SMITHSONIAN CONTRIBUTIONS TO ANTHROPOLOGY VOLUME 8 [Whole Volume] Cultural Chronology of the Gulf of Chiriqui, Panama Olga Linares de Sapir Research Associate, Smithsonian Institution SMITHSONIAN INSTITUTION PRESS City of Washington 1968 A Publication of the SMITHSONIAN INSTITUTION United States National Museum LIBRARY OF CONGRESS CARD 68-60007 UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON, 1968 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 1.04,02. Price $4.50 Preface This field work was part of a much larger program lasting three years entitled "Inter- relationships of New World Cultures" and organized by the Institute of Andean Research with financement from the National Science Foundation (Ekholm and Evans, 1962). Evidence of interconnection during the Formative period and at other points in time was sought in particular regions of the Intermediate area between Mesoamerica and South America. The Panamanian region surveyed is the Pacific coast from the Costa Rican border to the Colombian border. It was covered in two consecutive seasons. The first, in February through May of 1961, included the western coast and principal offshore islands from Punta Burica to Punta Mariato in the southwestern corner of the Azuero Peninsula. The second, in 1962, extended the work to cover coastal Darien. This is a report of the work done in the first season. The participants of the project were Dr. Charles R. McGimsey III of the University of Arkansas, as principal investigator, and two assistants, besides me, who were also graduate students at the time: Mr. William Bishop, geologist, and Mr. Freeman Mobley, ethnologist, both from the University of Arkansas. In four months we visited over 60 sites and tested 12 of these. To each of the sites that we recorded we assigned a letter-number combination. The letters designated the district, and the numbers, the sites in the order found. The site of El Cangrejal was thus given the code number SL-1, with SL standing for the district of San Lorenzo in the province of Chiriqui, and number 1 standing for the first site found in this district. Island sites could not be assigned to districts, so they were simply given a number preceded by the letters IS which stand for "isla." Thus, Isla Villalba became IS-7, Isla Palenque became IS-3, and Islas Las Secas became IS-11. Reported in this study are the four most important sites where the most concentrated work was done. These are the sites mentioned above, all of them located in the Chiriqui Gulf. Three are on islands, the fourth on the mainland. Distances between them vary. The closest to each other are Isla Villalba (IS-7) and El Cangrejal (SL-1), separated by 10 kilometers of water channels. The farthest away is Las Secas (IS-11); it is 37 kilometers from Isla Palenque (IS-3), 53 kilometers from Isla Villalba (IS-7), and 55 kilometers from El Cangrejal (SL-1). The three island sites are located on ridges in back of the beaches, often the highest spot available. The mainland site is on a high bank near the edge of an estuary. All four sites consist of middens formed by the deposition of abundant cultural materials: discarded pottery, stone tools, and the remains of marine and land fauna. These were choice loca- tions and, to judge from the refuse, the occupation of them was continuous. Listed in summary form, the main aims of this report are as follows: (1) To estab- lish a preliminary chronology for the Gulf of Chiriqui, based on local sequences of pottery; (2) to trace, in time and space, the cultural contacts of the inhabitants of these sites with PREFACE each other and with peoples on the mainland; (3) to reconstruct as much as possible the manner of living and the subsistence patterns of the occupants of these sites; (4) to sum- marize ethnohistorical accounts of the area in an effort to tie these in with the termmal part of the archeological sequence; (5) to place the Chiriqui phases in proper chronological relationship to other Panamanian sequences; and (6) to relate the chronology thus estab- lished to those of other lower Central American areas. Two approaches were used in our pottery classification. The first was a subdivision of the decorated sherds into types, differentiated from each other primarily by decorative treatment and secondarily by paste attributes. Their percentage frequencies were plotted by levels. Since the plain pottery does not always appear to show any distinctive differences in temper or paste, it was not divided into the full, conventional pottery type description, but instead was sorted into groups named A-J. Only one plain ware was distinctive enough to be described fully and was labeled Tarrago Bisquit Ware. Perhaps a larger sample from more sites in Chiriqui, or from sites with deeper stratified refuse, will eventu- ally permit the detailed refinement of the plain wares, but for this study they receive secondary treatment even though the percentages were always calculated and actually show some interesting trends in the site seriation charts. There are also some small samples of pottery types that fit into the "Classical" Chiriqui wares known for so many decades from the large museum collections; they are listed under their well-known published names. The pottery type description follows the style established in earlier Smithsonian publications on the archeology of South America, with each type subdivided into the major categories of paste, surface, form, decoration, and chronological position. To these I have added one other category: geographical distribution and comparative materials. For the color identifications of paste or surfaces, words rather than the Munsell Color Chart have been used, for no more precise a code is needed to describe pottery with only simple paints, no polychrome, and variable firing. Hardness is stated in the Mohs scale. It is perhaps interesting to record how the rim profiles were drawn. To facilitate and increase the accuracy of the rim profile drawings, a section from each rim was cut with a blunt-edged carbide geological power saw at right angles to the lip. Unless the rim was badly mutilated, I always cut a section on the right side of the rim exterior so that, when drawn, all rim profiles faced in the same direction. By cutting the rim it was easy to trace the correct profile because it lay flat on the paper. A second method of analysis used in this report, first separately, and then in conjunction with the type approach, is a modal analysis of all appendages (supports and handles) found in each level of the pits. I have chosen to use "mode" to mean "a small group of inseparable attributes" because this is useful in making geographical comparisons. If, for example, a tripod foot shares a cluster of salient attributes with a tripod foot in another area, one can be reasonably certain that some kind of historical relationship, rather than mere chance, accounts for the similarities. In this study, modes have been kept separate from types by using different, though not mutually exclusive, criteria for classification. The main criterion used in the type approach is surface decoration; in the modal approach, the main criterion for distinguishing modes is shape. These criteria could just as well be reversed in another area. The result is that modes and types often crosscut, so that the same mode may appear in more than one type. This may have interesting cultural and chronological implications. The potters, conforming to standardized stylistic patterns (types), were able to isolate features (modes) in this pattern and reproduce them in another context. Chronologically, a mode that crosscuts two types may sometimes indicate con- temporaneity as, for example, with Tripod Foot Mode e, which appears in two different but contemporaneous types. A separate modal approach, based on shape elements rather than on painted motifs, is particularly suited to an analysis of the Gulf of Chiriqui collections because this area belongs to a primarily plastic tradition. Paints and slips were used, to be sure, but vessels of the Gulf were not elaborately painted as they were in the Cocle-Azuero region. Rather, PREFACE it is the shape of handles, supports, and rims, and the manner in which these were decorated with plastic motifs, such as applique strips, incisions, and punctations, that make Chiriqui pottery so enormously varied. Using a modal analysis had advantages of a very practical nature. Handles, feet, and supports are so distinctive that they can be recognized easily in any context. Since the surface collections (Ranere, Appendix 2) were made without any regard to randomness, they contain a disproportionate number of such fragments. A modal approach serves to relate these to the excavated collections. The distribution of certain modes beyond the Chiriqui Gulf has also offered important insights into the influences that came there from Costa Rica and from the central provinces of Panama. In short, modes were used in this study as timemakers and as indices of geographical distribution. From them we could draw inferences concerning wider cultural contacts. I have used several criteria to interdigitate the site sequences and to mark off units of time into phases: (1) the appearance of new types in the stratigraphic columns; (2) marked contrasts in the percentage frequencies of pottery types; (3) the presence or absence of particular modes; (4) the association of types; and (5) the presence of trade wares from known areas. The use of these criteria was facilitated by the deep stratified sequence of Pit No. 3 at Site IS-3. A very distinctive type of pottery was strongly represented in the bottom levels, different types were associated with the intermediate levels, and the most recent levels were marked by the appearance of several new types. In addition, modes were absent from some levels and present in others, and trade pottery from several phases of the Cocle-Azuero area came in at different levels and served to confirm my phase divisions. The other three sites, none of which had the complete sequence represented, were therefore interdigitated with reference to IS-3. The assumption made was that all four sites were close enough geographically to show the same general chronological trends. Contrasts in the percentage frequencies of pottery types proved useful in coordinating pit levels within a site and in differentiating between units of levels in which certain types were heavily represented from units in which they were not. The pottery types are described by phase for easy reference, moving from the oldest to the most recent in the sequence. The modes are described separately. Stone tools and shell ornaments found at these sites were not used in establishing a tentative chronology because they did not show enough changes through time. Along with bones and mollusks, however, they are essential to our cultural reconstructions. A description of surface collections from 26 sites in the adjacent districts of San Felix and Remedios has been appended (Ranere, Appendix 2) to the body of the report, which describes our excavations, to enlarge the scope of this analysis. In this appendix Mr. Ranere has contributed ethnohistorical data and new insights into the archeological material, making his section an important part of this work. I would like to thank first Dr. Charles R. McGimsey III, principal investigator of the project, for inviting me to collaborate in the field work and for asking me to report and interpret the results. The success of the expedition was also due to Mr. William Bishop and Mr. Freeman Mobley, who were indefatigable in their search for sites and acute in their observations of the Panamanian countryside. The authorities of Panama, as well as Dr. Alejandro Mendez, Director of the Museo Nacional, and Dra. Reina Torres de Araiiz, Professor of Anthropology at the University of Panama, greatly supported the work by their kind cooperation. The entire survey was facilitated by cartographic aids provided by Ing. Amado Araiiz of the Subcomit6 del Darien and of the Pan American Highway Program. Ample laboratory space to analyze the collections was generously provided by Professor John Otis Brew, Director of the Peabody Museum of Archeology and Ethnology at Harvard University. A grant from the National Science Foundation and a Thaw Fellowship in the Peabody Museum permitted survival during the summer and fall of 1963, when the first, unrevised version of this monograph was being prepared as a doctoral dissertation (Linares, 1964). PREFACE Drs. George Cowgill, James Giffo-rd, and John Ladd were knowledgeable in their advice on how to handle particular problems in the analysis. Mr. Anthony J. Renere worked with me on the Chiriqui collections and expanded this study in his Appendix No. 2. Mrs. Alexandra W.de Garcia-Brycfe advised me on the latest sources for Latin American Archeology. Miss Margaret Currier of the Peabody Museum Library was helpful in the initial research. Mr. Arsen Charles took the photographs on plate 14. From Professor Gordon R. Willey of Harvard University^ I have received endless encouragement through the years and haost of my knowledge of New World Archeology. My intellectual debt to,him is enormous. Expert advice was provided in 1964 by several investigators of the Museum of Com- parative Zoology, Harvard University. Dr. W. J. Clench, Curator of Mollusks, guided my identifications of the shellfish material; Dr. E. E. Williams, Curator of Reptiles and Am- phibians, sorted out at a glance any reptiles represented on a mass of tiny" fragments. All mammal bones were identified by Mr. Charles Mack. Mr. William Bishop, geologist for the expedition, made the rock identifications while in the field, and the material identifications in the descriptions of stone artifacts are all his; his report on the geology of the coast, from which Mr. Ranere and I have quoted exten- sively, has been of additional value. Mrs. Charles R. McGimsey did the field cataloging of all materials collected and took the photographs of stone artifacts for plates 19 and 20. Dr. Clifford Evans and Dr. Betty J. Meggers of the Smithsonian Institution have spent countless hours over the last years teaching me to classify pottery, making suggestions about plans of research, and helping me in every phase of preparing this manuscript fot publica- tion. I cannot thank them enough for their generous help. Dr. Alexander Wetmore, Research Associate of the Smithsonian Institution, provided various maps from his vast collections that were the basis for the final maps produced in this report. To Mrs. Marcia P. Bakry, Scientific Illustrator of the Office of Anthropology, Museum of Natural History, I wish to express my appreciation for preparing the final copies of the maps and seriation charts. I have profited greatly from conversations with Dr. Claude F. Baudez of the Musee de rHomme, regarding his archeological investigations in nearby Costa Rica. Recently, Dr. Roberto de la Guardia of the Museo Chiricano in David, Panama, brought me up to date on finds in Chiriqui. In typical Latin American style, my work became a true family enterprise. Sr. Frank Linares Danz, my father, loaned a jeep to the expedition, and Sr. Guillermo Tribaldos, Jr., my grandfather, made his boat available to us. Sra. Olga Tribaldos de Linares, my mother, did the drawings of the artifacts, prepared an original version of the pottery type fre- quency charts, and typed this report twice, first as a dissertation, then as a manuscript for publication. Their interest in my work is immensely gratifying. I also thank my husband, J. David Sapir, for his patience and continuous encouragement. O. L. de S. Department of Anthropology University of Pennsylvania Philadelphia March 24, 1967 Contents INTRODUCTION 1 Geography of Chiriqui 1 . History of archeological work in Chiriqui 4 THE SITES . 7 Isla Palenque (IS-3) . . 7 Location ;....,..,. 7 ' Excavations 7 The pottery type sequence 9 El Cangrejal (SL-1) 10 Location 10 Excavations 10 The pottery type sequence 11 Villalba (IS 5-7) 12 Location 12 Excavations 13 The pottery type sequence 14 Las Secas (IS-11) 14 Location 14 Excavations 15 The pottery type sequence 16 ANALYSIS OF CULTURAL REMAINS 17 Burica Phase 19 Isla Palenque Maroon Slipped 19 Unclassified plain wares 23 San Lorenzo Phase 23 Arayo Polished Line 23 Banco Red Line 26 Caco Red Slipped 26 Cangrejal Red Line 27 Castrellon Red Slipped 30 Centeno Red Banded 31 Horconcitos Red Banded 32 Linarte Zoned Red Line 32 Pan de Azucar Red Line 33 Zapote Red Banded 35 Unclassified plain wares 36 Chiriqui Phase 37 Cavada Applique and Red Banded 37 Tarrago Bisquit Ware 38 Villalba Red Streaked ' 41 CONTENTS Page Unclassified pottery types ^^ Alligator Ware '*"' Black-on-Red Negative '*'* Chocolate Incised '*'* Red on White ^^ Red Slipped Bisquit Ware ^^ Unclassified plain wares Appendage modes Supports '*° Tripod feet *^ Mode a ^^ Mode b '^^ Modec 48 Mode e 48 Modef 49 Mode g ^" Mode h 49 Mode i ' ^" Ringstands ^^ Mode a ^0 Mode b, c, d ^1 Pedestal bases 5^ Mode a 51 Modeb 51 Mode c 51 Moded 51 Mode e 51 Modef 51 Mode g 51 Modeh 53 Mode i 53 Modej 53 Modek 53 Model 53 Mode m 53 Strapped feet 53 Modes a, b, c 53 Handles 54 Double handles 54 Mode a 54 Mode b 55 Mode c 55 Strap handles 55 Mode a 55 Mode b 55 Mode c 55 Mode d 56 Round handles 56 Mode a 56 Mode b 56 Ladle handles 57 Horizontal handles 58 Miscellaneous handles 59 Miscellaneous applique motifs 59 Stone artifacts 60 Celts 60 Flakes 60 Manos 61 TABLES Page Metates 61 Net or line weights 61 Pebble polishers 61 Pecking stones 62 Pounding-anvil stones 62 Projectile points 62 Sandstone saws 62 Scrapers 63 Sharpening stone 63 Bone and shell ornaments 63 THE SEQUENCE, CULTURAL INVENTORY, AND WAY OF LIFE OF THE GULF OF CHIRIQUI INHABITANTS 65 The sequence 66 Burica Phase 66 San Lorenzo Phase 66 Chiriqui Phase 66 The cultural inventory and way of life 68 Burica Phase 68 San Lorenzo Phase 68 Chiriqui Phase 72 EUROPEAN CONQUEST AND ASSIMILATION OF THE INDIANS OF CHIRIQUI 75 Early explorations of coastal Chiriqui 75 17th-century explorations of coastal Chiriqui 78 Mainland Indians of Chiriqui 78 Inferences 80 CHRONOLOGICAL SUMMARY AND COMPARISON WITH OTHER AREAS 83 Chronology of Chiriqui 83 Relationships with neighboring Panamanian areas 87 Chronological links with other areas 89 Final remarks 91 LITERATURE CITED 93 APPENDIX 1 (Tables 1-7) 97 APPENDIX 2: Analysis of pottery surface collections from the Pacific districts of Punta Burica, San Felix, and Remedios in Chiriqui, Panama. By Anthony J. Ranere ... 107 PLATES 121 Tables APPENDIX 1 1. Frequency of pottery types from Pit No. 3, Site Isla Palenque (IS-3) 97 2. Frequency of pottery types from Pit No. 1 and Pit No. 3, Site El Cangrejal (SL-1) . . 99 3. Frequency of pottery types from Pit No. 1 and Pit No. 2, Site Villalba (IS-7) .... 100 4. Frequency of pottery types from Pit No. 1, Site Las Secas (IS-11) 101 5. Frequency of miscellaneous sherds, necks, supports, handles, and rims at Sites IS-3, IS-7, IS-11, and SL-1 102 6. Frequency of stone artifacts at Sites IS-3, IS-7, IS-11, and SL-1 103 7. Animal bones from Sites IS-3, IS-11, and SL-1 105 APPENDIX 2 1. Distribution of pottery types in the district of Punta Burica 117 2. Distribution of pottery types in the districts of San Felix and Remedios 118 3. Distribution of modes (handles, supports) in the districts of Punta Burica, San Felix, and Remedios 118 4. Distribution of plastic decorative motifs in the districts of Punta Burica, San Felix, and Remedios 119 xii ILLUSTRATIONS 5. Number of sherds in each pottery type of the test excavation at Site BA-8 in the Punta Burica district Illustrations FIGURES 1. Map of Panama, showing area of research in the Gulf of Chiriqui in relation to selected sites in the Cocle-Azuero area 2. Map of the central part of the Gulf of Chiriqui, showing location of excavations and geographical features 3. Site map of Isla Palenque (IS-3), showing location of excavations and topographical 10 features 4. Pottery type sequence of Pit No. 3, Site Isla Palenque (IS-3) 5. Site map of El Cangrejal (SL-1), showing location of excavations and topographical features ' ^ 6. Pottery type sequence of Pits No. 1 and No. 3, Site El Cangrejal (SL-1) 12 7. Map of Isla Villalba (Isla Muertos), indicating location of excavated Site IS-7 and location of surveyed Sites IS-5 and IS-6 13 8. Pottery type sequences of Pits 1 and 2, Site IS-7 on Isla Villalba (Isla Muertos).... 14 9. Map showing Site IS-11 on Isla Cavada of the Las Secas complex of islands, with location of excavations and geographical features 15 10. Pottery type sequence of Pit No. 1, Site IS-11 on Isla Cavada in Las Secas Island complex 16 11. Rim profiles of Isla Palenque Maroon Slipped, thick variety, Burica Phase 18 12. Rim profiles and side views of Isla Palenque Maroon Slipped, thin variety, Burica Phase. 20 13. Burial urns of Isla Palenque Maroon Slipped, thick variety, from the sites of Las Tinajas and Dolega, reported by other investigators 22 14. Rim profiles of Arayo Polished Line, San Lorenzo Phase 24 15. Painted exterior designs and a few rim profiles of Arayo Polished Line, San Lorenzo Phase 25 16. Reconstructed vessels of Banco Red Line, San Lorenzo Phase 26 17. Rim profiles of Caco Red Slipped, San Lorenzo, and Chiriqui Phases 27 18. Rim profiles and exterior of Cangrejal Red Line, San Lorenzo Phase 29 19. Rim profiles and interiors of Castrellon Red Slipped, San Lorenzo Phase 30 20. Painted designs on the exterior surface and rim profiles of Centeno Red Banded, San Lorenzo Phase 32 21. Type sherds of Horconcitos Red Banded, San Lorenzo Phase 32 22. Body wall profiles and decoration on Linarte Zoned Red Line, San Lorenzo Phase. , 34 23. Type sherds of Pan de Azucar Red Line, San Lorenzo Phase 35 24. Rim profiles and designs on the interior of Zapote Red Banded, San Lorenzo Phase. . 3 6 25. Rim profiles and exteriors of Cavada Applique and Red Banded, Chiriqui Phase. . . 38 26. Rim profiles, side views, and adornos of Tarrago Bisquit Ware, Chiriqui Phase .... 39 27. Rim profiles, exterior and interior views, Villalba Red Streaked, Chiriqui Phase. . . 42 28. Sherd of Alligator Ware, Chiriqui Phase (especially diagnostic of "Classical Chiriqui Culture") 43 29. Sherd of Black-on-Red Negative, Chiriqui Phase (especially diagnosfic of "Classical Chiriqui Culture") 44 30. Tripod Feet Modes from all three Chiriqui Gulf Phases 47 31. Ringstand Modes of San Lorenzo and Chiriqui Phases 50 32. Pedestal Base Modes of the San Lorenzo and Chiriqui Phases 52 33. Strapped Feet Modes of the San Lorenzo and Chiriqui Phases 53 34. Double Handle Modes of the Chiriqui Phase 54 35. Strap Handle Modes of the San Lorenzo and the Chiriqui Phases 56 36. Round Handle Modes of the Chiriqui Phase 57 30. ILLUSTRATIONS Page 37. Ladle Handle Modes of the San Lorenzo Phase 58 38. Horizontal Handle Modes shared by the San Lorenzo and Chiriqui Phases 58 39. Miscellaneous Handle Modes from the San Lorenzo and Chiriqui Phases 58 40. Applique motifs of the San Lorenzo and Chiriqui Phases 59 41. Shell pendants from Las Secas Site (IS-11), Chiriqui Phase 63 42. Seriated sequence of selected stratigraphic pits in the Gulf of Chiriqui area, showing ceramic trends and presence or absence or pottery types in each of the three phases . 66 43. Distribution of Appendage Modes in the San Lorenzo and Chiriqui Phases 67 44. Occurrence of stone artifacts in the San Lorenzo and Chiriqui Phases 69 45. Distribution of moUusk remains by site 70 46. Map of the Gulf of Chiriqui area, showing selected modern town and villages, arche- ological sites on the mainland reported by other investigators, and surface sites in the Burica, San Felix, and Remedios districts 84 47. Chronological chart of regional archeological sequences in western and central Panama, with approximate time periods indicated 88 48. Rim profiles of Balsa Polished, groups 1-3, Burica Peninsula Ill 49. Rim profiles of pottery type Quebrada Baules Brushed, groups 1 and 2, Burica Peninsula Ill 50. Three-Toed-Tripod Foot belonging to Scarified Ware of La Goncepcion complex from Site BA-9 in the Burica Peninsula 112 51. Applique figurine from Site BA-9, Burica Peninsula, resembling figurine from Aguas Buenas complex in Costa Rica 113 52. Polychrome trade pottery at Site BA-8 114 53. Double Handle Mode from Site SF-1, San Felix district; a Chiriqui Phase marker . . 114 54. Decorative Modes from the districts of San Felix and Remedios 115 55. Appendage Modes in sites of the Burica Peninsula, associated with Isla Palenque Maroon Slipped, Burica Phase 116 PLATES 1. Typical estuaries south of David, showing water level at low tide. 2. Typical mangrove vegetation on western shores of the Estero de Horconcitos. 3. Views of islands in the Gulf of Chiriqui. 4. Views of fields with vegetation cut and burned for maize planting in the San Felix district. 5. Rim sherds of Isla Palenque Maroon Slipped, Burica Phase. 6. Isla Palenque Maroon Slipped, Burica Phase. 7. Type sherds diagnostic of the San Lorenzo Phase. 8. Vessels of Banco Red Line, San Lorenzo Phase. 9. Cangrejal Red Line vessel and rims, San Lorenzo Phase. 10. Type sherds diagnostic of the San Lorenzo Phase. 11. Type sherds diagnostic of the San Lorenzo and Chiriqui Phases. 12. Type sherds of Linarte Zoned Red Line, San Lorenzo Phase. 13. Type sherds diagnostic of the Chiriqui Phase. 14. Trade pottery from other Panamanian regions, and sherds of Classical Chiriqui pottery. 15. Tripod Feet Modes of the San Lorenzo and Chiriqui Phases. 16. Tripod Feet Modes of local and trade origin, Chiriqui Phase. 17. Strapped Feet and Pedestal Base Modes and Ringstands, diagnostic of the San Lorenzo and Chiriqui Phases. 18. Handle Modes of the San Lorenzo and Chiriqui Phases. 19. Stone artifacts of the San Lorenzo and Chiriqui Phases. 20. Stone artifacts of the San Lorenzo and Chiriqui Phases. 18. 18. Cultural Chronology of the Gulf of Chiriqui, Panama Introduction GEOGRAPHY OF CHIRIQUI Within the small crescent-shaped Isthmus of Pan- ama, which is approximately 75,500 square kilometers, there are varied climatic and vegetational zones. Despite this variety, the deep jungles and tropical savanna, the lowlands and the highlands of Panama can all nonetheless be considered variations of an essentially tropical environment. The province of Chiriqui covers 8,960 square kilo- meters in the southwestern comer of Panama (fig. 1). In contrast to the lowland tropical forests of the neighbor- ing northern province of Bocas del Toro, it offers two, rather than three, of the prevalent Panamanian climatic zones: highlands with appreciable plateaus and valleys and broad coastal plains with savanna vegetation. The highlands are formed by the continu- ation of the Talamanca Range from Costa Rica into the western part of the isthmus, where it is called the Continental Divide. This mountainous backbone extends across Chiriqui and Veraguas, stopping at a point northeast of the town of Penonome in the province of Code. In the westernmost section it is capped by deposits from Miocene and Pleistocene volcanoes. The now inactive Volcan Baru in Chiriqui is 3,470 meters high, the highest spot on the isthmus; the rest of the Continental Divide averages 1,000 to 2,000 meters in height. Mountain flanks and plains are composed predominantly of Oligocene, Miocene, and Pleistocene marine sediments plus recent alluvium (Bishop, 1961). The mountains of the Continental Divide determine to a large extent the amount of rainfall and conse- quently the type of vegetation found on either side of the western section of Panama. Warm, rain-filled clouds formed over the Caribbean hit against the Continental Divide and drop their waters along the 284-757 O 68—2 Atlantic coast. There are places in the province of Bocas del Toro where the mean average rainfall a year is 3,500 millimeters. The Pacific coast of Chiriqui is drier than the Atlantic coast; this is true of all of the isthmus as well. Along Chiriqui, through the southern part of the provinces of Veraguas and Code, runs a coastal belt of semideciduous forests and savanna, varying between 5 and 40 kilometers in width. The rainfall here av- erages 1,500 to 2,500 millimeters; it is concentrated in the marked rainy season extending from May to December. The combination of sufficient rainfall and temperatures in the eighties turns the coastal plains of Chiriqui into Panama's richest agricultural lands. In addition, there are large rivers and innumerable streams to keep the land well watered even during the dry season. From descriptions by the chroniclers of the 16th century, one can be reasonably certain that the rich coastal savannas of Chiriqui existed in aboriginal times. The body of water south of the coastal plains is called the Gulf of Chiriqui (fig. 2). In the central part of this Gulf, some 16 kilometers south of the city of David, there is a complex embayrnent of estuaries stretching in an east-west direction for 40 to 50 kilometers. This embayment is formed by the drainage into the sea of several large rivers that flow down from the Cordillera Central (Bishop, 1961). Most important among these are, from east to west, the Rio Madronal, the Rio Chorcha, the Rios Gualaca and Esti, the Rio Chiriqui, the Rio Cochea, and the Rio David. Minor rivers flowing directly from the skirts of the Volcan Baru into the Gulf are the Rio Piedra, the Rio Chico, and the Rio Platanal. 1 SMITHSONIAN CONTRIBUTIONS TO ANTHROPOLOGY VOLUME 8 Punia Mariato FIGURE 1.—Map of Panama, showing area of research in the Gulf of Chiriqui in relation to selected sites in the Cocle-Azuero area. Re- search area is shown in hachure. Great estuaries are formed at the mouth of these rivers; through them flow sluggish, semibrackish waters that are constandy depositing fine silt and clay in the bordering swamps (pi. la). Today, these mosquito-ridden swamps are uninhabited by man, but the calm waters that border them are rich in fish, and the adjacent mud flats teem with mollusks and shellfish. They may have provided fishing and gather- ing grounds for people who lived elsewhere (pi. \b). The outer fringes of the estuaries are protected from the action of tides by numerous islands and sandbars. When the tide is low, great expanses of mud and sand are exposed as tidal flats and beaches. Only temporary encampments of fishermen dot these beaches today. Judging from the absence of archeological sites, there were no large prehispanic occupations in this area either. One of the largest estuaries of the Chiriqui Gulf is the Estero de Horconcitos, formed by the rivers Chorcha, Madronal, and Corrales. On the western shores of its mouth there are only swamps and man- groves (pi. 2), whereas high savanna lands and some hills, the largest of which is Cerro Pan de Azucar, render the eastern shores habitable by man. The site of El Cangrejal (SL-1) was found in these savannas (fig. 2). The Estero de Horconcitos drains into the protected Bahia de Muertos, flanked on the west by the islands of Sevilla and Mono, on the north by the low savannas of the mainland coast, on the east by the higher savanna lands near the town of Horconcitos, and on the south by the large island, Isla Brava. On the western corner of this bay is Villalba (more commonly known as Isla Muertos), where Sites IS-5, 6, and 7 were found (fig. 2). Two different channels can be taken from the Bahia to reach the open sea and oflFshore islands. Boca Brava runs between the easternmost tip of Sevilla Island and the westernmost tip of Isla Brava. Heavy surf and treacherous sandbars make this a dangerous route to navigate in any type of craft. The channel that curves around the eastern part of Isla Brava and passes south of Isla de Cedro is a far more navigable route. On the mainland facing it is the village of Boca Chica, connected with the larger town of Hor- concitos by a narrow dirt road. Once the open sea is reached by way of either of the above channels, the waters become choppy and the winds strong. Solitary islands and groups of islands dot the sea at varying distances from the mainland. Isla Palenque, where we found one of the most im- portant sites (IS-3), is just off the southernmost ex- tremity of Isla Brava. Farther to sea is a complex of WHOLE VOLUME INTRODUCTION • = Modern towns A = Sites excavated ■ - Site reported by other investigator jn^-r^ 2.00° f •CERRO = ^>../""" PUNTA Barriles .v..>-^"/< ... S ^""^-'^ ^. / C i. ^•*"^\IA,5 "■••"/, 2 0 00 n;\.., - 'J^o, Palenque i^San Panda Jose Islands 8°00' GULF OF CHIRIQU iS-lh Las Secas Cavada Jfe^S' \ |30' 115 82°i00 FIGURE 2.—Map of the central part of the Gulf of Chiriqui, showing location of excavations and geographical features. SMITHSONIAN CONTRIBUTIONS TO ANTHROPOLOGY VOLUME 8 islands called San Jose. About 10 kilometers south of Isla Brava is found the very large and heavily wooded island called Parida, surrounded by small islands and tiny rock promontories. The composition of these islands has been described in detail by the geologist of the expedition. Bishop (1961, pp. 7 and 8): Rocks composing the islands southeast of the estuary and the mainland east of the estuary are highly deformed, and they appear to be older than the strata of the estuary and the marine plain. They consist of shale, siltstone, sandstone and graywacke. Deformation has resulted in severe fracturing, faulting, and recementation of the rocks. Minor recrystallization has occurred in some instances. The strata have an average strike of N 70° W, with a dip to the southwest ranging from vertical to horizontal. On the islands of San Jose and Parida there are excellent exposures of pillow lava and columnar jointed lavas. Pillow lavas result when the lava is extruded into the ocean from submarine fissures. Between the "pillows" of lava there are deposits of chalcedonic quartz and green chart and other minerals. These minerals result from the chemical reaction between the sea water and the low temperature components of the lava. Columnar jointing results from the rapid cooling of lava when it occurs as a shallow sill. At one location in Parida island the pillow structure could be observed to grade into columnar jointed andesite. Much of the strata of these islands were sedimentary. The close association of the pillow and columnar jointed lava with the sedimentary rocks would suggest that the lava was intruded into the sedimentary strata during or shortly after their deposition. The vulcanism which gave rise to these lavas was probably associated with the severe deformation which produced the present physiography. If the deformation was Recent in age it would explain why this area of Panama possesses a submergent appearing physiographic character whereas the marine plain to the west and north definitely establishes a recent emergent history. The shore line of these islands consists of narrow boulders, gravel, or sand beaches and vertical sea cliffs. Far out to sea, about 40 kilometers from the island of Parida and 24 kilometers from the mainland, is the island complex of Las Secas, formed by two large and three small islands (fig. 2). Site IS-11 was found on Isla Cavada, one of the two large Secas Islands. To reach the site one must land on a narrow beach that cuts across the island and climb for about 100 meters on the steep sides of the nearest hill to the west, appropriately called Cerro Las Huacas. At the time of this survey, the U.S. Geodetic Service had cut a faint trail through the deep vegetation to its lookout tower on the top. Long distances separate Site IS-11 from the other sites in the Gulf. The mainland Site SL-1 is roughly 55 kilometers away, while Isla Palenque (Site IS-3) is approximately 37 kilometers to the northwest. To reach Isla Villalba (Sites IS 5-7), one must navigate around the bend of Isla Brava through the straits of Boca Chica for at least 53 kilometers. HISTORY OF ARCHEOLOGICAL WORK IN CHIRIQUI In the year 1838, a native of the town of Santiago de Alanje, together with a companion, found a large graveyard somewhere in the Cordillera de Chiriqui (Lothrop, 1919). Excited by this find he persuaded the famous General Morazan to organize a party and return to exploit the graveyard, but they were never able to relocate it. This is the first documented instance of what later became a lucrative sport among Pana- manians and foreigners alike, "la huaqueria." No more was heard of large Indian cemeteries until August of 1859, when the town of Bugavita had a gold rush of its own. In an effort to clear a garden plot, a man pulled a tree from the ground and found adhering to its roots spectacular gold ornaments, more of which lay in the ground below. The news of the treasure spread fast, and within a week the natives of Bugavita had dug more than 225 pounds (or 102 kilos) of golden objects (op. cit. p. 35). Only a few of these objects survived in museums; the bulk were melted down and sold as raw gold. The looting of graves continued through the years. Collectors became more eclectic in their tastes, incorporating into their collections not only gold images but also pottery vessels and stone artifacts. Finally thousands of Chiriqui pots found their way into museums. One of the largest collections of Chiriqui antiquities in existence was made by an indefatigable explorer, J. A. McNeil, in the 19th-century style. He personally supervised the opening of thousands of graves; the contents were sent to the U.S. National Museum, where they were classified later by William H. Holmes (1888). Holmes was the first person to publish an exhaustive classification of Chiriqui pottery. He divided the McNeil collection into two plain wares and nine painted wares. Today these wares are called "types": groups of vessesls sharing clusters of associated features. Holmes said of the criteria he used to separate wares (Holmes, 1888, p. 84): The characters upon which the classification is based are some- what heterogeneous and include material, color, shape, finish, ornamentation, method of manufacture, and evidences of use. With few exceptions, the criteria used by Holmes to separate wares are perfectly sound, though limited. WHOLE VOLUME INTRODUCTION They are purely descriptive criteria, devoid of chron- ological information. Holmes recognized the limi- tations of his classification. There was little he could do to extract chronological information from the McNeil collection because "Unfortunately our ob- servations in the field are not sufficiently accurate to enable us to utilize associations or methods of oc- currence in the graves as a means of classification" (op. cit. p. 84). There is a sort of pseudochronology in Holmes' thoughts about the evolution of art forms. He was an exponent of the notion that art underwent a necessary process of simplification, from representational forms that copy nature and are hence ideographic, to art forms that are geometric and nonimitative because they are produced by mechanical agencies. He did not, however, let this belief interfere with the ex- cellence of his detailed observations of Chiriqui antiquities. A second work in classification, this time a monu- mental volume, was written by George Grant Mac- Curdy in 1911. It is based on the large Chiriqui collections of Yale University and on sizable private collections, such as those of George G. Heye and Minor C. Keith of New York. MacCurdy (1911, p. 47) respected Holmes' classi- fication, but recommended two changes of nomen- clature: changing Holmes' name Terra Cotta or Bisquit Ware to Armadillo Ware, and Holmes' Tripod Ware to Fish Ware. More fundamental is MacCurdy's addition of three new wares: the Salmon- Colored and the Handled Ware (added as subdivisions of the Unpainted Group) and the Chocolate Incised Ware (added to the Decorated Group of wares). MacCurdy noticed that the Scarified and Chocolate Incised Wares differed from the rest of the wares, but did not attribute this difference to time or trade. Since his time, the Scarified Group has proved to be a Formative period pottery complex preceding by many hundreds of years the other Chiriqui Wares. The Chocolate Incised Ware is either a Costa Rican trade ware found in Chiriqui or a local imitation of a neigh- bor's product. The Polychrome Ware of both Holmes and MacCurdy is definitely trade from the Cocle- Azuero region. A third publication on Chiriqui ceramics, this time a short but very valuable article, was written by Cornelius Osgood in 1935. Osgood took a second look at the Yale collections and, while passing through Panama, inspected large private collections in Boquete, Chiriqui. Of the Holmes-MacCurdy combined total of 14 wares, Osgood retained only 4 wares that accounted for at least 95 percent of all the Chiriqui pottery in the collections he studied. These are Armadillo-Terra Cotta Ware, Fish-Tripod-Handled Ware, Lost-Color Ware, and Alligator Ware. The rest he explained as trade items or as problematic pieces. He traced the Polychrome Ware to Code; Red Line Ware, White Line Ware, and Chocolate Incised Ware to Costa Rica. The origins of the Serpent-Black Incised Ware, the Maroon Ware, and the Scarified Ware, he left as problematic. With admirable concise- ness, Osgood listed the salient characteristics of the four Chiriqui Wares that he retained and plotted their known distribution. The majority of them were found in Bugavita, Divala, and in an area 30 miles west-northwest of David. Wolfgang Haberland of the Hamburg Museum has split some of Osgood's groupings and included others, emerging again with 14 wares belonging to what he calls the "Classic Chiriqui" complex of A.D. 1200 to 1500 (Haberland, 1958). Following is a list of Haber- land's major revisions: a. He split the Lost Color Ware into four groups: Negative Ware, Black on Cream Negative, Black on Red Negative, Red on Red Negative. b. He reinstated the Red Line and Chocolate Incised Wares, plus the Black-Incised or Serpent Ware (which he renamed Punched Band Ware), as Classic Chiriqui pottery types. Osgood had discarded them as being of Costa Rican origin. c. He eliminated the Scarified Ware from the list because it is older than the Classical Chiriqui pottery types. d. He added four new wares to the list: Red on Orange, Redbrown Applique, Black Ware, and Un- painted Ware. The value of all the distinctions made by Haberland cannot be ascertained from the present study. Two of the phases I establish for the Gulf of Chiriqui precede in time his Classic Chiriqui complex. The most recent phase in the Gulf is Classic Chiriqui, but it is repre- sented by only two of the popular types associated with this period on the mainland: Bisquit (or Arma- dillo) and Fish-Handled-Tripod Ware (renamed Villalba Red Streaked by me, called Red Painted Ware by Haberland). The absence of other "Classic Chiriqui" types in our most recent phase may be due to the fact that Negative, Red on Orange, Choco- late Incised, and Red-Brown Applique are typical mostly of the Highlands of Chiriqui (Haberland, 1958, p. 345). Splitting, regrouping, and adding new wares to the Chiriqui complex reflects the obvious fact that little is known of the chronological or geographical sig- nificance of the Chiriqui ceramics. This is a situation which may be remedied in part by the present study. The Sites Four archeological sites were investigated in the Gulf of Chiriqui. The location of each site and its physical appearance are described, and excavation procedures at each pit are reported as fully as possible. The pottery type sequence yielded by these excava- tions is summarized following each site description, although pottery type descriptions constitute a sep- arate chapter. Cultural remains other than pottery are included in Appendix 1, tables 6 and 7. ISLA PALENQUE (IS-3) Location Palenque is a small island, 2.4 by 1.6 kilometers, located at 82° 15' west longitude and 8° 10' north latitude, immediately south of Isla Brava. These two islands are separated by a shelf less than a kilometer in length and only one fathom deep. Isla Palenque is surrounded by waters of relatively shallow depth (1)^ to 2% fathoms). It is not, however, in a protected location as are the islands in the Bahia de Muertos. Like the islands farther out to sea, it receives the full impact of tides and winds on its southern side. Although neighboring island complexes may be far away, they are easily accessible from Palenque. Isla Villalba is 20 kilometers away and LasSecas is 37 kilometers out to sea. To reach the mainland site of El Cangrejal (SL-1), one must navigate for 22 kilometers. In aboriginal times there must have been continuous contact between the inhabitants of all of these islands. Canoes, and perhaps even rudimentary sailboats, probably traveled back and forth much in the fashion that the scattered fishermen of today visit each other in crafts essentially unchanged since prehispanic days. Site IS-3 is located near the center of the island on a hill 50 meters high (fig. 3). The archeological site is recognizable by four closely spaced artificial mound'" built up in part by the deposition of one to two meters of cultural debris (pottery, bone, stone artifacts, and shell) on a naturally raised ridge one meter high. The ridge is about 100 meters long by 20-25 meters wide and runs northeast-southwest. Another ridge of the same height runs off the west end of the first ridge for about 50 meters. There are also visible middens on this second ridge. The area between the ridges is strewn with scattered potsherds. Judging from excavations at Pit No. 1, located 40 meters to the north of the ridged area, the depth of the deposit is a meager 20 cm. in the flat area surrounding the mounds. The mounds are covered with grass, while the sur- rounding area is heavily forested. We were informed that there is a freshwater supply from a sizable stream on the island, although we did not determine its distance from the site. Excavations Four pits were excavated at Site IS-3 (fig. 3) but only the largest. Pit No. 3, has been included in this report because it is the only one with considerable depth and cultural refuse. This pit, placed on the farthest mound at the northeast corner of the ridged 7 SMITHSONIAN CONTRIBUTIONS TO ANTHROPOLOGY VOLUME 8 Pit-3 Pit-4 / Old lake bed ? I I I 0 15 20 M FIGURE 3.—Site map of Isla Palenque (IS-3), showing location of excavations and topographical features. area, measured two meters north-south by one meter east-west. The arbitrary levels of 10 cm. in thickness were measured from the east wall of the pit, which was one level higher than the west wall. No natural stratigraphy was visible at the time of the excavations. The following description has been reconstructed in retrospect from field observations on the physical appearance of each level as it was being excavated, which permit subdivision of the deposit into sections. These sections form natural units in terms of the consistency, the color, and the contents of the soil. In addition, these units often correspond to sharp breaks in the amounts of cultural materials present in the levels, though not necessarily to the phase divisions as drawn in this report. Description follows the order of excavation, that is, from top to bottom. LEVELS 0-30 cm.—Soil was dark brown, soft, porous humus, and with a slight silt content. In the uppermost level there was a large quantity of frag- mented rocks, two or three fists in diameter, plus many roots. At 30 cm. the soil turned slightly more yellow, but kept its soft consistency. In the two top levels (0-20 cm.) the amount of cultural material was impressive: a gunny sack of potsherds per level, plus several small bags of stone artifacts. Some bone fragments, but no shell, were also present. A fractured vessel was found in Level 10-20 cm. It was laid, face down, 20 cm. from the east wall of the pit and 45 cm. from the south wall. In Level 20-30 cm. the amount of pottery decreased sharply to about one-half that of the upper levels. This sudden decrease may represent a temporary abandonment of the site by some of its inhabitants. WHOLE VOLUME THE SITES The three top levels correlate with the third and last phase of the pottery sequence at IS-3. The 30 cm. depth marks, therefore, a cultural change in the deposit. LEVELS 30-90 cm.—At 30 cm. the soil turned slightly lighter and more compact in texture due to a higher clay content. This trend continued to a depth of 90 cm. where the soil turned definitely hard and compact. Fractured stones were numerous, although less abundant than in the top levels. At 90 cm. a type of laminated shale rock, called locally "laja muerta," first appeared. The amount of cultural material found in the Level 30^0 cm. was almost twice that of Level 20- 30 cm., confirming the observation that the latter level marked a decline in intensity of occupation. Two nearly complete vessels and one tall pedestal base were found in Level 40-50 cm. One vessel (p. 26) was leaning against the east wall of the pit at a point 36 cm. from the south wall. Next to the west wall and 40 cm. from the south wall lay a fractured but nearly complete shallow tripod (p. 29). Adjacent to it was a tall pedestal base (p. 52). From the next level down we extracted several fractured shallow bowls, one inside the other, 76 cm. from the south wall, very close to the east wall. The presence of nearly complete vessels in three levels of this pit does not necessarily imply that the stratigraphy of Pit No. 3 was disturbed. It clearly was not, judging from the pottery type sequence (see dis- cussion in the following pages). Furthermore, a pre- liminary sorting of the sherds from Pit No. 2 (not in- cluded in this report) showed the same sequence of pottery types. The unit of levels between 30-90 cm. differs from the preceding unit by having soils that are slightly more compacted due to higher clay contents. Like- wise, it differs from those in the unit below by a relatively looser texture and by the absence of "laja muerta" (shale). This unit corresponds only in part to the cultural division which, on the basis of pottery type distribution, has been drawn at the 130 cm. mark. LEVELS 90-160 cm.—The soil of Level 90-100 cm. had a burned yellow color due to the large amount of "laja muerta" (shale). Its consistency was hard and rocky, becoming so compact at Level 110-120 cm. that a small, sharpened shovel had to be used to break the ground. Patches of nearly white soil could be observed throughout these levels, resulting from decomposition of limestone rock. Level 140-150 cm. contained very hard, almost pure igneous rock, weathered buff to tan in color. Preserved in it were exfoliated fractures typical of the exposed bedrock of these islands. Levels 150-170 cm. consisted of deeply weathered bedrock of either igneous or metamorphic origin and were practically devoid of cultural material. The amount of cultural material declined from 1,083 sherds in Level 90-100 cm. to 460 sherds in the next lower level. There was an increase again in Level 110-120 cm. to 960 sherds, followed by a second decline in the next level down. The next two levels contained between 100 and 200 potsherds; only a handful of sherds was found in Level 150-160. The unit of levels 90—160 cm. does not correspond in its entirety to the earliest phase at the site. Only the levels below 120-130 were included in this phase. The Pottery Type Sequence FIGURE 4; APPENDIX 1, TABLE 1 Site IS-3 (Isla Palenque) is a stratified site with a sequence of three phases distinguished from one another by the presence or absence of certain pottery types. The diagnostic types of the earliest phase, the Burica Phase, are Isla Palenque Maroon Slipped which occurs in frequencies of 30 to 40 percent and Plain Ware C which occurs in similar percentages. Levels 130-150 cm. correspond to the Burica Phase. The middle phase, San Lorenzo, is defined by large percentages of pottery types decorated with red lines, red bands, or a red slip: Arayo Polished Line, Banco Red Line, Caco Red Slipped, Cangrejal Red Line, Castrellon Red Slipped, Centeno Red Banded, Horconcitos Red Banded, Pan de Azucar Red Line, and Zapote Red Banded. Four of these types are restricted mainly to this phase; the others also occur in the next phase. The plain wares do not show any significant correlation. Levels 30—130 cm. corre- spond to the San Lorenzo Phase. The most recent of the phases, the Chiriqui Phase, is marked by the appearance of three new pottery types: Tarrago Bisquit Ware, Villalba Red Streaked, and Cavada Appliqu^ and Red Banded. Levels 0-30 cm. correspond to the Chiriqui Phase. 10 SMITHSONIAN CONTRIBUTIONS TO ANTHROPOLOGY VOLUME 8 EL CANGREJAL (SL-1) Location The Estero de Horconcitos bends east at 8° 20' north latitude to receive the waters of the Rios Madronal and Corrales and to carry them south to the Bahia de Muertos. The archeological site (SL-1) is on the mainland coast by the eastern edge of the Estero de Horconcitos, only some 100 meters from the edge of the water (fig. 2). Its exact location is 82° 13' west longitude and 8° 19' north latitude. The inhabitants of the site could have sailed directly south, hugging the eastern shores of the calm Bahia de Muertos to reach the eastern tip of Isla Brava, roughly 12 kilometers away. By sailing another 10 kilometers they could have skirted the tip of Isla Brava and reached Isla Palenque. Instead of sailing all the way, they could have walked south, reached the area where the town of Boca Chica stands today, and from there on sailed to Palenque Island. They could have also easily sailed southwest to reach the northeast tip of Isla Muertos, 10 kilometers away. The site (fig. 5) consists of four roughly ovoidal mounds in a half-moon arrangement. In part they are formed by natural undulations in the savanna; in part they have been built up by about one meter of cultural refuse. Two of them were sterile of cultural material, but potsherds and other artifacts were found on the surface of the flat, unmounded areas, so that the total site area is more than 200 meters square. The site itself is between a grove of "corozo" palms to the north and the dropoff line to the Estero de Horconcitos on the west. The mound tested is only 16 meters from the edge of the bank that drops off gradually to meet the calm waters of the estero. A brook called Quebrada del Caco runs about 130 meters to the east of the site. About 90 meters to the northwest is a waterhole called locally "El Zapote." Both of these could have kept the people of SL-1 well supplied with fresh water. Excavations Two small pits were excavated in Mound A, the nearest mound to the bank. The surfaces of Mounds B and C, were found to be sterile. Pit No. 2, on the flat area to the northeast of Mound A, yielded a small number of potsherds. Pits 1 and 3, included in this report, were dug in artificial levels of 10 cm. No natural stratigraphy was observable. Test Pit 1, measuring 1 by 1 meter, was placed on the northern half of the mound. LEVELS 0-60 cm.—The soil was brown in color and slighdy compact, tending to break into very small clods. Well-rounded and fractured small stones occurred throughout, but were especially abundant in the top three levels. The amount of cultural material yielded by these levels was uneven. Only 215 potsherds were found in the first level, while the second level (10-20 cm.) had the largest number of potsherds in the entire pit—1,093 sherds. The other four levels in this unit produced between 200 and 500 potsherds each. Fragments of mammal bones and a small number of shells were found throughout. In Level 30^40 cm. there was a concentration of bones, mainly of fish, in a circular area 20 cm. in diameter in approximately the center of the pit. These bones were associated with pieces of carbon and fired earth, suggesting that this had been a small cooking hearth. In the southwest quarter of the pit, at Level 50-60 cm., a concentration of small univalve shells of an edible variety was encountered. There were a few rocks at the margins of this shell concentration, but the association between rocks and shells may be accidental. The upper 60 cm. of the deposit corresponds to the last, or more recent, of the two subphases postulated by this study for Site SL-1. LEVELS 60-90 cm.—At 60 cm. a sharp change occurred from the brown, slightiy compact soil of the upper levels, to a yellowish, hard, and compact soil that tended to break into very large clots. Com- pactness increased until, in the bottom level, it was a red-brown almost impenetrable clay. The number of potsherds declined from 342 in the preceding level to 110-158 in Levels 60-80 cm. The bottom level had only 22 sherds. The rare shells found in these levels were often so disintegrated that they appeared as mere white lenses on the soil. This unit of levels corresponds to the earliest of the two subphases established at SL-1. Test Pit 3 was placed 3 meters to the south of Pit No. 1 in an effort to increase the sample of the later subphase. It was begun with the dimensions of two meters east-west by one meter north-south, but heavy downpour forced us to reduce the area to one-half at the third level down. Reducing the dimensions facilitated covering the excavated area with a small tarpaulin and also accelerated the work. The sample size in both pits, considered together, is large enough to establish a sequence of deposition for D Q 0 D □ □ 3 a D H UJ O J I I 0 5 10 15 20 25% Scale B 0 - iO 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80- 90 90-100 100-110 liO-120 I20M30 130-140 140-150 D D D UNCLASSIFIED PLAIN TYPES- I > I ■ I I 0 10 20 30 40 50% Scale A FIGURE 4.—Pottery type sequence of Pit No. 3, Site Isla Palenque (IS-3). D □ I 1 I—\ Q 0 a a Scale A <- D D D EZD ill CL CL LU CO o CO 5 d k i Q ] dn m 0 Q D D □ D D D □ L: D D D nn D Q LiJ Q LU _l Q LU Q Q LU iD 2 < 0 LU < GO X 0 m CC Q- LU CO z y^ h- Q 0 _I UJ Q 04J 0 UJ > 0 1- UJ ocf) a. a: < Q. z cr < 1 < cr LU 1 ——^ Q ^ n 10 R < 0 ^ 0 0 a 1 ti D 0 I 0 Q a 0 I □ n n a I D D D D D D Q Q D D D n D CAR Q Ul CL (O Q UJ 3 UJ 2 CL 0 0 LU NJ Z 0 }— Z -J Z < — _i _J < < "i ^ -I _j CO 0 CD -0 -" LU Ul •z. UJ Q Q (r Q 0 Q cr Q LU H- Ul 0 LU C3 LU Z (T cn cr cr cr z cr < < 0 < a_ 0 X 0 D 0 0 Ul z 0 N UJ z LU _l 1- cr Q < UJ z QC _J en LJ O o O UJ QC < CD LU < 3 O CO O CD < cr cr CD < Q UJ cr Q LU V < CD 1 < U) < \x. -J \- CO UJ 3 0 Q -1 UJ CL Q Q- Z < < CD < Q Q < Ul > cr < 0 (C WHOLE VOLUME THE SITES 11 ^ _, — — ~~-^ -^ /^ / y • / y y / Pit-2 '^ Mound A ^- 1. ,x o' Previous ^ excavation Ujl /■■ Pit-I • CO' UJI / Mound \ NQ 2 B •"''"■ \. Pit-3/ K.' / Previous ^ .. ^ ^^ excavation