island, near the lighthouse,
Haiti (identified on sheet as “Thrinnx ruendlnndiann Becc.
var glnuca Ekm. et Burret”) (S).]
Small to medium palm, 1-10.5 in tall; caudex
columnar, slightly tapering upward, ashy gray,
smooth, but leaf and inflorescence scars prominent,
becoming rimose with age, (5-)8-35 cm in diameter
at the base; leaf sheath 28-60 cm long, at first
long linguiform, soon breaking apart, and separating
from the petiole apically, becoming somewhat
netlike, densely tan to white velutinous at first,
soon becoming glabrescent; petiole 27-84 cm long,
at first densely white lepidote abaxially, soon glabrescent,
(1.0-) 1.5-2.0 (-2.1) cm wide at the sheath,
0.7-1.6 (-2.0) cm wide apically; hastula ovateobtuse,
rarely pointed, erect, thin, 0.2-0.8 cm long,
ca 3 cm wide, densely velutinous, ciliate, often with
dense conspicuous tufts of white or silvery indumentum
when first exposed, soon becoming glabrescent;
abaxial hastula rather inconspicuous,
densely, white lepidote; unexpanded bud leaf
(blade portion) densely appressed lepidote on all
surfaces; expanded leaf usually glossy green adaxially;
variously lepidote abaxially, with spreading,
NUMBER 19
interlocking, fimbriate, hyaline scales (especially
when freshly expanded), becoming glabrescent in
age, otherwise glaucous, blue green as a result of
numerous minute white dots in rows associated
with sunken stomata, the abaxial leaf color variously
interpreted, depending on age and exposure;
blade ca 75-150 or more cm in diameter, nearly
circular in outline, the segments rarely lying in the
same place (except in juveniles), the outer or lower
segments often folded up and forward toward the
apex of the blade; palman ca 12-42 cm broad; segments,
rigid or lax, ca 33-58 in number, narrowly
trullate to rhombic, pinched or not about midway;
the longer middle segments 55-75 cm long, (2.3-)
3.5-4.8 cm wide (at widest part), widest at point of
fusion, apically bifid, 2-8 cm deep.
Inflorescence 55-150 cm long, erect or arcuate,
equaling or greatly exceeding the leaves in length;
primary bracts silvery white to tan lepidote, often
with apical tufts of long white scales; primary
branches glabrous, pendant, or tufted-arcuate (when
not fully exposed in some specimens), 9-21 in number,
(8-)9-30 (-35) cm long at anthesis; ultimate
branches 20-50 in number, 5.5-14.5 (-20) cm long,
subtended by a narrowly triangular bract which is
conspicuously tufted apically with white hyaline
scales; flowers white; becoming pale yellowish or
even pale orange with age, situated on inconspicuous
disklike pedicels; bracteoles triangular; perianth
lobes mostly acute, triangular, apiculate, rarely attenuate
or abruptly rather long apiculate; stamens
mostly 6 in number; filaments broadly connate basally
into a cup or ring often equaling the perianth;
anthers (0.8-) 1.0-1.8 (-2.0) mm long, exceeding the
pistil in length; stigmatic area infundibuliform,
often ciliate, the ovary often becoming orange post-
FIGURE %.-Thrinax morrisii: A, leaf sheath: B, leaf segments and a portion of the inflorescence
with its bumplike pedicels. Notice the white undersurface of the folded down segment. (Photographs
from the 0. F. Cook files.)
90 SMITHSONIAX CONTRIBUTIONS TO BOTANY
anthesis: fresh riDe fruit (3.5-)4.0-4.5 (-8.0) mm in L. H. Bailey 1009 (BH). GREES CAY. July 1903, Ti’. C. Coker
\--,. pressed globose’ mahogany brown’ (2’3-)2*7-4’2 FLORIDA. MO%ROE COU~TY. Key Largo: May 1923, J . K.
(-5.0) mm in diameter, mostly Only partly intruded
by a conical cavity basally.
Fmall, C. A. Mosier & J . B. DeTfTinkeler 10950 (US); Jan.
1925 L. H. L- E. 2. Bailey 6162 (BH); Aug. 1929, J. K. Small
SPECIMENS EXAMINED OR CITED.-LESSER ANTILLES. Barbuda.
Near river landing: May 1937, H. E. Box 668 (BH,
US). Southwestern part of island: South of the lagoon, sea
level, Apr. 1956, A . C. Smith 10459 (BH, US). Near sand
dunes: Jan. 1947, L. R. Hutson 2 (BH). ANGUILLA. Blowing
Point: Dec. 1890, D. Morris (K. not seen, syntype of T. morrisiij;
Nov. 1891, H. A . A . Sicholls (K, not seen, lectotype
of T. morrisii); Jan. 1950, I. V e l a 3749 (US). Long Bay Hill:
June 1974, R. TY. Read 74-213 (US).
PUERTO RICO .4ND OFFSHORE ISL.4NDS. PIIFRTO
RICO. Dry limestone hills west of Ponce: 0. F. Cook “1005”
(US, holotype of T. ponceana); Nov. 1902, A . A . Hellpr
6125 (G, US); Jan. 1949, L. H. Bailey 603 (BH). Ponce to
Penuelas: Mar. 1906, Ar. L. Britton L J . F. Coruell I300 (US).
Tallahoa, near Ponce: May 1932, L. H. Bailey 30 & 30s (BH).
Near Fajardo: May 1885, P. Sintenis 1324 (US). Yanco: May
1935, F. H. Sargemt 666 (US). Salinas de Cab0 Rojo: Feb.
1885, P. Sintenis 100 (G). Guanica: Mar. 1915, A’. L. Britton,
J . F. Cowell c!v S. Brown 4917 (US); Jan. 1886, P. Sintenis
3500 (G, US). Road from Utuado to Arecibo: July 1901,
L. M. Underwood k R. F. Griggs 850 (US, holotype of T.
prheceps): Mar. 1914, AT. L. Britton C J . F. Cowell 2049 (US).
South of Arecibo: May 1932, L. H. Bailey 64 (BH). Camny
to Quebradillas: Mar. 1914, AT. L. Britton S J . F. CorL3ell 1953
(US). Lares: June 1901, L. M. Underwood ‘3 R. F. Griggs 38
(US); Apr. 1913, AT. L. Britton, E. G. Britton ck 14’. E. Hess
2750 (US). Between Toa Baja and Espinoza: Apr. 1921, E. E.
Barker s. n. (BH). West of Vega Baja: Mar. 1922, A’. L. Britton,
M. ,S. Brown 6 C. E. Chardon 6811 (US). Corozal: Mar.
1923, N. L. Britton & E. G. Britton 7829 (BH, US). Near
Bayam6n on Mogotes: June 1968, R. W. Read 2058, 2059
(US). Guajateca Beach: Jan. 1962, Bro. Alain 9090 (IJ).
h5oNA ISLAND. hfar. 1955, E. I>. Little L* Af. L. Kuns 16528
(US). VIEQUES ISLAND. Feb. 1914, J. A . Shafer 2796 (US).
TURKS AND C.4ICOS ISL.4NDS. FIVE CAYS. Prolidenciales:
July 1954, G. R. Proctor 9181 (IJ).
BAHAMA ISL.4NDS. GREAT BAHAMA. Apr.-May 1905,
L. J. K. Brace 3709 (US). NEW PROVIDENCE. Nov. 1937, L. H.
Bailey 1054 (BH); Nassau: hfar. 1903, A . H. Curtiss 101 (US):
July 1903, IV. C. Coker 539 (US). HOG ISL.AND (Paradise Island).
Freshwater Pond: Feb., 257 (? syntype of T. bahawensis);
July 1962, R. W. Read 840 (US). ELEUTHERA. Feb.-Mar.
1907, E. G. Britton 6515 (US). ASDROS. Big Cabbage Creek:
Dr. 6 Mrs. J. D. Nortlwop 668 (US, holotype of T. bahainensis).
Margin of Fresh Creek Lagoon: Mar. 1966, E. Y. Dawson
26650 (US). Key near Mastic Key: Kov. 1937, L. H. Bailey
1046 (BH). Near Fresh Creek: J . K. Small 6 J . J . Carter
8757 (US). Deep Creek: Aug.-Sept. 1906, L. J . K. Brace 5183
(US). West Side, in Loggerhead Creek region: h’ov. 1937,
-
& C. A . Mosier “a” & “c” (BH). Lower Matecumbe Key: .4pr.
1932, 0. F. Cook 6 J . T. Presley (US). Long Key: Jan. 1963,
R. TI’. Read 797 (US). Grassy Key: Jan. 1963, R. T V . Read
798 (US). Ram Rod Key: Apr. 1932, 0. F. Cook 6 .I. T.
Presley (US). Saddle Bunch Key: Mar. 1898, C. I-. Pollard,
G. A’. Collins &- E. L. Morris 23 (US). Big Pine Key: Apr.
1932, 0. F. Cook L* J. T. Presley (US); Aug. 1935, I.. H.
nailey 42 (BH); Ang. 1929, J . K. Small & C. A . Mosier “b”
(BH); Feb. 1937, E. P. Killip 32025 (US); Feb. 1935, E. P.
Killip 31436 (US); Jan.-Feb. 1940, R. F. Martin, 1328 (US):
Mar. 1936, E. P. Killil, 31717 (BH US); June 1963, R. W.
Read 923 (US). No Name Key: June 1879, A . H. Cttrtiss
2679”** (NY, holotype of 7 . microcarpa, US); May 1891,
J. H. Simpson 268 (US); 1881, A . H. Curtiss S o . H (NY, US);
June 1895, A. H. Curtiss (US): July 1927, H. O’Seill (US);
Nov. 1929, 0. F. Cook (US); Apr. 1932, 0. F. Cook 2 J . T.
Presky (BH, 1;s). Marquesas Keys: north shore of the largest,
1898, C. ,S. Sargent (AAH, not seen, holotype of T. keyensis);
Hotel, Miami: Feb. 1925, L. H. L* E. 2. Bailey 6350 (BH);
hfar. 1925, L. H. Bailey 6174 (BH). C. T. Simpson’s place,
Little River: Feb. 1925, L. H. Bailey 6342 (BH); .4pr. 1932,
0. F. Cook &,I. T. Presley (US); June 1928, 0. F. Cook (US).
WD.4 Plant Introduction Station, Miami: Feb. 1952, H. E.
Moore 6010 (RH); Sept. 1962, R. TT’. Read 782, 781, (US):
R. I$’. Read 1437 (BH). Fairchild Tropical Garden, Coral
Gables: June 1962, R. TV. Read 771 (US); Apr. 1965, R. TI’.
Read I409 (BH, FTG, Voucher for chromosome nnmber);
June 1969, DeArmand Hull H-41 (BH, Voucher for M Sc.
Thesis). SOURCE UNKNO~VN. Florida 1913 (US).
HAITI. NAvASS.4 ISLAND. Oct. 1928, E. L. Ekinnn H. I0839
( S , holotype of T. ekmanii; US); June 1956, C. R. Proctor
15468 (IJ, US): Feb. 1960 S . Kiem (BH).
CUBA, PRO\,. CAMAGUEY. Cayo Cruz: Oct. 1909, ,I. A. Shafer
2800 (US). Cay0 Romano: Pnnta Jncaro, Oct. 1909, J . A .
Shafer 2602 (US). Cayo Ballenato Grande: Mar. 1909, J . A .
Shafer 947 (US). PROV. MATASZAS. Curva de Gnerro between
Caliso and San Miguel de 10s BaAos: Dec. 1931, Bro. Ledn
15328 (BH). San Miguel de 10s BaAos: Mar.-Apr. 1931, J . T.
Roig (BH). PROV. HABANA. Lomas de Capasse: June 1930,
Bro. Ledn 14583 (BH). Loma de la Janla, Capasse: Dec.
1929, Bro. Ledn 11723 (BH). Loma Peregrina, Capasse: May
1928, Bro. Ledn 13336 (BH). Sikrra de .4nafe: Dec. 1936, Ri-o.
Ledn 16667 (BH, US). Sierra de Esper6n: Sov. 1932, L. H.
Bailey 1088 (BH). PROV. PINAR DEL RIO. Xft. Guanajay: Dec.
1904, Baker L- Van Hermann 4245 (B, destroyed, F; XY,
Lectotype of T. punctulata); 1859-1860, Wright 3965 (AAH;
B, destroyed; F; GH, lectotype of T. dritdei; NY, US). Sierra
de Esper6n (“0 Anafe,” Mt. Gnanajay of Van Herman):
AUg. 1935, I . H. Bailey ff (BH). CULTIVATED. Royal Palm
KUXIBER 19 91
Dec. 1936, Bro. Ledn 16667 (BH). Sierra de Anafe: Dec. 1911,
P. Wilson 11504 (US). Vicinity of Sumidero: July 1912, J . A .
Shafer 13414 (US). Sierra del Sumidero: Mar. 1932, L. Howell
for Bro. Ledn 15563 (BH). La Guira: Aug. 1912, J . A . Shafer
I3752 (US). Hills between Rio Cayaguateje and Sierra Guane:
Nov. 1911, J . A. Shafer 10469 (US). Santa Cruz de 10s Pinos:
Arroyo Piedras, July 1930, Bro. Ledn I4590 (BH, US).
Vinales: Dec. 1930, E. P. KiZlip I3596 (US); Apr. 1930, Bro.
Ledn 14369 (US). Sierra de la Guassa: Mar. 1932, L. Howell
for Bro. Ledn 15562 (BH). Mogote de la Baudea: Dec. 1930,
Bro. Ledn 14721 (BH). Cayos de San Filipe: Oct. 1932, Bro.
Ledn 15860 (BH, CS). Arroyo Mamey-Pan de Quajaib6n:
Aug. 1934, G. Sateson for Ledn 16142 (BH). Sierra del
.41ic6n: Mar. 1932, L. Howell for Bro. Ledn I5561 (BH).
Sierra del Rangel: ArroJ-o Aspiro, Dec. 1930, Bro. Ledn
14724 (BH).
DIscussIoN.-Almost from its conception, Thrinax
mol-l-isii has been treated as a distinct but poorly
known species restricted to two tiny remote islands
at the extreme northeastern margin of the Caribbean
Sea. As a matter of fact it was the most thoroughly
known of any Thl-inax described up to that
time. It was the only one for which the description
and illustrations had been drawn up from excellent
material, including mature leaves, inflorescence,
and mature fruit. Treated simply as a dwarf, narrowly
endemic species on the islands of Anegada
and Anguilla, the name Thrinax morrisii frequently
appeared in the literature but never in synonymy.
It wasn’t even associated with T. Pumilio, a name
which means “small,” from which it was said to
differ (Wendland, 1892: 104) “in the greater number
of segments of the lamina and in the shortness
of the lamina itself.”
Thl-inax mon-isii was first discovered in December
1890 by D. Morris, then Assistant Director of
Kew, near a place called Blowing Point on the
southwestern part of the island of Anguilla. In a
report published by Norris (June 1891:131) the
palm (without a name) was described as “present
in fairly large quantities, and the fan-shaped leaves
were used for thatching native huts.” Morris explained
that “the chief interest attached to this
palm is connected with its dwarf habit. The largest
and apparently most matured specimen did not
measure more than about 30 to 35 inches in height.”
He was unable to find fruit or flowers at the time,
but “30 to 40 plants . . . were carried away . . . to
be established at the Botanical Stations at Antigua
and St. Lucia.” Figure 56 is an old photograph,
possibly Antigua. It is not known if these are from
the original Morris collection, but I suspect that
they are, since the photograph dates from between
1902 and 1931.
Sargent (1896) published a new species of
Thrinax from the Florida Keys which he called T.
microcal-pa in allusion to the small fruits (I/s inch
in diam.). Instead of 30-35 inches, however, as in
T. morl-isii, this new species measured 20-30 feet
in height. He also described (p. 162) a palm growing
on the Marquesas Keys (Florida), to which he
erroneously attributed specimens of fruits with a
“fleshy succulent pericarp.” For this, he later (1899)
published the name T. keyensis. Thl-iizax micl-o-
FIGURE 56.-Small plants of Thrinax morrisii, possibly from
the oiigiiid collection, in the Botanic Garden, Antigua.
(Photo b) M. XI. Joseph, Dominica, in the L. H. Bailey collection
at the Bailm Hoitoiium. Cornell 19373
~~
depicting a group of plants in a botanic garden, I - I
92 S\fITHSONIAN CONTRIBUTIONS TO BOTANY
carpa was said to differ, from this species (p. 87)
“in its taller and more slender trunk without a
basal enlargement, in its shorter spadix and smaller
fruit with darker colored seeds, and in its smaller
and thinner leaves.” But characters such as “taller,”
“more slender,” “shorter,” “smaller,” “darker,” and
“thinner” are not considered adequate in the present
treatment. Sargent (1902:80) separated the two
species in a key only on the basis of seed and leaf
size.
Thrinax praeceps and T. ponceana appeared
simultaneously in 0. F. Cook’s synopsis in 1901.
FIGURE 57.-Thrinax morrisii growing naturally on Big Pine
Key, Florida. (Photograph from the 0. F. Cook collection,
now deposited in the Department of Botany, Smithsonian
Institution.)
Considering the results of studies on Thrinax parviflora
in Jamaica, the following observations by Cook
(1901:537) are certainly of interest:
Several species of Thrinax, of which T. Morrisii Wendland
may serve as an example, have been described chiefly with
reference to the relative size of the leaf segments and the
extent of their separation. If the palms under observation
near Ponce belonged as was believed, all to one species, it
is not only true that the individual Thrinax passes all the
stages from the narrow and grass-like almost completely
separated segments of the very young plant to the more
than half united leaf of the large tree, but it also appears
to be true that under unfavorable conditions a Thrinnx may
not be able to attain full maturity of size and form but
may at the same time produce flowers and seeds. In the narrow
chinks and crevices of the bare rocks were very small,
stunted trees, obviously of great age, while but a few feet
distant a deeper fissure might hold vegetable debris and
moisture sufficient to nourish vigorous specimens several times
the size of their less fortunate companions. The stunted trees
retain in proportion to their size, but apparently with little
reference to their age, the small deeply divided leaves of
young plants and have short few branched inflorescences,
another difference of supposed systematic importance.
Yet examination of the Cook types and his original
diagnoses reveal little more than quantitative
differences between his two new species, and no
reference to T. mol-risii other than that quoted
above. Unpublished notes in the Washington files
left by Dr. Cook (in the Department of Botany,
Smithsonian Institution) state that T. praeceps
“differs from T. ponceana in longer and slightly
more slender and more tomentose leaf sheaths;
smaller, smoother leaves with longer petioles and
smaller, more acute ligule.” These “diagnostic”
characters are, as has been repeatedly stated, the
real problem in palm taxonomy. The only sure way
of separating them remains the admonition by Cook
in the same paper (p. 526) where he stated that
a locality definitely indicated would often go further toward
establishing the identity of a species than much of the
descriptive matter prepared for this purpose. For the present
at least, the geographical idea should be kept uppermost in
systematic studies of the palms, since it is generally much
easier and far more logical to extend the limits of supposed
species, than to cope with the confusion caused by the miscellaneous
reporting of species far outside their natural
ranges.
The test of time has shown that good diagnostic
qualitative characters, as in many other groups of
plants, seem to be the easiest means of distinguishNUMBER
19 93
ing taxa in the Palmae; however, this has required
a thorough understanding of ecological tolerances
and phenotypic plasticity within the species.
The following year (1902) Cook described yet
another new species, this time from the Bahamas.
Based on specimens collected by Northrop at Big
Cabbage Creek, Andros Island, Thrinax bahamensis
was said to compare favorably with Sargent’s
T. keyensis “in view of the short pedicels, distinctly
lobed calyx, broad filaments and short styles.” But
“this species is evidently much smaller in all its
parts than Thrinax Keyensis” (p. 20). And “in
comparison with Thrinax Ponceana . , . the leaves
of the present species are smaller with the petioles
less flattened and more distinctly ribbed on the
upper side near the apex.”
In 1907, Beccari treated ten “good” species and
three dubious species in the genus Thrinax. I now
treat six of the ten good species as conspecific with
T. morrisii. Two of these, T. drudei and T. punctulata,
were described as new and apparently distinguished
solely by the extent of the intrusion of
the testa into the center of the seed, for they were
placed in two subgenera separated by this character.
While seemingly a good distinction betweeen T.
radiata and T. morrisii, this character appears to
be variable in T. parviflora subspecies parviflora
in Jamaica and requires further investigation with
more adequate material in T. morrisii in Cuba, especially
since the fruits upon which the species in
Cuba were based are immature. I find no other
significant differences between the type of T. diudei
(Wright no. 3965) and the type of T. punctulata
(van Hermann no. 4245). Beccari (1913:407) felt
the resemblance between the two species to be great
enough to suggest that ‘I. . . it is necessary to obtain
better specimens with flowers of both to establish
these species on more sure diagnostic characters.”
Descriptions of new species of palms, based on
single incomplete specimens, have usually proved
to be extremely questionable at best.
When N. L. Britton and Percy Wilson (1923)
presented their findings, they did not accept Cook’s
two species but treated them under the earlier
T. microcarpa of Sargent and distinguished this
species from T. morrisii by the (p. 116) “mature
leaves silvery beneath” for the former and “leaves
green on both sides, or faintly glaucescent beneath”
for the latter.
Thrinax ekmanii was described by Max Burret
(1929) based on collections made by E. L. Ekman
on the island of Navassa. A limestone island with
steep cliffs rising high out of the sea, Navassa lies
just off the westernmost penisula of Haiti between
Hispaniola and Jamaica. While Thrinax morrisii
does not occur on either of the larger islands it does
occur naturally, as presently circumscribed including
T. ekmanii, on this tiny island. Interestingly Navassa
supports another palm, Pseudophoenix sargentii.
Although no really good diagnostic characters
are known which can be used to maintain T. ekrnanii
as a distinct species the most conspicuous
thing, characterizing both palm taxa on Navassa,
is the extremely silvery white condition of the
abaxial surface of the leaves. Until further extensive
observations of this character in the field are
accomplished the nature of its variability throughout
the entire range of the species must be treated
as uncertain.
COMMON NAxxEs.-Known variously in the literature
as “brittle-thatch,” in Florida; “palmita”
or “miraguano,” in Cuba; “buffalo-top,” in the
Bahamas or “buffalo-thatch” in the Caicos Islands:
the name “pimetta” was noted on specimens from
Barbuda. Various herbarium specimens collected in
Cuba bear the names, “miraguano de sierra,” “guano
blanco de sierra,” and “guanita de sierra.” It is
called “broom palm” in Anguilla.
Appendix I
Binomials Published in the Genus Thrinax
(Arranged in chronological order)
T. parvipora Sw., Prodr. 57. 1788.
T. radiata Mart. ex Desf., Cat. PI. Hort. Reg. Paris, ed. 3,
T. argentea Lodd. ex Desf. loc. cit. nom. nud.
T. argentea Lodd. ex J. A. & J. H. Schult., Syst. Veg. 7 (2):
1300. 1830. = Coccothrinax argentea (Lodd. ex J. A. &
J. H. Schult.) Sarg. ex. K. Schum.
T. pumilio Lodd. ex J. A. & J. H. Schult. loc. cit. 1301. nom.
incert.
T. pumila Fulch.? ex J. -4. & J. H. Schult. loc. cit. nom. nud.,
in synon.
T. radiasta Lodd. ex J. A. 8s J. H. Schult. loc. cit.
T. gracilis Hort. ex J. A. & J. H. Schult. loc. cit. pro syn.
T. elegans Hort. ex J. A. & J. H. Schult. lac. cit. pro syn.
T. aurantia Fulch. ex J. A. & J. H. Schult. loc. cit. nom. nud.
T. m,ultiflora Mart., Hist. Nat. Palm. 3:255. 1838. = Cocco-
T. aurata Hort. ex Mart. ihid p. 257. pro syn.
T. barbadensi.p Lodd. ex Mart, loc. cit. = Coccothrinax sp.?
[not necessarily C. barbadensis Becc.].
T. inaritima Lodd., Cat. Palm. 1845. nom. nud.
T. montana Lodd. loc. cit. nom. nud.
T. stellata Lodd. loc. cit. nom. nud.
T. ferrziginea Lodd. loc. cit. nom. nud.
T. excelsa Lodd. loc. cit. nom. nnd.
T. mexicana Lodd. loc. cit. nom. nud.
T. rupestris Lodd. loc. cit. nom. nud.
T. ? chitco Mart., Palm. Orhig. p. 45. 1846. = Chelyocarpus
chuco (Mart.) H. E. Moore.
T. yurnguana 4. Rich. ex Sagra, Hist. Fis. Pol. Nat. Cuba, pt.
2. 11:278. 1850. nom. illeg. = Coccothrinax niiragziama
(H. B. K.) Le6n [as to type, not as to plant described,
which = C. yuruguana Le6nI.
T. miraguano Mart., Hist. Nat. Palm. 3:320. 1853. = Coccothrinax
miraguanza (H. B. K.) Le6n.
T. Imiragziaina (H. B. K.) T\'alp., Annal. Bot. Syst. 5:818.
1858.
T. e x c e h Lodd. ex Griseh., F1. Brit. 1%'. Ind. p. 515. 1864.
T. martii Griseh. et H. Wendl. ex Griseh., Cat. PI. Cub. p.
221. 1866. = T. radiamta Lodd. ex J. A. & J. H. Schult.
T. rigida Griseh. et H. Tt'endl. ex Griseh. loc. cit. = Coccothrinax
rigida (Griseb. & H. Tt'endl.) Becc.
T. crinita Griseb. et H. TVendl. ex Sauv., Anal. Acad. Ci.
Habana 7:563. 1871. = Coccothrinax crinita (Griseb. &
H. Wendl.) Becc.
31. 1829. nom. nud.
thrinax sp.
T. acuniinata Griseh. et H. TVendl. ex Sauv. op. cit. = Coc-
T. grandis Hort. ex Kerchove, Les Palm. 258. 1878. nom. nud.
T. garberi Chap., Bot. Gaz. 3: 12. 1878. = Coccothrinax argeti-
T. graniinifolia Hort. ex Kerchove, Illust. Hort. 31:187. 1884.
T. arborea Hort. ex Kerchove, ibid. p. 187. nom. nud.
T. tunicata Hort. ex Kerchove, loc. cit. nom. nud.
T. elegantissima Hort. ex Hook., f. Kew Rept. 1882:66. 1884.
T. inorrisii H. TVendl. ex Anonymous, Gard. Chron. ser. 3,
T. morrisii H. TVendl., Gard. Chron. ser. 3, 11:104. 1892.
T. microcarpa Sarg., Gard. & For. 9:162. 1896. = T. tnorrisii
T. floridana Sarg., Bot. Gaz. 27:84 1899. = T. rndiata Lodd.
T. keyensis Sarg. ihid. p. 86. = T. tnorrisii H. Tt'endl.
T. praecefx 0. F. Cook, Bull. Torr. Bot. Club 28:536. 1901.
T. poncenna 0. F. Cook, loc. cit. = T. tnorrisii H. Tt'endl.
T. bahamensis 0. F. Cook, Mem. Torr. Bot. Club 12:20. 1902.
T. wendlnndiana Becc., TVebbia 23265. 1907. = T. radiata
T. drudei Becc. ihid. p. 269. = T. ttiorrisii H. Tt'endl.
T. tessellnta Becc. op. cit. 271. = T. pnrviflora Sw.
T. punctulata Becc. op. cit. p. 280. = T. morrisii H. Tt'endl.
T. harrisiatin Becc. in Fedde, Repert. 6:94. 1908. = T. parvi-
T. rex Britt. & Harris, Bull. Torr. Bot. Club 37:352. 1910.
T. longistyla Becc. in Urh., Symb. Antill. 7:170, 1912. = Coc-
T. ekrnnnii Burret, Kungl. Sv. Vet-Akad. Handl. ter. 3, S(7):
T. altissima Hort. ex Bailey, Gent. Herb. 4:149. 1938. nom.
T. niirantiaca Fulch.? ex Bailey, ibid. nom. nud.
T. compacta Griseh. & Tt'endl. ex Bailey, ibid. nom. nud.
T. havenensis Hort. ex Bailey, ibid. nom. nud.
cothrinax miraguama (H. B. K.) Becc.
tutu (Jacq.) Bailey.
nom. incert.
pro syn.
10:700. 1891. nom. nud.
H. TVendl.
ex. J. A. & J. H. Schult.
= T. morrisii H. TYendl.
= T. morrisii H. Tt'endl.
Lodd. ex J. 4. & J. H. Schult.
flora Sw.
= T. excelsa Lodd. ex Griseh.
cothrinax sp.
27. 1929. = T. ~norrisii H. TVendl.
nud.
94
Appendix I1
Nomina Incerta et Dubia
Thrinax pumilio Lodd. ex J. A. & J. H. Schult., Linn. Syst.
T. purnila Fulch. ? ex J. A. & J. H. Schult. loc. cit. in synon.
T. pumilio Lodd. Cat. Palm. 1846. nom. nud.
Porothrznav H. TVendl. ex Griseb. Cat. PI. Cub. p. 221. 1866.
Veg. sec. 7 (2):1301. 1830. nom. incert.
nom. nud.
The original description of Thrinax pumilio by
J. A. & J. H. Schultes (1830) was based on material
from a young plant sent to the Munich Herbarium
by Loddiges. No specimen has since been found
there that could be identified with the description.
It could represent almost any palmate genus since
Loddiges included plants of both Coccothrznax and
Thrinax under Thrinax. It is impossible to apply
the very ample description to any known species
or for that matter to ascribe it to any genus. A
description drawn from a young palmate leaf with
only 10 to 14 segments is impossible to identify
lacking a specimen. The statement that a hastula
is not present is disconcerting, for even very young
Thrinax and Coccothrinax exhibit this organ.
Moreover it would seem that a Thrinax leaf blade
2 feet in diameter should have more than 14 segments.
The additional statement (p. 1301) that
T. pumilio “Differt a Thr. parviflol-a et argentea
praesertim defectu prominentiae in centro frondis
a prolongatione stipites” suggests a generic difference
rather than a specific one. The mention of the
embryo must be disregarded, since it came from
Martius and no doubt does not apply to the taxon
in question because it is a juvenile plant by
definition.
Specimens cited at the time of the publication of
Porothrinax pumilio by Grisebach (1866:221) are
merely associated with the name and are not to be
considered as the type. The mention of the seed
having “the albumin centrally perforate, the margin
not at all ruminate” is taken from associated specimens,
not from type material and thus must be
ignored in typifying Thrinax pumilio. Porothrinax
H. Wendl. ex Griseb. was not accompanied by a
separate diagnosis and is therefore not validly published
(4rticles 41 and 42 of the International
Code of Botanical Nomenclature).
Thrinax section Porothrinax ascribed to Drude
(1887) and validated by C. S. Sargent (1896) is a
new name based on Thrinax microcarpa and should
not be confused with the earlier invalid name
published by Grisebach.
95
Literature Cited
Aiton, W.
1789.
Aiton, W. T.
1811.
Asprey, G. F.
1959.
Hortus Kewensis. Volume 3. London.
Horfus Kewensis. 2nd ed., volume 3. London.
Vegetation in the Caribbean Area. Caribbean Quarterly,
5:245-263.
Asprey, G. F., and 4. R. Loveless
The Dry Evergreen Formations of Jamaica, 11: The
Raised Coral Beaches of the North Coast. Journal
1958.
of Ecology, 46:547-570.
Asprey, G. F., and R. G. Robbins
1953. The Vegetation of Jamaica. Ecological Monographs,
23 : 359-4 12.
Ayensu, E. S.
1967. Aerosol OT Solution, an Effective Softener of
Herbarium Specimens for .4natomical Study. Stain
Technology, 42: 155-156.
1938. Thrinax-The Peaberry Palms, Gentes Herbarum,
4:129-149, figures 75-84.
1939. Coccothrinax in the Southern Greater Antilles.
Gentes Herbarum, 43247-259.
1949. Palms Uncertain and New. Gentes Herbarum, 8:
95-99.
The Classification of Tropical American Vegetationtypes.
Ecology, 36:89-100.
Le Palme Americane della Tribu delle Corypheae.
Webbia 2:l-288.
Palmae Novae Antillanae, XXIV. In Refiertorium
Novarum Specierum Regni l’egetnbilis (auctore F.
Fedde), 6:94-95.
The Palms Indigenous to Cuba. The Poinona College
Journal of Economic Botany, 2:402-407, figures
1933. Asiatic Palms-Corypheae. Annals of the Royal
Botanic Garden, Calcutta, 13:327-336, plates 24, 28,
29. [Published posthumously, revised and edited
by Prof. Ugolino hlartelli; dated 19311.
Bailey, L. H.
Beard, J. S.
1955.
Beccari, 0.
1907.
1908.
1912.
164-167.
Bentham, G., and J. D. Hooker
1883. Genera Plantarum. Volume 3.
Bond, J.
1961.
Britton, N. L.
1908. Beccari’s American Palms. Torreya, 8:238-241.
1909. Botanical Exploration in the West Indies and
Florida. Journal of The Sew York Botanical Garden,
10:99-114.
Birds of the West Indies. London.
Britton, N. L., and TY. Harris
1910. Studies of IVest Indian Plants, 14: Hitherto Undescribed
Species of Jamaica. Bulletin of the Torrey
Botanical Club, 37:352.
Britton, N. L. and P. Wilson
1923. Scientific Survey of Porto Rico and the Virgin
Islands, New York Academy of Sciences, 5:116-117.
Browne, P.
1756. The Civil and Xatural History of Tamaica. London.
Burret, M.
1929. Palmae Cubenses et Domingenses .4. CL. E. L.
Ekman. Kungl. Svenska l’etenskapsakndeniiens
Handlingar, series 3, 6 (7):27.
1878. .4n Enumeration of Some Plants. Bomtanical Gazette,
3:12.
1883. Flora of the Southern United States. Supplement,
2nd ed. New York.
1901. A Synopsis of the Palms of Puerto Rico. Bulletin
of the Torrey Botanical Club, 28:.525-569, plate 45.
1902. Palms from the Bahamas. In Alice R. Sorthrop,
Flora of New Providence and Andros. Memoirs of
the Torrey Botanical Club, 12:20-21.
1937. Hurricane Palms in Florida, Including a New
Genus Simpsonin. Science, 85:332-333.
The Xatural His,tory of Palms. Berkeley and Los
iingeles: University of California Press.
Chapman, A. W.
Cook, 0. F.
Corner, E. J. H.
1966.
Desfontaines, R.
1829. Catalogus Plan tarurn Horti Regii Parisiensis. 3rd
ed. Paris.
Drabble, E.
1904. On the Anatomy of the Roots of Palms. Transactions
of the Linnean Society of London, Botany,
series 2, 6:427-490.
Drnde, 0.
1887. Palmae. In Engler and Prantl, Die Satiirlichen
Pflanzenfamilien, 2 (3):34.
Gatin, C. L.
1906. Germination des Palmiers. Annales des Sciences
A’aturelles, Botnnique, Paris, series 9, 3: 192-314.
Beitrzge ziir Anatomie der Palmen- und Pandanaceen-
Tl’urzeln. Botanisches Centralblatt, 83:339-345.
1864. Flora of the British West Indian Islands. London.
1866. Catalogus Plantaruin Cubensiuiii. Leipzig.
1884.
1889.
Gillain, G.
1900.
Grisebach, A. H.
Hooker, J. D.
Kew Report for 1882. London.
Thrinax excelsa. Curtis’s Botanical iVfagazine, tabula
7088.
96
KUMBER 19 97
Hooker, J. D., and T. Thomson
1855. Introductory Essay. In Flora Indica. xv + 285 pages.
London.
Thrinax rex-.4 Regal Palm of Jamaica. Principes,
Howard, R. A.
1960.
Jost, L.
1887.
4:133-137.
Ein Beitrag zur Kenntniss der Athmungsorgane der
Pflanzen. Botanische Zeitztng, 45:600-606; 617-628;
633-642.
Kerchove de Denterghem, 0.
1878. Les Palmiers. Paris.
Kunth, C. S.
1841. Enumeratio Plantarum. Volume 3. Stuttgart and
Turbingen.
L e h , Hno.
1939. Contribncion a1 Estudio de las Palmas de Cuba,
111. Memorias de la Sociedad Cubana de Historia
Natural, 13: 153.
1946. Flora de Cuba, 1. Contribuciones Ocasionales del
Museo de Historia Satitral del Colegio de la Salle,
8: 260-262.
1956. Identifying Coccothrinax. Principes, 1:16.
Ligon, R.
1657. A True L- Exact History of the Island of Barbados.
Linnaens, C.
1753. Species Plantaritm. Stockholm.
1845. Palms in the Collection of Conrad Loddiges and
Sons, Hackney Sear London with Their Native
Countries. London.
Loddiges, C., and Sons
Loveless, A. R., and G. F. .4sprey
1957. The Dry Evergreen Formations of Jamaica, I: The
Limestone Hills of the South Coast. Journal of
Ecology, 45:799-822.
Martius, C. F.
Moore, H. E.
1838. Historia Suturalis Palmarttm. Volume 3. Leipzig.
1963. The Typification and Species of Palma Miller
(1754). Gentes Herbarum, 9:242-244.
Report of a Botanical Mission to the \Test Indies.
Royal Gardens, Kew, B d l e t i n of Miscellaneous Information,
1891:131.
Morris, D.
1891.
Morrow, L.
1965. Morphology and Anatomy of Coryphoid Palm
Flowers. .4 thesis, Cornell University.
Parthasarathy, M. V.
1968. Observations on Metaphloem in the Vegetative
Parts of Palms. American Journal of Botany, 55(10):
1152-1167.
Pfister, R.
1892.
Ray, J.
Beitrag zur vergleichenden Anatomie der Sabaleenblatter.
51 pages. Dissertation, Zurich.
1688.
1963. Palm Chromosomes. Principes, 7:85-88.
1964. Palm Chromosome Studies Facilitated by Pollen
Culture on a Colchicine-lactose Medium. Stain
Technology, 39:99-106.
1965. Chromosome Numbers in the Coryphoideae. Cytologia,
30:385-391.
196515. Palm Chromosomes by Air Mail. Principes, 9:4-10.
1966. Coccothrinax jamaicensis-The Jamaican Silver
Thatch. Principes, 10: 133-141.
1968. A Study of Thrinax in Jamaica. .4 thesis, The University
of the It’est Indies, Mona, Jamaica.
1911. Der Spaltoffnnngsapparat de Plamenblatter. Sitzungsberichte
der Kaiserlichen Akademie der Wissenschaften
. . . Klasse, 1 (120):1049-1086.
1896. Thrinax in Florida. Garden and Forest, 9:162.
1896. Silva of North America. Volume 10. Cambridge,
Mass.: The Riverside Press.
1899. Sew Sorth American Trees. Botanical Gazette, 27:
1902. S i h of A’orth America. I’olume 14. Cambridge,
Historia PEantarum. Volume 2. London.
Read, R. W.
Rudolph, K.
Sargent, C. S.
83-87.
Mass.: The Riverside Press.
Sauvalle, F. .4.
1871. Flora Cnbana. Angles de la Academia de Ciencias
MPrlicas, Fisicas y Saturales de la Habana, 83152.
Schreber, J. D. C. (editor)
1791. Cnroli a Linne‘ . , . Genera Plantarum. Editio
Octavo. Frankfurt.
Schultes, J. A. and J. H. (editors)
1830. Caroli a Linne‘ , . . Systema Vegetabilium. Volume
7, part 2.
Schumann, K.
1901. Neue Arten der Siphonogamen 1899, VIII. Just’s
Botanischer Jahresbericht, 27 (1):469.
Shreve, F.
1914. A Montane Rain-Forest. Carnegie Institution of
Washington, Publication number 199.
Sloane, H.
1696. Catalogus Plantaruni Qztae in Insula Jamaica.
London.
1725. A l’oynge to the Islands Madera, Barbados, Niezies,
S. Christopher. Volume 2. London.
Small, J. K.
192.5. Silk-top Thatch-Thrinax parviflora. Journal of
the A’ew York Botanical Garden, 26:49-54.
Statistical Methods. Iowa State College Press.
Palmae. Volume 3 in Systernatische Anatomie der
Monoktyledonen. Berlin.
Caroli Linnaei , . . Genera Plantai i i m . Editio nona,
volume 2. Gottigen.
Snedecor, G. W.
Solereder, H., and F. J. Meyer
1948.
1928.
Sprengel, C. (editor)
1830.
98 5hIITHSONI4S COSTRIBUTIONS TO BOTANY
Swartz, 0.
1788.
1797.
Xoua Genera 3 Species Plantarum seu Prodromus.
Stockholm, Uppsala.
Flora lndiae Occidentalis. Volume 1. Erlangen.
Tomlinson, P. B.
1960. Essays on the Morphology of Palms, I: Germination
and the Seedling. Principes, 4:56-61.
Palmae. Volume 2 in Anatomy of the Monocotyledons.
Oxford: The Clarendon Press.
1961.
Vasey, G.
1876. A Catalogue of the Forest Tree? of the United
States. U.S. Department of Agriculture Report,
number 11.
Venkatasubban, K. E.
1945. Cytological Studies in Palmae, Part I: Chromosome
Numbers in a Few Species of Palms of British India
and Ceylon. Proceedings of the Indian Academy
of Sciences, 22:193-207.
New or Noteworthy Plants. The Gnrdners’ Chroiiicle,
ser. 3, 11:104, figures 20, 21.
Flrendland, H.
1892.
TVilldetiow, C. L. (editor)
1799. Caroli a Linne‘ . . . Species Plantartc~n. Editio
Quarta. Berlin.
Das Anatomische Verhal ten der Palmenblatter den1
System dieser Familie. 40 pages, Dissertation. Erlangen.
Zawada, K.
1890.
& U.S. GOVERNMENT PRINTING OFFICE 1075-. 546-365!33