AN ARCHAEOLOGICAL SURVEY OF THE P’UU BASIN. WEST CENTRAL TAIWAN, REPUBLIC OF CHINA Dissertation for the Degree of Ph.’ D. MICHIGAN STATE UNIVERSITY RICHARD BROWN STAMPS 1975 This is to certify that the thesis entitled AN ARCHAEOLOGICAL SURVEY OF THE P'ULI BASIN, WEST CENTRAL TAIWAN, REPUBLIC OF CHINA presented by RICHARD BROWN STAMPS has been accepted towards fulfillment of the requirements for Ph.D. degree inAnthropologY .\ r \V I . \ X» 1" .\ \ 5" L3 ‘ \ (-‘ 0" \z r.. I, \_.y x \ 3k .\) -t‘ "\\ \‘v\ v, \I \’ ‘ 5‘) '4 \ ._‘ \g \ :3 luwx (':N J Major professor Date 20 May 1976 0-7 639 test th Taiwan culture ABSTRACT AN ARCHAEOLOGICAL SURVEY OF THE P'ULI BASIN, WEST CENTRAL TAIWAN, REPUBLIC OF CHINA By Richard Brown Stamps The purpose of the research herein reported was to test the hypotheses of Dr. K. C. Chang, that in central Taiwan (1) there was a shift to slash and burn, root agri- culture by Cord Marked pottery peoples at c. 9000 B.C., and (2) that another shift to grain agriculture by Lungshanoid peOples occurred at c. 2500 B.C. As a result of the research the first hypothesis is strongly questioned, while the second hypothesis was supported. AN ARCHAEOLOGICAL SURVEY OF THE P'ULI BASIN, WEST CENTRAL TAIWAN, REPUBLIC OF CHINA By Richard Brown Stamps A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Anthropology 1975 Dr. Josef Maxwell : knowledg Larry R3 encourag expected for his Proiect, and Out their a1 their 1N alSo to COOnQra. the pol. ACKNOWLEDGMENTS I wish to thank my dissertation committee, Dr. Joseph Chartkoff for theoretical guidance, Moreau Maxwell for technical guidance, Bernard Gallin for his knowledge of China, and the chairman of the committee, Larry Robbins, for his overall support. guidance and encouragement above and beyond what was required, eXpected or even hOped for. I express Special appreciation to Dr. K. C. Chang for his invitation to participate in the Choshui and Tatu Project, as well as his guidance and suggestions both in and out of the field. Thanks to Misses w. H. Chang and w. S. Lin for their aid in the preparation of the pollen slides and their introducing me to the procedures involved. Thanks also to Dr. T. C. Huang and Mr. T. C. Chung for their c00peration and significant contributions in the study of the pollen record. Thanks also to Professor Y. C. Hsu and his staff at the Radiocarbon Laboratory. Department of Physics, N.T.U., for processing the radiocarbon samples. I wish to thank all of the staff at the Institute of Ethnology, Academia Sinica, e8pecially Y. Y. Li and ii S H. W8- of the 0 23111 Co T‘ Taiwan 8 Huang of V T.'. f ment of and rela animal 1 Mr 0 than I was in artists apprecia oloéists C‘ H. T: compani fer the S. H. Wang for their help and guidance in making my portion of the overall project work smoothly. Thanks to P. H. Liu and C. W. Liu for my introduction to P'uli and central Taiwan archaeology. I also wish to thank Professor W. S. Sung and S. C. Huang of the Department of Archaeology and AnthrOpology, N.T.U. for sharing their experience and knowledge with me. I wish to acknowledge Dr. T. T. Shih of the Depart- ment of Geography, Normal University for his help with maps and related problems, but most for his encouraging smile. Thanks to J. S. Liang for his information about animal life in central Taiwan. I give Special acknowledgment to Mr. P. C. Lin and Mr. Huang for their assistance as guides and workers while I was in P'uli. The work of L. C. Lin and Y. C. Tung as artists and C. T. Hu as a lab assistant was also greatly appreciated. Last but not least, thanks to my fellow archae- ologists on the projects S. C. Huang, S. C. Lo, P. K. Sun, C. H. Tsang and C. K. Ho for their ideas, guidance and companionship. Ideas came from each, but I accept reSponsibility for the contents of this work. iii Chapter 1. TT $.50 III. IV. VI. QPPendifl Append I Append i4 5.10110); TABLE OF CONTENTS LIST OF FIGURES . . . . . . . . . . . . . Chapter I. INTRODUCTION . . . . . . . . . . Statement of the purpose Review of literature II. HYPOTHESES TO BE TESTED . . . . Hypothesis 1 Hypothesis 2 Hypothesis 3 III. METHODS FOR TESTING HYPOTHESES . Statement of approach Test implications IV. DATA COLLECTION METHODS . . . . Sampling zones Survey procedures Test excavations V. DATA ANALYSIS . . . . . . . . . Results of the stratified survey Summary of survey results Chronology of P'uli Basin archaeological remains Pleistocene and Holocene vegetational sequence VI. CONCLUSIONS . . . . . . . . . . Cultural history of the P'uli Review of the hypotheses Significance of this study Appendix A: Ceramic Classification . . . Appendix B: Stone Artifacts . . . . . . Appendix CI Inventory of Sites . . . . . Bibliography.............. iv 27 39 47 100 118 1&8 16h 179 ‘IJ on L "3 (D 153 ll. 13. 14. 15 16. 17 18. 19. 20. 21. 22. 230 24. LIST OF FIGURES Page Landforms in the P 'uli Basin . . . . . . #0 Six Sampling Zones . . . . . . . . . . #2 Sites Located During the 1972- 73 P'uli Basin Survey . . . . . . . . . . Rear Pocket Site Po Lo 38 e o e e e I o e e o o 0 “‘9 Site P. L. 38, profile drawing of west wall in test trench (between 8 and 9 meters south) 0 O O O O I O O O O O O O O O O O 51 Site P.L. 39 e o e e e e o o o e e o e o o e 53 Site P.L. 37 e e e o o e o e e e e e e o o o 55 Site P.L. 250 e e e e o e o e e e e e 59 Site P.L. 25, profile drawing of east wall in 12881; pit A O O O O O O O I O O O O O O 61 Site POL019 e e o o e e e o o 614’ Site P.L. l9, profile drawing of south wall in teSt pit B O O O O O O O O O O O O I O 67 Site P. L. 29 o o o e e e e e e o 69 Site P. L. 29, profile drawing of west wall in teSt pit B O O O O O O O O I O O I O O 71 Site P.L. 21, Ta-ma-lin . . . . . . . . . 75 Site P.L. 21, profile drawing of west wall in teSt pit C e o e e e o e e o e e o o e 77 81133 P. L. 30 o o o e o o o e o o 81 Site P. L. 30, profile drawing of intrusive pit between burial cist A and F . . . . . 81 Sites Located in Each of the Sampling Zones. 87 Radiocarbon Dates from the P'uli Basin . . . 88 Time Periods in Central Taiwan and the P 'uli Basin . . . . . 89 Pollen Diagram of 19 Samples from 3 Sites of the P 'uli Basin . . . . . 95 Profiles of the Tamalin Light Brownish Gray / Coarse Pottery . . . 137 Profiles of the Tamalin Light Brownish Gray / Medium; Tamalin Light Brownish Gray / Fine; Tamalin Light Orange / Coarse and Tamalin Light Orange / Fine Pottery . . . . . 138 Profiles of the Tamalin Dark Gray /'Medium; Tamalin Brownish Gray / Coarse: Tamalin Brownish Gray /’Medium and Tamalin Brownish Gray / Fine Pottery . . . . . . 139 250 26. 27. 28. 29. 3o. 31. 32. 33 34. 36. 37. 38. 39. 40. #1, Profiles of the Tamalin Light Orange / Medium: Hsu Light Brownish Gray / Coarse; HsiaOpushe Brownish Orange / Fine: Shuiwaku Yellow Orange / Coarse and Shuiwaku Light Brownish Gray / Fine Pottery . . . . Profiles of the Shuiwaku Light Brownish Gray / Coarse Pottery . . . . . . . . . Surface Decorations on Pottery, arranged by ceramic classes with provenance data for each example . . . . . . . . . . . . . Surface Decorations on Pottery, arranged by ceramic classes . . . . . . . . . . . . Surface Decorations on Pottery, arranged by ceramic classes . . . . . . . . . . Example of a Pointed Shaped Hoe . . . . Examples of Oval Shaped Hoes . . . . . . Example of a Rectangular Sha ed Hoe . . Waisted or Necked Shape Hoe A), Trapezoi Shaped Hoe (B) o e e o e 0 Examples of Spatula Shaped Hoes . . . Examples of Grooved Net Sinkers . . . Examples of Notched Net Sinkers . . . Examples of Adzes from the P 'uli Survey Examples of Triangular, .Tanged and Serrat d d 6 Arrow POintS o o e o o e o e o 0 Examples of Willow Leaf and Triangular Shaped Arrow Points . . . . . . . . . . Examples of Rectangular Knives . . . . . . . Examples of Stone Artifacts: Knives - A- rectangular; B, C-crescent shaped: D-semi- lunar; Unknown - E, F-boot-shaped objects vi 140 141 142 144 146 152 153 154 155 156 157 158 159 160 161 162 163 reporte: activiti attempt Sition t central Shift in dcing th relation More Spe: Changes . 13310118an If SOT W of Deep}. lion cau Way of 1 Or were n“. b”liver I?“ v 3Qlwan IQ CHAPTER I INTRODUCTION Statement of Purpose The purpose of this paper and the research herein reported is to investigate the origins of agricultural activities in the uplands of west central Taiwan. An attempt is made to shed light in particular on the tran- sition to an agricultural way of life by the people in central Taiwan, and, in general, by extension to a similar shift in other parts of Taiwan and all of East Asia. In doing this I have tried to shed light on the changing relationship between humans and their physical environment. More Specifically, an attempt is made to study the reported changes in Taiwan's vegetation over the past fifteen thousand years. Have changes in the vegetation occurred? If so, when? Were these changes related to the movements of people into the area? Were these changes in the vegeta- tion caused by peOples shifting from a hunting and gathering way of life to a settled village agricultural way of life? Or were there other causes? It has been reported (Tsukada 1966, 1967; Chang and Struiver 1966; Chang 1969, 1970 A) that several shifts in Taiwan's vegetational cover have occurred in the past activit gically the Lung archaeo] general out in C ROWevEr, used in 1 archaeolC testing a of archa. basic gc; fifteen thousand years and that these changes have been induced by human activity. Dr. K. C. Chang of Yale University and the Academia Sinica in Nan Kang, Taiwan, noting the reported changes (Tsukada 1966) in Taiwan's vegetation at approximately 9000 and 2500 B.C., developed a hypothesis for further testing that explained these changes. He described them as first a shift from hunting and gathering activities to fruit and root agriculture (9000 B.C.) and later a shift to grain agriculture (c. 2500 B.C.). These hypothesized shifts in subsistence activities were attributed to peOples known archaeolo- gically on Taiwan as the Corded Ware (pottery) peOples and the Lungshanoid peOples. It was the purpose of the archaeological research herein reported to test this general hypothesis. Specific hypotheses tested in this paper are Spelled out in Chapter II. It Should be pointed out at this time, however, that the eXplicate hypotheses testing approach used in this paper is novel among studies of Taiwan archaeology. It was felt that in using this hypothesis testing approach we could maximize our return of data and information while answering questions with the minimum use of archaeological, financial and manpower resources. It was felt that this approach would not only meet one of the basic goals of archaeology by develOping a cultural history First, look at questic Stateme third c, hypothes needed 1 fourth c and the final Cy} includes C‘JSSion DPOPOSed modified this stu, The appel CEramic , and the A d of the area but that it would also go beyond this to test some eXplicit hypotheses important in develOping law-like principles for explaining human behavior. It is hoped that well-tested hypotheses will aid in the search for the principles governing the trajectory of human develOpment. The general outline of the paper is as follows: First, a review of previous research in the literature to look at the work that has been done and the kinds of questions that are being studied. The second chapter is a statement of certain Specific hypotheses to be tested. The third chapter deals with methods for testing the proposed hypotheses, the test implications for each and the data needed to accept or reject the test implications. The fourth chapter deals with the data collection methods used, and the fifth chapter with the analysis of that data. The final chapter deals with the conclusions of the study. It includes a cultural history of the P'uli Basin with a dis- cussion of subsistence activities and a discussion of the pr0posed hypotheses where they are supported, rejected or modified in light of the recovered data. The importance of this study for Asian prehistory in general is then discussed. The appendix includes additional information regarding the ceramic classification (A), the lithic classification (B), and the sites located (C). archae stage the £0 Because groundv it will directs from th Nationa Academia Work. at the W Review of the Literature In an attempt to evaluate the state of the art of archaeology dealing with Taiwan in 1972 and to set the stage for the research problems discussed in this paper, the following review of the literature is presented. Because the purpose of this review is merely to lay the groundwork for this report and is not an end in itself, it will be brief. Those interested in more detail are directed to Sung Wen-hsun's l953/5A index: the Bulletins from the Department of Archaeology and AnthrOpology, National Taiwan University, and the Institute of Ethnology, Academia Sinica: and the introduction to K. C. Chang's 1969 work. A In the present review, an attempt is made to look at the work that has been done and the research goals that motivated those researchers. First we encounter the early period of discovery and testing. This is followed by a period of early synthesis. Next we will look at the post- wOrld War II period. Then the work and questions from the 1960's and 1970's will be examined. The earliest reported discovery of prehistoric remains in Taiwan was made by a Japanese school teacher in 1896. Stone implements were found at Chih-Shan-yen on the northern outskirts of Taipei. This led to the discovery of the Yfian-shan Shellmound in 1897. At this early date basic island anthro Taiwan but th basic questions about the cultural affinities of the islands‘ early inhabitants were asked. The Japanese anthropologist Torii Ryuzo (1897:118) wrote, "The Taiwanese Stone Age remains are certainly prehistoric. but the problem is who left them? Were these peOple Malays, Negritos, or Papuans? This should be further studied." In 1930 Professor Utsurikawa NenozB and other staff members of the Taihoku Imperial University investi- gated the site of K'en-ting in southern Taiwan. Stone burial cists were found at K'en-ting. This search was one of the first major excavations on the island. For the next thirteen years several sites along the west coast were located and surface collections taken (Chang l969:7: Sung 1953-1954). Similarities were observed between pottery, both black and painted, found on Taiwan and pottery from the mainland (Kokubu 1943: Kanaseki 1943). In 19A3 Kano Tadao published a general synthesis of the sites and materials that had been collected on Taiwan to that time. He divided the prehistoric peOples of the island into seven cultures, according to Specific traits (an index fossil approach), and then suggested possible areas from which they had come. He described seven strata or cultures from early to late: (1) Cord-marked pottery, probably introduced from the Asian mainland: (2) check impressed pottery, (& tc ea A; [ V'Qf‘ I 7.15 l-j migrat they "r A. .“e et and art: QOQCIQSI introduced from Central China: (3) black pottery, introduced from the eastern coastal regions of China: (A) the stepped adz, introduced possibly from Fukien: (5) Proto-Dongsonian stratum, intro- duced from Indochina before Chinese influence: (6) Megalithic Cultural stratum, probably related to Megalithic culture of French Indochina eSpecially of Cambodia: and (7) Philippine Iron Culture stratum, confined to the east coast Afhang 19 69: His idea was that the prehistoric inhabitants of Taiwan migrated in waves from the Asia mainland to Taiwan where they maintained their cultural identity, some of them into the ethnographic present (Kano 1955: Miyamoto 1956). The important question of the day seemed to be from where did Taiwan's prehistoric peOpleS come? How were they related to the modern aborigine groups? After the close of World War II, archaeological information was gathered from several sources. A Depart- ment of Archaeology and AnthrOpology was established at National Taiwan University (NTU), and several important surveys and small excavations were carried out there. Reports from earlier Japanese excavations began to appear (Kanaseki and Kokubu 1953). By 195A K. C. Chang, a student of archaeology at NTU, was able to use excavated materials to deve10p the first major time-Space synthesis of the prehistoric cultures in Taiwan. Chang grouped the pre- historic peoples into seven cultures, Showing type sites and artifact inventories attributed to each. The general conclusion was that ”most of the prehistoric inhabitants time Cult Vari of Formosa were immigrants from the mainland of China. The formation of the prehistoric culture layers of the western coast of Formosa, therefore, cannot be understood unless the cultural stratigraphy of the mainland has been reconstructed." (Chang 1954:162). The major objective was to develop temporal sequences of spacially different cultural groups and then to Show the origins of these cultural groups and how they related to each other. The growing importance of Taiwan in the prehistory of East Asia in general and the islands of the western Pacific more specifically was seen by the publishing of the Special Taiwan Section of the 1963 Asian Perspectives. Tapics covered included geology and ecology of Taiwan prehistory: recent excavations of key sites: external relations to the east, the north and the south and the position of Malayo-polynesian languages of Taiwan. In 1963 K. C. Chang reviewed the contributions of individuals and institutions to the study of Taiwan's archaeology, and he noted that work was done "under the auspices of colleges, schools, museums, or the provincial and local Commissions for Historic Research" (Chang 1963a: 195). He suggested that research to that date had been directed toward the following three aims: (a) to group archaeological assemblages under a time-Space framework: (b) to reconstruct the culture and society of prehistoric peeples in various periods and regions of the island: and (c) and area rel; sites, i of these been tee Tap’enk‘ and the these st (c) to compare the island's prehistoric cultures and cultural elements with those in the adjacent areas of the Far East to determine their historic relationships [Chang l963a:l957. As recently as 1963 progress toward answering the first two questions was Slow. Of the several hundred known sites, few had been extensively excavated, and only a few of these were stratified sites. Important sites that had been tested included Fengpitou (Tsuboi 1956: Chang 1969), Tap'enk'eng (Liu 196A), Shuiyuanti (Sung and Chang 195A) and the Yflanshan shell mound (Chang 1956:377: 1969). From these stratified sites only general sequences could be detected: i.e.. cord-marked pottery from the lower levels followed by the red pottery and then by the black pottery. No radiocarbon dates from Taiwan were available, and all sequences were relative. Chronological control was weak at its best. Reconstructions of the culture and society of prehistoric peoples also suffered from the lack of large- scale excavations that might show things like settlement patterns, activity use areas, social stratification and subsiStence patterns. In 1963 no pollen studies had been made, and with the exception of a study done by the geologist Lin Chao-chi little interdisciplinary work had been carried out. Of the above-mentioned research aims, the greatest success was in showing the relationship between Taiwan and the mainland. Both Kano (1952:181) and Chang hi ceramic influem contrar; gaisticz $‘nré- ' Luau Wl‘ from the Shown, Chang had provided convincing evidence in the form of ceramic and lithic comparisons to show that Taiwan was influenced by the mainland. At the time, this was contrary to popular theory which related Taiwan lin- guistically with insular Southeast Asia. Chang states that with the excavation reports and information coming from the mainland the relationship with Taiwan was clearly shown. It is now solidly established insofar as the western coast of the island is concerned, that (a) its pre- historic cultural phases have as many horizons as there are on the opposite coasts of mainland China, and (b) each ceramic horizon of Taiwan is a local facies of a corresponding and identical horizon of the South China coast [Chang 196Aa:l9z7. Ling Shun-sheng, a scholar at the Institute of Ethnology, Academic Sinica, independently came to a similar conclusion. After an extensive study of several cultural traits, including rafts (Ling 1956), kava-drinking (Ling 1958) and ancestral temples (Ling 1959), he concluded that South China was the ancestral home of the Malayo-Polynesians. He determined that these peeples were forced out or assimi- lated by the southward and eastward advance of Sino-Tibetan peOples. Remnants of these early peoples are the Naga in the west and the Taiwan aborigines in the east (Ling 1959: 182-18A), Another important synthesis of the archaeological, linguistic and cultural data of the Taiwan aborigines groa 330‘ Y 1 ~ 010.51 S In t: 9“ g. fa s; at" OP is te I’TT '13 l>-J o o m U 0 (‘7 IO appeared in the Spring, 1966 Bulletin 9: the Institute of Ethnology. Raleigh Ferrell, after a study of aboriginal language groups of Taiwan, divided them into three major groupings: Ataylic, Paiwanic and Tsouic. These major groupings were then compared with the three major archae- ological cultures known at that time. His conclusions were: Archaeological data point clearly to a direct South China derivation for the overwhelming majority of the Formosan peoples and cultural traits. Close examination of cultural and lin- guistic data which Show the present-day tribes to fall into three distinct groupings, also gives surprisingly eXplicit clues as to the possible affinity of each of the major groupings with one of the three major prehistoric traditions on the island. This in turn permits us to assign a tentative area of origin on the South China main- land for the Speakers of the various present-day Formosan languages. The Atayalic/Cord-Marked Pottery Horizon shows clear affinities with the South and Southwest China region, and the Tsouic/ North Formosan Proto-Lungshanoid (Yflanshan) Culture has unmistakable northern elements, and may represent the more northerly of the Austronesian mainland peeples who earlier occupied the entire eastern coastal region of China and probably extended as far northward as modern Japan and Korea. The Paiwanic/South Taiwan Lungshanoid- Geometric Horizons are probably from an area between the Ataylic and Tsouic areas on the mainland. Their culture was basically that of pre-Han Southeast China, and their Spread to Formosa was part of the large-scale movements of Lungshanoid agricultural peOples from the Northwest China nuclear area into mainland Southeast Asia and the Pacific islands during the first and second millennia B.C. [Ferrell 1966:1237. Some changes have been required in the above scheme due to the addition of earlier dates, and Ferrell (1969:21-22, 73) has taken a more cautious stance on identifying specific moderl h owe ve 11 modern groups with prehistoric cultures. This 1966 work, however, was representative of the research problems being studied at the time, principally the question where had the people come from? In W. H. Sung's 1967 report, Q-nggepi, A_£§g- historic Site at the Southern Tip g: Formosa, we see another example of detailed archaeology. This study, though, is still basically developing typologies and establishing time-Space relationships. The basic questions are still: Who were these pe0ple? From where did they come, and when? Who are their descendants? How are they related to other prehistoric groups? From Sung's report of the O-Laun-pi region, it appears that two groups of people have occupied the area. The first group, called the Stone Cist peOple, shows certain material cultural Similarities with the earlier phase of the Lungshanoid culture at Fengpitou. Similar artifacts include the long-necked bottle, the high pedestal with cutouts, the hache pediforme stone knives, the thin plano-convex stone hoe, the c-type tridacna bracelet and the discoidal tridacna core. Sung also points out certain similarities with the coastal region of eastern Taiwan (Sung et a1. l967:#6). Possibly the most interesting result of this research is the isolation of two different types of stone burial cists: the rectangular 12 (presumably extended burial) cists that seem to fit into the Lungshanoid era, and the square (flexed burial) cists that are very Similar to those used in the ethnographic present by the Parilarilao tribe in the village of Kuraluts in the foothills northeast of O-Laun-pi. The case for cultural continuity is strong. The finds from this area are interesting in light of the stone cists found by the author in the P'uli Basin (see pp. 34, 36, A9-50) . Prior to the important l96h-l965 Yale-National Taiwan University expedition to Taiwan, K. C. Chang noted the following specific problems (Chang 1969:9-12): First there was a lack of extensively excavated sites. Few relatively complete cultural assemblages were known for any of the prehistoric cultures, and those that were delineated had been defined on the basis of single traits. There was a need to understand better the relationship between the cultural traits and the natural environment. Second, Chang listed the problem of the very loose island- wide time-space framework. At that point he noted, "East Asia archaeology in general remains a pre-carbon-lh stage insofar as chronological assessment of data is concerned" (Chang 1969:10). Third, Chang looked at the island of Taiwan as if it were the crossroads between the Asian mainland and the insular area of the western Pacific. He 13 suggested that it could be an important area for studies of the processes of cultural growth. He then listed seven Specific questions regarding the relationship between prehistoric Taiwan cultures and possibly related cultures in Asia and the western Pacific. An attempt was made in the 1964-1965 expedition to answer some of these questions and firm up the chronological picture of the area. This expedition's results appeared in 1969. They have given us the most complete synthesis of Taiwan's prehistory to date. Detailed results from the excavation of Fengpitou in the southwest and Tapenkeng in the north (the Taipei basin) indicate that a common Corded-Ware Culture came from the earliest levels at both Sites (thus supporting the earlier work of Kano 219527). This level was followed by the Lungshanoid culture in the southwest and the Yfianshan Culture in the north. The subsistence base for the Corded-Ware cultural horizon was not clearly understood, although a small number of stone implements, apparently used for cultivating purposes, and woodworking tools combine with the palynological evidence at Jihyfiehtan, showing the burning of primary forests. This also suggests an early form of swidden agriculture, possibly root and fruit crops. By 2500 B.C. when the Lungshanoid and Yaanshan cultures appear on the scene, intensive agriculture base (i devote: of the a very Treistm influen Tonkin, related areas. Phases c on the x the Cor corded-h the wide POSsible cord-ma: China) 8 19703175 an atten reS‘Jl‘ts + T a be: 1A (probably millet and rice) combine with some hunting, fishing and mollusks-gathering to form the subsistence base (Chang 1969:191-19A). An important section of Chang's 1969 work is devoted to comparative studies showing external affinities of the various cultural manifestations on Taiwan. He makes a very convincing case (although not totally accepted: see Treistman 1972) for the Yfianshan Culture's receiving influence from the South China Sea Coast and the Gulf of Tonkin, and for the Lungshanoid Culture's being directly related to the cultures of the southeastern China coastal areas. He goes so far as to delineate local Lungshanoid phases on Taiwan and trace them to their "close relatives on the mainland" (Chang 1969:238). Possible origins for the Corded-Ware Culture are not as clearly known. The Corded-Ware Culture is apparently a local manifestation of the widespread cord-marked pottery horizon of east Asia. Possible correlations exist with the Jomon in Japan, cord-marked pottery in Southeast Asia (including southern China) and the Kalanay complex in the Philippines (Chang 1970:175-176). Another important section of his report deals with an attempt to bring together in an interdisciplinary manner results of geological and palynological information germane to a better understanding of Taiwan's prehistory. C. C. Lin'5 and 2 cent: cool: .1 (D (I) (D (I) 15 Lin's 1964 work dealing with the changes in the shorelines and Tsukada's 1966, 1967 study of pollen samples taken in central Taiwan are combined in a comparative study of the ecology and archaeology of Taiwan. In conjunction with K. C. Chang's 1964-1965 research, Matsuo Tsukada collected six lacustrine cores in an attempt to study the past climatic changes of the island. The most detailed results came from the Jihyflehtan (Sun Moon Lake) core. The beginning of the Recent period, dated to about 11,000 BP. began with the ”destruction of primeval forests, probably caused by early human activities" (Tsukada 1966:5A6). Although no direct connection could be made, it has been suggested that these "early human activities" were performed by the peOples of the so-called Cord-Marked pottery cultures (Chang 1969:191-194). Another important change in the floral situation of Taiwan appears to have occurred around 4500 BP. At this time there was a Sharp increase of pollen from chenOpodiaceae and large-sized grass which Tsukada (1966:5A6: 1967:50) considered to be cereal pollen. The finding of grass pollen, associated with an increase of subtropical secondary forest, suggested to Tsukada evidence for agriculture. This sudden change in the vegetation seems to correlate well with the influx of Lungshanoid and Yfianshan cultures from the Asia mainland (Chang and Struiver 1956). AS other rese Ir C. C. Lin industry 1 Taiwan whi (Raiiocarb communicat farther. ChOppers a: tance of t} date at whj important If Um west cc Ral of the more Ill'ith their inclusiVe r that "Scien lilfancyn (F 750 cOnneCt C‘Qltures as One data to diTr I the Atam: 16 1966). As this work was being done on the west coast, other research was being carried out in eastern Taiwan. In January and February 1969, Sung, Wen-hsun and C. C. Lin located a pre-ceramic, chipped flake tool industry in a series of sea caves on the east coast of Taiwan which they named the Changpinian. A 15,000 BP date (Radiocarbon 1973) was recovered, although Sung (personal communication) would like to push this date back even farther. Finds included flake implements, hand axes, choppers and several bone and antler objects. The impor- tance of this find is that it pushes back the earliest date at which man was found on the island and is the first important pre-ceramic site yet found. Similar sites on the west coast have not been located. Raleigh Ferrell (1969) seemed to have noted some of the more complex problems in relating modern groups with their prehistoric ancestors. In his concise but all- inclusive review of Taiwan archaeology, he pointed out that "Scientific archaeology in Taiwan is still in its infancy" (Ferrell l969:4). In this light he did not try to connect specific tribes with specific prehistoric cultures as he had previously done in 1966. Once again Ferrell used cultural and linguistic data to divide Taiwan's aborigines into three main groups, the Atayalic, the Tsouic and the Paiwanic (with subgroups invoke TY“. diversit; ample ‘tin day langu arrived a 1959:73), the fairl; complexes , One to f0 Slbilitles (l) a CORZMOQ Tigrat .SOuiCu -bOQiC Eroto-r proto-t had PUt fol-HI 17 Paiwanic I and Paiwanic II). These major groups are further subdivided into seventeen languages. Although the language picture for such a small area as Taiwan is so diverse, Ferrell felt that it was not necessary to invoke many separate waves of migration to explain the diversity. "Four or five thousand years in situ would be ample time to produce the differences seen in the present- day languages, even had the ancestral Formosans all arrived at once and Spoken one single language” (Ferrell 1969:73). But because the archaeological evidence indicates the fairly sudden appearance of not one but three cultural complexes, he suggested that any number of migrations from one to four could be used to explain the present-day aboriginal Situation. He suggested the following pos- sibilities: (1) all the Formosan langua es develOped from one common ancestor in Taiwan: T2) two separate migrations, proto-Ataylic and proto-Paiwanic- Tsouic: (3) three movements, proto-Ataylic, proto- Tsouic, and proto-Paiwanic: and (4) four migrations, proto-Ataylic, proto-Tsouic, proto-Paiwanic I, and proto-Paiwanic II [Ferrell 1969:737. He then went on to state that at that time (1969) there was no real scientific support for the "colorful but unlikely earlier hypotheses” which he (1966) and others (Kano 1952) had put forth. The emphasis was upon eXplaining the cultural and linguistic diversity in light of the limited archaeological evidence at hand. and was Richard his arch; southeaS‘ tive stud the Yfians Prehistor was conrlI to be car: A: new inform Finds Of 9. early cult. soPhistica‘ Southeast A by world pr CUITUre, t}: mOCIEIS fOr Heine-Gelde data. It It AmePican nc Point K. c in the 1118*; 18 By 1970 W. H. Sung had finished his work at Olampi and was working on the Chang-pinian sites near Taitung. Richard Pearson (1972:317-330) had finished and published his archaeological survey of an area along the coast of southeastern Taiwan. K. C. Chang had finished his compara- tive study between the Lungshanoid site of Fengpitou and the Yfianshan site of Tapenkeng and published his general Prehistory of Taiwan (Chang 1969). The basic groundwork was complete and the stage set for more in-depth research to be carried out. Archaeological advances in the 1960's had brought new information and new insights into Asian archaeology. Finds of early pottery in Japan (Kamaki 1965: Kotani 1969), early cultivated plants in Thailand (Gorman 1969) and sophisticated bronze metallurgy in parts of mainland Southeast Asia (Solheim 1968) called for new interpretations by world prehistorians regarding the development of human culture, the rise and Spread of civilization. Earlier models for explaining culture change, for example Robert Heine-Geldern's (1932) migratory waves, no longer fit the data. It was realized at this time that European or American models of cultural dynamics could not be indis- criminately applied to all Far Eastern situations. At that point K. C. Chang (1970:MS) stated, "We are at the crossroads in the history of Far Eastern archaeology." of Eas t prepose: small we to condu area loo. Surroundi Choshui a: and Envirc from the C from the U aim of the cultural ec significam Hm Specifg exPeC‘ted I‘é 19 In an attempt to understand better the prehistory of East Asia in general and Taiwan in particular, Chang prOposed an intensive research project of a relatively small well-defined area in central Taiwan. The idea was to conduct an interdisciplinary "saturation“ study of an area looking at the ecosystems of the present and the past. The project was a modest model of Richard MacNeish's work in the Tehuacan Valley in Mexico. It was a new and imaginative approach for Taiwan and indeed all of the surrounding areas. By 1971 K. C. Chang was able to organize the Choshui and Tatu River Valley Project in AnthrOpological and Environmental Research. Support for the project came from the Chinese, from the National Science Council and from the United States National Science Foundation. The aim of the project was to investigate intensively the cultural ecology and history of a small, ecologically significant region, within an interdisciplinary framework. The specific objectives of the proposed research and its expected results were as follows: a. Intensive studies of the geology, geomorphology. zoology, botany. and soils of the Choshui and Tatu river valleys. Studies in each field are important in their own right, but together they provide or begin to provide the basic data for the construction of the present ecosystem of the area. b. An understanding of the major changes of the natural environments since the beginning of the Fle gec 6f sit lin anc pla« of seti resc ecos hist thei. refe: and 1 resou f" - SVMT bee ments and hills, t5; geomorphol T'le Vegeta temperature the high e1 Plain Where "9 See diff mm utilizi flm PeSult ?F0bab1y th 11101.12". ta 1 n tr 20 Pleistocene. involving the investigation of geology, analysis of old soils, identification of bony remains of animals from archaeological sites, and pollen analysis. Data along these lines provide the basis for reconstructing ancient ecosystems and for understanding man's place in them. c. Description and analysis of the cultures of the present inhabitants, emphasizing their settlement patterns, their different patterns of resource use, and their place in the current ecosystems. d. Reconstruction of past cultures and their histories, and understanding of the process of their develOpment and change, eSpecially with reference to their roles in ancient ecosystems and their various patterns of use of the natural resources [Chang 197laMS7. The Choshui and Tatu river drainages were chosen for study because of the great variations of the natural environ- ments and available resources. The area includes mountains, hills, terraces, basins and plains; i.e., all the major geomorphological planes found on the island as a whole. The vegetation, as a reSponse to the different altitudes, temperature and rainfall patterns, ranges from arctic in the high elevations to subtrOpical on the western alluvial plain where the rivers empty into the ocean. In this area we see different ethnic groups occupying different zones and utilizing a variety of resources. This is partially the result of known historical movements but was also probably the result of differential adaptation of dissimilar environmental niches. Chang (197l:MS) in comparing the Chinese and assimilated plains tribes of the lowlands with mountain tribes from the central mountain range found signifl 134-135 of the 3 rice an: in the ? hectares for swee (1972:74 vary in 1 The E perce of di half Villa distr high 1 tend “h -.e Suggeg is an impo aFChaeolog; areas have Change in E between mar ' man 8 mate. mental r98 ( 21 significant differences. The 1959 records (Ch'en 1966: 134-135) show that in the relatively low and flat land of the P'uli Basin, 4,042 hectares of land were used for rice and 689 hectares were used for sweet potatoes, but in the high mountain valleys around Jen-ai only 942 hectares were used for rice, while 62h hectares were used for sweet potatoes and #50 hectares for millet. Treistman (1972374) also noted that ”settlement patterns appear to vary in relation to altitudinal zones”: The Bunun, who inhabit the zone above 1000 m (30 percent live above 1500 m) live in small hamlets of dispersed homesteads, while the Tsou, over half of whom live below 1000 m, have centralized villages with satellite hamlets. The Atayal, distributed in the same zones, have occupied the high river valleys, where small diSpersed villages tend to have alliances within each drainage system. The suggestion is that differential use of natural resources is an important criterion that distinguishes cultural groups in the area (an idea that needs further testing through archaeological methods). The second point is that different areas have different potential resources, which may cause change in a group's adaptive mechanisms through which it secures a subsistence from the environment. The interplay between man and his environment is a delicate one, dif- ferences and changes on either end cause differences and changes on the other end. Through a study of that part of man's material culture by which he c0pes with his environ- mental resources we can better understand the peOple themselve different different what may 2 T} interactic ologically the first located, 5 chronologi exoavation Selected a< graphical ( remains, e1 The first stage PeSpOnsible‘ Basin on th Int Effects on In 1972 she W Whe f aWWPOIOE divers]: ty f 22 themselves. We must also be aware of the fact that differential adaptation by one group of peOple into two different environmental niches will produce through time what may appear to be two different groups. Thus the theoretical focus of the project was the interaction between man and his environment. Archae- ologically the project was divided into two stages: during the first stage (1972-1973) prehistoric sites were to be located, sites were to be ordered into a series of local chronological sequences and key sites selected for excavation in the second stage. The key sites were to be selected according to their chronological spread, tepo- graphical diversity, stratification, abundance of natural remains, etc. The author was invited by Dr. Chang to join the first stage of the Choushui and Tatu Project and be reSponsible for the archaeological survey of the P'uli Basin on the upper reaches of the Tatu river. Interest in the ecology of Taiwan and environmental effects on cultural groups was also noted by Judith Treistman. In 1972 she published a short but searching article in Science wherein she questions the ”schemata” devised by ethnologists, linguists, archaeologists and physical anthrOpologists for explaining the cultural and linguistic diversity found in aboriginal Taiwan (Treistman 1972:7h-76). Tr range in N remains ar A radiocar' date of Ad occupation to represei and into t1 from the m, previously instead th; Shows the that is ty She felt t upland Sit indigEDOQS (Treistman change cou Taiwan and 1:10:13 migr; Ve w I ork COnce qJQStiOnS of fUrthEF 23 Treistman, working in the high central mountain range in Nan-t'ou Hsien (county), found archaeological remains around the village of T'ungp‘u (elevation 1000 m). A radiocarbon sample from one of her test pits yielded a date of A.D. 785 1110 for the upper portion of the occupation layer. She felt that this date was too early to represent peOples who were "pushed out of the plains and into the mountains by a swelling p0pulation emigrating from the mainland" (Treistman 1972375), as had been previously suggested by Chinese historians. She suggested instead that "The cultural ecology of aboriginal Formosa shows the adaptive dichotomy between uplands and plains that is typical of Southeast Asia” (Treistman 1972370). She felt that there was sufficient time depth in the upland sites to allow for ”internal differentiation and indigenous cultural change apart from external pressures" (Treistman 1972:75). The implication was that internal change could account for the cultural diversity found on Taiwan and that scholars need not invoke waves of p0pula- tions migrating from the mainland. Keeping the preceding review of archaeological work concerning Taiwan's prehistory in mind, the following questions are recognized as still unanswered and in need of further study. 2“ --What refinements can be made on the local chronologies, especially the west coast plain area roughly between Taoyuan and Fengyuan, the central Taiwan area, the area south of the Choshui down to the Tainan area, the whole Central Mountain range, as well as the Ilan plain and the whole east coast? --How do these local chronologies relate to the established chronologies from the Taipei Basin and Fengpitou? --What has the environmental picture been for the past 15,000 years? Has it been stable or changing? Have there been significant changes in the flora or fauna? --What was the relationship(s) between humans and their environment? --What changes did different local ecological settings cause in human adaptive patterns to these different areas? --What was the geographic distribution of recognized cultures? --What changes, if any, do we see in the cultural remains through time? If there is stability, why? if there is change, why? --Were there Chang-pinian-like peOples on the west coast? 25 --Was the early corded-ware culture found all over Taiwan or was it restricted? If restricted, why? --What was the influence from the major Lungshanoid and Yuanshan centers in southern and northern Taiwan on the central west coast, the central mountain range and the east coast? --Were there exchanges of materials and ideas? What kinds of exchanges? --What were the subsistence bases for the different cultural groups, both temporally and Spacially? --What different levels of socio-economic develop- ment were reached in different areas and times? Why? --What were the different settlement patterns through time? What were the different locational factors? Why did man settle at particular sites? With these broad, general questions in mind for Taiwan as a whole, let us now turn to some more specific questions from the author's P'uli Basin study area. --When did man arrive in the P'uli Basin and surrounding hills? Before pressure from Han Chinese agriculturists, as suggested by Ferrel (1969.4) and Treistman (1972:7u-76), or later, as is generally assumed by traditional Chinese scholars (Hsieh 1960; C. Y. Chang 1953)? It is quest 26 --What was the ecological situation of the Basin before humans arrived? --What changes did humans make in the environment? Were they as significant as suggested by Tsukada (1966)? --What were the subsistence bases for the cultural groups living in the Basin? --What cultural changes occurred through time? It is hOped that the answers to some of the Specific questions concerning the P'uli area will add input in attempting to answer these broader questions, especially the questions dealing with the rise of agriculture. In the next chapter certain specific hypotheses will be proposed; the testing of these hypotheses should help answer some of the above questions. CHAPTER II HYPOTHESES TO BE TESTED Dr. K. C. Chang has suggested several significant ideas that have been helpful in attempting to understand the prehistory of Asia. These hypotheses are related to the question of agricultural origins and as such are of fundamental importance to a major problem area in anthrOpology. At this time three of Dr. Chang's ideas will be formulated into specific hypotheses for further testing. Hypothesis 1: As the result 2: swidden-type root and fruit agriculture in the uplands of central Taiwan, there was a change in the ecosystem at about 9000 B.C. In Tsukada's 1966 report of "Late Pleistocene Vegetation and Climate in Taiwan (Formosa)," he found that the vegetation for approximately the last 65,000 years could be divided into four pollen zones. Concerning the most recent pollen zone which he estimates to have begun about 9000 B.C., he says "Zone R (Recent) begins with the destruction of primeval forests, probably caused by human activities, and with the rise of the subtrOpical and warm temperate species" (Tsukada 1966:506). Several issues of importance are raised here. Was the change in vegetation 2? earl Occu Chan OC‘Jc-f-r-A-nmm-«H- _.. ks 28 noted at Jihyuehtan an isolated occurrence or was it an island-wide phenomenon that could be repeatedly demon- strated from other pollen studies in the surrounding area? Was the change caused by either natural phenomena like forest fires or human activity like swidden agriculture? If so, what evidence do we have to support it? This issue is important not only for understanding Taiwan's prehistory but for understanding the history of the greater areas of East Asia in general. Chang (1969: 193, 217—218) has used the Jihyfiehtan pollen profile as suggested evidence showing the post-Pleistocene shift to early forms of agriculture. In discussing the corded-ware occupation and the problem of early forms of agriculture, Chang (1969:217-218) notes: --that at about 9000 B.C. the vegetational history at the Jihyfieh-t'an area underwent a clear and persistent change accounted for by the continuous burning of the primeval forests and the continuous replacement of secondary growths. Accidental burnings of the forest (such as forest fires) cannot account for the persistence of the new pattern of growth, and forest clearance by human hands is not likely to have been undertaken except for plant cultivation. Since the sequence at #200 BP shows evidence of intensive farming and of grain cultivation, the contrast seems to suggest that the earlier form of agriculture was charac- terized by the cultivation of root and fruit crops. During this period-~from 11,000 to #200 BP--the only known prehistoric culture on the island of Taiwan was the Corded Ware culture described above, and the conclusion naturally follows that this Culture was associated with an early form of root and fruit cultivation. ,_’_A———- Char hor‘ one Saue crOp Sine: Why £ One b 29 Chang recognized that one cannot generalize about the horticultural activities of the whole island from only one palynological profile and that the Corded-Ware Culture in central Taiwan is not well established, but he felt that the concluded relationship ”is credible on the grounds of comparative significance” (Chang 1969: 218). With these reservations in mind, Chang has suggested that the Corded Ware peeples were clearing the forests and cultivating root and fruit crops (Chang l968a:82-83; l968b:522-523; 1970:62-65). Because the pollen profile suggests a vegetation pattern that is not the primary forest or the later cereal or grasses, Chang follows Sauer's (1952.103-104) suggestion that root and fruit cr0ps were cultivated prior to grass craps. At this point it is necessary to ask why the single pollen profile from Jihyflehtan was so important, why Dr. Chang was willing to put so much emphasis on this one bit of limited data. The answer is seen, however. when we reflect back to the mid-sixties and remember the particular ideas that were being discussed. Gorman and Solheim 's (1968) finds from Southeast Asia were suggesting a new, separate and independent center for the beginnings of agriculture. The suggestion was made that the Southeast Asian finds might even predate those from the Near East. The primacy of the Near East as the hearth for world ciVlliZ of agri it woul primacy agricul thus the there a Southeas also the going on adaptati Chang (3.} because ( Were cons the idea - which was Therefore, did TSUkad Taiwanus IV sampling e partiCular With Sol-he agrlCul till“! 30 civilizations was being questioned. If an earlier center of agriculture and metal working could be located in Asia it would rip into tatters previous assumptions about the primacy of the Near East as the center for the origin of agriculture, metal working, urbanism, civilization and thus the whole of Old World prehistory. Nor only was there a suggestion of a separate and earlier center in Southeast Asia for the beginnings of civilization but also that there was a whole different set of dynamics going on dealing with culture change and ecological adaptation. Thus, even though it may have seemed that Chang (and even Solheim and Gorman) was grasping at straws, because of the potential implications of the finds, they were considered important. It was in this context that the author conceived the idea to gather further data and test Hypothesis 1 which was fast becoming an accepted but untested fact. Therefore, it was important to determine first, number one, did Tsukada's pollen profile from Jihyflehtan truly reflect Taiwan's past vegetational history or was it a case of sampling error? And, secondly, if it occurred at that particular date, could this information be put together with Solheim and Gorman's data to develop an Asian agriculture horizon? Jihyfiehta archaeolc n u (1966:536 was cause: prehistor: period was remains at the vegeta Ware PEOII Preceded a to more €X' Posed root Um Subsis< Fengpitou g beWeen an corded Warf “Otes; COpd_m; archae, 9 is Sung IS turban: 31 IHypothesis 2: The swidden root and fruit agriculturists reSponsible for the 9000 B.C. vegetational change at Jihyflehtan were related 3g the Corded Ware peOples seen archaeologically at Fengpitou and Tapenkeng. Chang (1969:217-218) suggests that Tsukada's (1966:5h6; 1967:49) observed shift in the pollen profile was caused by human activity and that the only known prehistoric culture on the west coast dating to that time period was the Corded Ware culture best described from remains at Fengpitou and Tapenkeng. Thus Chang thought the vegetational change at Jihyfiehtan was caused by Corded Ware peOples. Because the 9000 B.C. change in vegetation preceded a later 2500 B.C. shift to cereal grain agriculture (see Hypothesis 3), Chang felt that the earlier shift was to more extensive use of root and fruit cr0ps. This pro- posed root and fruit agricultural pattern is suggested as the subsistence base for the Corded Ware peoples of Fengpitou and Tapenkeng. The importance of this connection between an early form of root crop agriculture and the Corded Ware pottery peOples is seen in the implications for the independent beginnings of agriculture in Asia. Chang notes: Cord-marked pottery assemblages were the only archaeological evidence of human activities on the island at that early date [Written before Sung's work at Changpin , and this forest dis- turbance strengthens the view that gardening mar S ; ThuS. 1 the caS domesti 1952xlC increase 0f the tc With £é£ug agricultu. 19593193-: 539-542), ObServed w. agricultuF; 32 probably had an early beginning in the cord- marked pottery horizon of coastal China and Southeast Asia [Chang 1968:827. Thus, the verification of this hypothesis would strengthen the case for Southeast Asia as an independent center of domesticated plants, as well as supporting the idea (Sauer 1952:103-104) that root crOps generally preceded grain crOps in the develOpmental sequence of tropic agriculture. Hypothesis 2: As a result of Lungshanoid agriculturists moving into the uplands of central Taiwan, there was a change in the local vegetation about 2500 B.C. Tsukada (1966:5463 1967:62) says that at 4200 :60 BP (Y-l612) in the Jihyflehtan pollen profile a dramatic increase in the frequency of grass pollen (about one-third of the total grass pollen may be cereal species) together with Liquidambar and chenOpodiceae indicate intensified agricultural activities." According to Chang (1968:5233 1969:193-199, 208, 220) and Chang and Struiver (1966: 539-5h2), the increase in grass pollen that Tsukada observed was due to the influx of Lungshanoid cereal agriculturists into Taiwan from the southeastern coast of mainland China. Chang's suggestion is that the Lungshanoid archaeological remains which he finds in central and south- western Taiwan were left by the people that caused the change seen in the Jihyfiehtan pollen profile. Inclusive in this hypothesis is the question of whether or not the observed cl": other area: The by the aut‘ri 33 observed change in the vegetation indeed took place in other areas of Taiwan or in just this one area. The next chapter deals with the methodology used by the author to test the hypotheses as stated above. At methodoloej so they ma; may be tes While other "a certain the second People liv statements World that hypothesis test impli exemplifie relected 0 direct Ost 119de (19 . CHAPTER III METHODS FOR TESTING HYPOTHESES At this point it is necessary to discuss methodologically how to test the above stated hypotheses so they may be rejected or supported. Some hypotheses may be tested by direct observation, "the sun is shining,” while others must be tested through indirect observations, "a certain group of people lived here in the past." Under the second method, because we cannot directly observe people living in the past, we must set out additional statements or test implications about conditions in the world that can be exemplified or demonstrated if the hypothesis is correct. Through direct observation these test implications are evaluated to see if they are exemplified in the world, and the hypothesis is accepted, rejected or modified in accordance with the results of the direct observations. As stated by Watson, LeBlanc and Redman (1971:8), to test a hypothesis: the invesitgator deduces from it propositions called test implications that must be exemplified in the world if the hypothesis is empirically correct. Ordinarily the investigator deduces several such test implications. Because the test implications follow from the hypothesis logically, they must be true of the world. However, the reverse is not necessarily the case: Because the test implication is found to be true of the world does not mean that the hypothesis may be empirically 39 fals COI‘I 35 true because it has not been shown to be false (disconfirmed). That is, the general form of the law does not say that the test implication is true if and only if the hypothesis is true: it leaves Open the possibility that the test implications could be true even if the hypothesis is false. It should also be pointed out that we cannot prove a hypothesis (or any proposition in a scientific paradigm) to be true, we can only "gain increasing confidence that it is not false. One can, however, demonstrate the falseness of a prOposition through its failure to have correct implications for the data'(Chartkoff 1975:3). In trying to assess the validity of the hypotheses stated in Chapter II as possible eXplanations of the observed phenomena, the following test implications were developed. Test Implications for Hypothesis 1: T. I. 1. If root and fruit agriculturists began swidden agriculture in the uplands of Central Taiwan around 9000 B.C., then we would eXpect to find a change in the vegetation reflected in the pollen record from the original primary forests to secondary growth. T. I. 2. If there were swidden agriculturists in the area at 9000 B.C., archaeologically we should find sites dating to that time period. T. I. 3. If the people were root and fruit swidden agriculturists, we should find prOportionally higher per- centages of swidden-related agricultural tools than 'vfa ce ti: r\ A: m. l 36 hunting-type tools or tools generally used in the pro- duction of cereal grain crOps. For example, there should be tree-cutting tools like axes or adzes for barking or cutting down trees and digging devices, with a general lack of reaping knives and grinding and milling devices. Test Implications for Hypothesis g: T. I. 1. If the root and swidden agriculturists in the uplands of central Taiwan are related to the Corded Ware peOples in other parts of Taiwan, similar cord-marked ceramics should appear in the archaeological record of the area. Similar ceramic vessel shape and decorations should be represented in the central Taiwan materials from that time period. T. I. 2. The lithic inventory ought to be similar also, if there is a direct connection. Test Implications f2; Hypothesis 2: T. I. 1. If there were an influx of village grain farmers into the uplands around 2500 B.C., we should notice a change in the palynological record to reflect an increase in grass or cereal pollen and related forms around this time. T. I. 2. If there were an influx of people at this time, there should be an increase in the number of sites and artifacts in the archaeological record. 51 hi es mi Si: te- LL11 Sh< rec in - area I‘ecc 37 T. I. 3. If there were an influx of peoples, the sites and artifacts at this time can be eXpected to differ from those of previous periods. T. I. h. If the p0pulation were involved in cereal grain agriculture, we should eXpect to find a prOportionally higher percentage of tools that relate to the production of cereals (for example, reaping knives, hoes, grinding or milling stones, etc.) than other types of tools. T. I. 5. If the village farmers in the P'uli Basin were related to the Lungshanoid peOples, we should find similar ceramic shapes, decorative motifs and manufacturing techniques. T. I. 6. If the population were an extension of Lungshanoid peOples from the west coast of Taiwan, we should expect to see similar lithic remains from both areas. In order to assess the above test implications, it was determined that certain classes of data would have to be collected. More specifically the major classes of data required included: verification of and locational informa- tion about sites: pollen samples for use in reconstruction of the vegetational history of the basin: radiocarbon samples to develOp the local chronology: ceramics to use in the development of an internal chronology within the area and for comparisons with other archaeologically recognized groups on Taiwan: and a sample of the stone 38 artifacts for comparison with outside groups and an indica- tion of the types of subsistence activities practiced by the people. In the next chapter the data collection methods used to obtain this information are discussed. pm in 801 f0} poi 331C 91" ac To ' ‘ {I‘ ’l[ ‘l l GflFERIV DATA COLLECTION METHODS In an attempt to fulfill the goals of the larger project as well as to examine the hypotheses set forth in Chapter II, the following classes of information were sought: site locational information (elevation, land form, soil, slope). faunal remains, soil samples (for pollen analysis and chemical analysis), datable materials and man-made remains. Stratified SamplipgiZones In attempting to locate the archaeological sites and recover the desired information, a stratified sampling program (Watson, LeBlanc and Redman 1971) was set up according to the different land forms found in the basin. To better understand the geomorphology of the P'uli area, a brief history of the physiography of the basins in central Taiwan is given here. In studying the physiographic history of the basins in central Taiwan, it is observed (Tomita 1951:3-5) that the Early Tertiary basins were formed by down-warping and faulting in an area of low relief that had been leveled in a previous subcycle of erosion. This down-warping and 39 7/V///////§ Z/Kfllxli //////4 E he @ \ rad/m. ’ ll [I‘ll Iii. # no ' MW L.£LGB£5~¢CD E 'A'PLANE— EH '0' PLANE- ’b'PLANE- 'e”/°LAN& 'C”PLAIV£- sac/ass e MOUN A/N El ’ FLANKS 7 a / i a 3‘ Fig. 1 Landforms in the P'uli Basin (after Tomita 1951) 41 faulting caused a series of mountain lakes to form. Fine- grained particles of clay and sand were carried into the lakes by slow-running rivers from the surrounding area of low relief. Then the whole island experienced an uplift that rejuvenated the streams and started the erosion cycle over again. Relatively coarse materials (5-8 cm in diameter) were deposited in the basins and formed the lower gravel beds. The basins filled and became dry land. The area was again affected by an uplift with faulting and the rejuvenation of the streams. The basins were now surrounded by high mountains, and the fast-moving streams carried boulders and coarse gravel (15-20 cm in diameter) into the flat basins forming extensive alluvial fans. The streams have continued to dissect the basin floors. Five major plains can be isolated (Tomita 1951:3-5). These five tOpographic plains and the surrounding slopes and mountains were chosen and designated as zones for the stratified sample (see Fig. l). The different plains, as described by Tomita (1951:4-5), varied not only in elevation but also in the history of deposition, the amount and direction of slope, material of which they are made up and the soil cover. Fig. 2 shows the zones one through six from which the samples were taken. Survey ngcegures Each of the six zones was surveyed on foot, and finds were plotted on 1: 25,000 maps and panoramic photos. 42 Fig. 2 Six Sampling Zones 43 The survey procedure consisted of walking each of the designated plots in a zigzag pattern. 0n the ground, surveys were supplemented with interviews of local farmers and workers. On several occasions they were able to take us directly to concentrations of artifacts or areas where stone coffin cists had been encountered during their labors. Plots were selected from each of the zones in a manner to insure that each variation of the zone was sampled. Some bias was noted in that the recovery rate from the areas of dense overgrowth con- sistently produced less than areas that had been cleared for farming. The effect of modern farmers was, however, a mixed blessing: several sites that had been previously reported (Liu 1956:55-57) could not be relocated. A partial eXplanation for the present absence of remains in areas where they were previously reported is that the land has been intensively cleared, re-landscaped and farmed. Many large features (stone cists, house remains, etc.) and even smaller artifacts have been destroyed and/or removed. Interviews with local Chinese farmers disclosed the fact that they did not recognize most of the lithic remains or pottery fragments as being man made. The initial reaction to this problem still stands: ”So much has been altered that it is felt that our sample [however largg7'will surely be biased and not 44 truly represent all of the activities that transpired here in the past. This problem must be recognized and dealt with in the interpretation stages of analysis” (Stamps 1972:2). Enough index artifacts remain to work with time-Space problems, but in most of the sites in- depth settlement pattern studies, activity area studies or studies of architecture will be very difficult. A total of twenty-eight sites were located. An explanation of the numerical system used for designating sites and provinces is presented in Appendix C. As sites were located in the survey effort, surface collections were made. At all but three of the sites (PL 21, PL 28 and PL 29), a total collection was made. At these three sites, artifacts were too numerous for a one hundred percent collection, so only a representative sample of the artifacts (tools and waste flakes) was taken. If the site displayed some unique feature or contained a large number of artifacts, test excavations were conducted. Test pits or trenches were located in areas of high artifact density to maximize the information retrieval. On certain occasions test pits were limited by the presence of modern features, such as tombs in the graveyards, modern buildings and agricultural crOps the farmers were unwilling to have disturbed. t... 45 Test Excavations Excavations were carried out using a light hoe and basket for soil removal where possible, and trowels and brushes when cultural materials were encountered. Materials from the test pits were screened through 1/4 inch screens and bagged according to levels. Excavations were begun in arbitrary 10 cm levels until the natural stratigraphic levels were encountered. Then levels were dug according to the natural stratigraphy, although if layers were greater than 20 cm thick, the artifacts were bagged into upper or lower portions of the level. Materials were washed, dried and labeled in P'uli before being boxed for shipment to Taipei for further analysis. Notes and drawings in the field were part of the preliminary interpretation, but most of the analysis was done in the laboratory at the Institute of Ethnology, Academic Sinica in Nankang. An attempt to recover bones for Faunal analysis by the zoologists participating on the project was foiled by the chemical make-up of the soil which destroys bone in a short time. As an illustration of this, the remains of a middle-aged woman buried thirty years ago were examined by the author during a bone-cleaning and reburial ceremony. It was noted that in just this short period of time only the long bones in the legs and arms, the pelvis, a few h (I) W (I) ’71 46 vertebrae, a few ribs and the skull remained. C. J. Shih (1953:14), in trying to explain the lack of skeletal remains in burial cists and bones in general from the site of Tamalin, suggested that they have been destroyed by the soil or by white ants. Fourteen charcoal samples were recovered in the process of excavation. These were immediately placed in new plastic bags and kept separate from the other remains. Samples were submitted to Dr. Y. C. Hsfi at the Radiocarbon Laboratory, Department of Physics, National Taiwan Univer- sity. The results appear in Chapter V of this paper. Soil samples were taken from each distinct level and the underlying sterile soil at each test pit. Three sites from diSpersed areas of the basin were chosen from which to extract pollen samples in an attempt to recon- struct the vegetation history of the basic (Chung, Hwang, Stamps 1973). A control profile was also taken at Wai- chia-tao-keng, Yu Chih on the southern rim of the P'uli Basin by the geologists on the project. Results of the pollen analysis appear in Chapter V. JIII . EL. .utdroi.u . .1. CHAPTER V DATA ANALYSIS This chapter includes the results of the stratified survey, a chronology for the archaeological remains found in the P'uli Basin and a description of the Pleistocene and Holocene vegetational sequences in the basin. Results pf the Stratified Survey. At this time we will describe each of the survey zones and the sites from those zones. ggpg 1. This sampling zone was made up of the mountain lepes that formed the perimeter of the basin. The elevation varied from 500 m at the edge of the basin up to 1924 m at the peak of Kuan-tao mountain on the northeast side of the basin. The watershed on these slopes was inward toward the basin. The mountains are composed of clay slate with varying degrees of soil cover, although it is generally shallow. The gradient of the slope varied, but it is generally quite steep and in some instances is a vertical cliff. Little of the original vegetation cover remains. Most of the area has either been used for slash and burn agriculture or to grow 47 48 hillside tea or fruit trees. Almost all large trees have been harvested by local lumber interests. Most of the hillsides are now planted with rows of fast growing pines which are harvested after about ten years and used in the construction industry (in a manner similar to our use of 2" x 4" or 2" x 6" beams). The original forest is characterized by broad-leaf evergreens with hard oak (Quercus glauca), rock oak (Quercus gilva), camphor trees (Cinnamomum Spp) machilus, acorn trees and several kinds of firs (Liu and Liu l959:note 3). Sites numbered 24, 27, 36, 37, 38, 39, and 40 were located in Zone 1 (Fig. 3). Site 24 was located on a steep lepe on the northern side of the basin just above 600 m elevation contour. There is a small stream one quarter of a kilometer to the west of the site. There was no sign of an occupation area, only a scattering of artifacts. Three stone hoe fragments, two adzes, three notched net Sinkers and two polished slate fragments were located in the surface collection. One of the polished slate pieces was perforated similar to the rectangular reaping knives (see Appendix B). The presence of the hoe fragments and reaping knives suggests some kind of cultivation. Modern Bunun farmers still practice slash and burn agriculture on similar steep slopes. The net Sinkers may have been used in the nearby stream. 49 , my'f/ 9‘ - . :23»? / « § . g3. ‘ d i "E ‘1 e/E-s \ X, . \p 42- \ LEGEND "’5 " : SLATE, ax sums FmMeA/rs .I'x' : 5mm: Haas (4N0 FfiA‘MéA/Ts) ’77:”: 71:67- TEL-3W ’ER':FI5H POND ”H’fi'house- ‘ ---—:F'aor mrH _ Gym 657 — 9:50”. 54mg. 0 30 Fig. 4 Site P.L. 38 do unuupP: . to La ill‘llllll‘l‘ 50 Site 27 is located on the top of a hill in the southeastern part of the Hsiao-pu-she area. Three stone hoes and one piece of polished slate were found. No other artifacts or features could be located. The area was probably cultivated at some time in the past. Site 38 is located on a hillside flat on the north bank of the Mei-chi River, seven kilometers east of the town of P'uli (Fig. 4). The elevation is 730 m above sea level. Soil here is a very dark brown. The soil cover is as deep as 1.5 m in some places. This area is currently being cultivated by a Chinese farmer who has capped a small Spring up the hill from the site. He mentioned discovering "many slab stone boxes” in the process of terracing and landscaping the area. The farmer remember them as being found in a row, end to end about 50 cm deep, but in no particular direction. He removed the slate slabs and used them for a floor in his house, a table and benches as well as a patio drying yard that was made up of about eighty pieces. Some of the larger pieces measured in cm: Length Width Thickness 186 52 3 154 35 3 140 36 2.5 84 45 2 9O 49 2 I 51 350W emmefiacAL 19:0er LEVEL I \— 0/5 Tabs-o \ $ R007 W : ,0 C;;, 54. c=g % =4”). e r {ig'fl"’l Q ...%. \ erk beam/v 301:. $ 09% ‘£.:. V ¢.’ (SEWR . (5)” . . ”v." ’o’e‘ Q (r 9%" € 0 .0: 9:33 5,, 9:. go; . .3, .vvv 8' 52:0 QMQV e o». mews" L --; 90‘ 090 x O 3.}8.’ égigggEHE.avacnnhaeces , a ’"' ’9” “Weasel O {’P.’ : Harm-av a” : 54.47:: at 64.47:- Humvee/73 1’54” : era/ve- mA- £54ka (II/vexas/A'reo) —o — ~ ’l?’:m£kx:k. .ScuL.:aAAmP%£hs Egfid Fig. 5 Site P.L. 38, profile drawing of west wall in test ‘ trench (between 8 and 9 meters south) 52 The source for these slate slabs is some 6 km away and their procurement, shaping and tranSportation represent no small expenditure of energy. A surface collection at the site produced several stone artifacts and a few pottery fragments. In an attempt to gather a more repre- sentative sample from the site and possibly locate one of the slab stone cists in situ, a 50 cm by 12 m test trench was excavated. Four levels were defined (Fig. 5): the uppermost level, when dry, was a light gray color disturbed by farming activities: the second level was a dark, moist, rocky level in which potsherds and several stone hoes were found: the third level was a lens of reddish clay which yielded sherds, a net sinker, some charcoal and two polished slate tool fragments: the fourth level was sterile, rocky soil. Slate slabs from a collapsed stone cist were found in a northeast by southwest orientation. All but one verticle slab on the southwest end had fallen over. The cist itself was full of dirt, with no human remains or grave goods. A total of 36 stone artifacts and 138 potsherds were recovered from the excavation. 0f the stone artifacts, 72 percent were rectangular or ovaloid stone hoes (Figs. 31, 32). Examples of the smallest rectangular hoes found in the P'uli survey came from this site. The ceramics from the site are of the SWK LBG/C class (see Appendix A, In 128). 53 1.. £5: * J;d’-Q=J':.4Kz#anz§§p.zgsrvvzs-—-Gh/fluwngsa.ewmuagge V’s/e'zkzgsfinvvze§;-egc:repeaggca4wer—'g: :4gggzaezaas- . Fig. 6 Site P.L. 39 ' 54 which dates to the early period of occupation in the basin. Pollen samples from this site were taken and analyzed (see pp. 94-99). The concentration of artifacts, the large number of stone cists and the color of the soil suggest to the author that this site was once a small hamlet. The good soil, the near-level land and the abundant water all suggest agricultural activities. This is supported by the presence of 72 percent stone hoes, only one net sinker and no arrow points. If the site had been used more for hunting activities or by hunters and gatherers, we would expect to find more arrow points and no ground cultivating tools. Site 39 is also located on a flat area overlooking the Mei-chi River. Site 39 is 2.5 km east of site 38 on the south side of the river (Fig. 6). The elevation is 770 m. It is also located just below a spring on the mountain. This spring has been channeled into rice paddies and sugar cane fields by Chinese farmers who presently occupy the site. The soil here is also dark and fertile. Measurements revealed that the dark soil was 60 cm deep in one area and 80 cm deep in another. Several slate fragments were found scattered around the area. These may have come from slab stone cists or from remains of prehistoric houses. The surface collection produced fewer artifacts than at site TUAG P’u ‘ J\ "l... EQE-QO s TONE‘ cxs'r ”H ” Hal/5b ox zed/z. mm ”#12:" FEATURE-s ”RP” km:- P4771}: Pig. 7 Site P.L. 37 56 38. Thirteen pieces of worked stone were recovered: 5 stone hoes, 5 sandstone fragments, 1 net sinker and 2 disk shaped pieces of worked stone (the waste centers of bracelets drilled with the use of large diameter bamboo and an abrasive). Two pecked stone bowls were also located at the site, but their origin was in question. Only three light brownish gray sherds were found. The site could have been a prehistoric hamlet or merely an outlying field worked by people living in other areas. Site 37 is situated on a flat point that juts out from the steep slopes above the confluence of two fast- running streams (Fig. 7). It is in the mountains of the southeast quarter of the basin. Actually, the P'uli Basin itself cannot be seen from this site-~only the rather narrow canyon that turns just before it joins the basin. The point that juts out from the mountain has been terraced and is currently planted in orange trees. When the original terracing was done, several stone cists were destroyed. Our surface survey located one rather small rectangular cist (101 cm long, 37 cm high, 45 cm wide) still intact in the side of a terrace. It was on a northeast, southwest orientation. Further investigation found it to be silted in and filled with fine grain soil but with no human remains or burial goods. Also located in the side wall of one of the terraces at this site was a dish-shaped impression r 4 57 115 cm across and 30 cm deep. Several pieces of charcoal were recovered from inside the feature. Its function could not be established. The only artifacts recovered from this site were two oval-shaped hoes. The site was probably occupied, as suggested by the presence of the hoes, the stone cists and the dish-shaped depression, but only by a few peOple probably for a short duration or seasonally. The charcoal sample that was recovered from the depression was dated by the NTU radio carbon laboratory as "modern” (NTU-198). For discussion of dates, see pp. 93- 94. Site 36 lies on a gently sleping peninsula that comes out from the southern mountains onto the flat basin floor. No artifacts were found, but broken slate slabs were found scattered over the lepe. The purpose of the slab fragments is unknown. Site 40 was located by Hwang and Liu, the geologists working on the project. This site is located on a high ridge in the southwestern quarter of the basin. Artifacts, including stone hoes (one spatula example), polished adzes, a small spear point and slate fragments were found on the crest of the ridge and down both sides. Both the soil cover and the cultural deposit were thin. An interesting fact about this site is that the nearest water was several kilometers away. The site was probably an outlying field l | '| ‘1, ill I {I II: If v 58 which was only occupied during times of Spring planting or fall harvest. ggpg g. This sampling zone is found only on the western half of the basin. It consists of slightly undulating tablelands that slope gently from west to east. These tablelands are made up of upper gravels, covered with deeply colored reddish laterite soil. The elevation of this zone is around 700 m above sea level. The original vegetation has been removed and the area is currently planted with rice, bamboo and shu-Shu. No archaeological sites were found in this zone. Liu (1956:56) refers to one example of chipped stone tools found in 1902 by a Japanese official. Attempts to relocate this site or any others were unsuccessful. The lack of sites may be due to a general lack of water. to some unknown characteristic of the red soil or some other unknown factor. aggeyg. Zone 3, like the previous zone, is an upper gravel bed tableland, covered with deep reddish laterite soil: however, it is lower in elevation (from 700-500 m) and is more eroded than Zone 2. It also SlOpes from west to east. Eight archaeological sites are found in this zone, although it is noteworthy that they are all on the border of this and other zones. Site 25 is located on the southwest corner of a small hill in the Hsiao-pu-She area on the north Side of 59 II I’ A :7557 FIT L. E- Q g N19 ”N'.’NE-T SINKERS Fig. 8 Site P.L. 25 60 the basin (Fig. 8). A small stream runs past the hill on the west side. The Site slopes gently to the southwest. Elevation of the site is 520 m. The surface collection produced 5 stone hoes, 12 sandstone fragments, 2 adzes, l reaping knife, a polished cube, but most surprisingly 17 notched net Sinkers. A 1 x 2 m test pit at the Site (25A) produced more of the same kinds of artifacts, but also 78 brownish orange potsherds. The sherds came from an undisturbed level at the bottom of the test pit (Fig. 9). These classes of sherds (HPS BO/C and HPS BO/M)l were also found at site 28 less than a half km to the northwest. Because both sites have the same classes of pottery, and because the HPS B0 pottery is similar to the dominant pottery (SWK LBG/C) at site 28, the two are considered to be contemporaneous. This is supported by the fact that similar lithic assemblages are found at the two sites. At site 25, 65 percent of the recognizable lithics recovered were notched net Sinkers and 33 percent stone hoes. The suggestion is that fishing activities were quite important at this site. In recent times carp, white fish, silver fish, loach, croaker and catfish have been caught in nearby lakes and streams. 1See Appendix A for an explanation of the abbrevia- tions of pottery types. 61 .1 .. . ‘ sifiu.yt «Mr: ,X .LY ,xx/ 2356: - . >023. kiwi , «4wa . . 6A. / c<< t u ,x t\ a no ,. . m p .2 W é a N p M w m a.“ e a m... , , a W .rm P x/ .w v cwm :— g 0C aft M“ «TA MTG l a , ,ci \¢:\,xt,,;gg «a x m.. . ._ VA», Ker/A y. < x ‘ / o . WW ,n/K\.W ”Ammowfl t/,x,,//\\/\A \A< (21 Fig. 9 Site P.L. 25, profile drawing of east wall in test pit A. I'llIII‘II 1 I lilll|llll|lllll!.(lll J 62 Less than one half km to the north is Site 28, situated on the crest of a low finger-like ridge that juts out into the Hsiao-pu-She area. On the point of this ridge is located the Chieh-ling-shih temple. The site begins immediately behind the temple and Spreads up the hill for approximately one half a kilometer. The elevation is 500 m, with a 4 percent gradient to the south. This Site was extremely productive and yielded 136 pieces of worked stone in the surface collection. Because of the large numbers of artifacts located at the site, the collection effort was divided into three zones. A comparison later showed, however, that the same types of artifacts, in similar percentages, were found from each of the zones. Stone hoes were most numerous, making up 52 percent of the total collection. Trapezoid, Spatula, rectangular, ovaloid and pointed Shaped hoes were all found at this site. Nineteen percent of the artifacts were notched net Sinkers, 2 percent were adzes, 2 percent were knife fragments and 1 percent was polished slate point fragments. A 1 x 2 m test pit near the center of the site (where 8 sherds had been found) yielded only 8 stone arti- facts but 62 potsherds. The ceramics came from one cultural level and were of three main calsses: PL GS/M, SWK LBG/C and HPS BOAM and F. Most numerous were the SWK LBG/C sherds lilIwafl A l I (A ll... 63 that date to the Shuiwaku Period. A few bits of scattered charcoal were found in the test pit, although not enough for a radio-carbon date sample. The presence of pottery and the wide range of artifacts suggest that the site was once a hamlet or small village site. Terracing by modern farmers, however, has probably removed evidence of any structures. No stone burial cists were found here. Agricultural activities seem to have predominated at the Site, as is suggested by the large number of stone hoes. The small number of reaping knives (only 2 percent), however, suggests either a different method of harvesting (as mentioned by Chen 1968, where the foxtail millet ear is plucked with the stock) or possibly different types of crOps being cultivated at this site. The fact that 19 percent of the artifacts were net Sinkers and only 1 percent were projectile points implies that fishing was much more important than hunting. Hunting, however, could have been carried on with the use of traps and snares, as it still is in certain aboriginal areas of Taiwan today (Chen 1968:30, 33-39). Site 35 is located on the eastern Slope of a small hill on the eastern edge of the Zone 2 tableland near Hsiao-pu-she. From the Site only two net Sinkers, two hoes and a waste flake were recovered. The Site represents no more than a campsite or more likely a field. 64 /\\ ‘ " at” “ ' 4.; eENo { ’14 ’- ‘E ” - r537 F/7'5 ‘ © .97er cueF/v A: Fig. 10. Site P.L. 19 65 Site 41 likewise is a small Site west of Hsiao- pu-she. Seventeen artifacts were recovered. 35 percent hoes, 53 percent net Sinkers (both notched and grooved) and the remainder, sandstone waste flakes. The site was probably an outlying work site, possibly a fishing station. Two different streams do run within a kilometer of the site. Site 44 is located on the leping tableland approxi- mately one kilometer west of site 28. Sixteen hoes and sandstone fragments were recovered. No pottery or worked slate were found. Most of the sites located in Zone 3 were found in a semicircle around the Hsiao-pu-she area on the north side of the P'uli Basin. A survey of the tablelands in Zone 3 on the southwest side of the basin located no sites. A survey of the north-south ridge on the east side of the Tao-mi stream did produce three artifact concentra- tions. The largest was site 19 located on the northern end of this ridge (Fig. 10). The site has an elevation of 470 m and towers over the surrounding area. Steep cliffs surround the site on the north, east and west. A surface collection produced both stone tools and pottery sherds. Test pits were put in at five different areas of the site. Test pit 19A was located on the northwestern part of the site. A large rectangular slate slab on the surface 66 indicated a possible structure or a burial cist in this area. Only one level was encountered. It ranged from the surface to a depth of 30-40 cm and was made up of brown soil mixed with pottery and charcoal. The charcoal was collected and submitted for dating. The results showed a "modern" date (NTU 204), which agreed with the find of a small metal pipe from this level. Protruding from the east sidewall of the test pit was a 60-cm-long piece of worked slate. After finishing the test pit down into sterile soil, the area around the long Slab was cleared. A 60 by 44 cm rectangular cover of worked slate was found 10 cm below the surface. When the cover was removed it eXposed a 54 by 55 cm box made of four upright slabs. At the time, two large ant nests were located inside. It was cleared to a depth of 55 cm, where three large stones were encountered. There was no slab base in this cist. Beneath the three large stones was the reddish, sterile gravel soil. This cist fits the description given the author of burial cists used by aborigines in the P'uli area in the early part of the century. From test pit 19A one rectangular, three oval and one waisted-shaped hoes, two adzes and 92 sherds were recovered. 0f the pottery sherds, 76 were H LBG/C, the major type found in the Modern Period. Eight SWK yo/M, four PL os/M and four swx LBG/C sherds were recovered from 19A. The second test pit, made up of 67 z:~——-£in5fiuaaé.5an_ LiA/Q ‘ 0 Rock 5 KSHA 0.5-0) Fig. 11 ‘Site P.L. 19 South profile, test pit B 68 squares B (Fig. 11) and C, also yielded only a Shallow cultural deposit (30-40 cm). The only features located were one large flake, three net Sinkers, several fire- cracked rocks and 259 potsherds. The dominant class of pottery again was H LBG/C. Test pit 19D was 1/2 m by 4 m. It also contained only one cultural level between the surface and the sterile subsoil at 55 to 60 cm. This test excavation produced 91 sherds (84 were H LEG/C), 1 grooved net sinker and l grooved pebble. Fifteen large rough sandstone cores and 65 flakes or fire-cracked rocks and 1 small grinding slab also came from this trench. Test pit 19E was a 1 by 2 m pit on the east side of the site. This excavation had only 1 stone hoe but several rough cores and flakes and 415 pottery sherds. H LBG/C accounted for 355 of the 415 sherds. Although the cultural materials were found deeper at 19E (down to 117 cm), no features or different stratigraphic cultural levels could be discerned. 0f the total 924 pottery sherds found at Site 19, 774 or 83 percent were of the H LBG/C class. One of the differences between this site and the other sites was the large amount of rough cores and fire-cracked rocks. Also different was the fact that no knives or projectile points were found. Perhaps this was because metal knives and points were traded in from the Chinese on the coast. Metal 69 MEI CH ’/ STREAM star \72. '\ 'd ‘5. .(c it L. ééENO ’14 '— '5 "- raver P1725 ”H ” - House- Fig. 12 Site P.L. 29 . 7O artifacts being more rare would likely not have been discarded but would have been reused or made into some- thing else. The large numbers of pottery sherds, tools, cores and flakes suggest that the site was important, but probably in the Modern Period. South along the ridge from site 19 are sites 32 and 33. The former, located on the top of the ridge, was a disturbed burial cist: and the latter, located halfway down the eastern slope, was where three H LBG/C sherds were found. app; 3. This sampling zone is assumed to have been develOped by streams flowing from zones 2 and 3 (Tomita 1951:4). Zone 4 has been subdivided by erosion into several scattered, narrow plains, all of which are almost completely flat. This zone is composed of gravel beds covered with 1-2 m of laterite soils. The soils are somewhat lighter in color, particularly near the surface. The elevation is around 450 m. Archaeologically, this was the most productive area of the survey: thirteen Sites are located on these flat river terraces. Site 29 (Fig. 12) is located on the southeast corner of the terrace in the Shui-wa-ku district. This terrace is the first one up from the basin floor. A vertical cliff drops some 20 m to the river below. The Hsiao-pu-she stream flowing from the north joins the Mei- chi stream flowing from the east at the base of the cliff 71 (f l How Nil/l}: ‘F—I’ . aw”””’#’#’- ,1 ..,———Z::::::j // / 1- a co (3 0 ’ //’_ // ‘“-——~ -—-£57’£&-fiZ;'L_:E- %:”<:;::i:c ,44:f\0:9 ‘Eéégggjo 7°'~/°o7'7'E-. ” . “~-->< . ' a: ‘-’ {9. I’ a, I! 5 'srq} "- :><: % , _____[______”'[[,,1.~ . G). e 4;, X I: ‘ 1%) X o {9' l_. — .35 ' - e - .. . X o» 0' 4' o .'d} " ‘5 agrees LEE-O @ :RA macaw/301v 6.4ch N70 I94 $ - sou. JAMI'Lé Q : flApxoaAm/v 6MP4£ 4’70 /75 a - Jock 5 : .5447: ”I " - ”z“ : LGV‘LS '0, an E M — Fig. 15 Site P. L. 21, profile drawing of west wall in test pit C 78 placed in the center of the mound. Excavations in pits A and B reached sterile soil at 80 cm. The stratigraphy and the artifacts from pits A and B showed no Sign of more than a one time period being represented. Test pit C, however, uncovered Six different stratigraphic levels above the sterile gravels that lay at a depth of 230 cm (Fig. 15). Test pit D was a 1-1/2 by 1 m enlargement on the southwest side of test pit C. Soil samples for pollen analysis were taken from test pits A, B and C. All but the earliest levels in pit C contain similar ceramics classes and probably date to the same time period. The ceramics from level 6, at the bottom of the pit, differ in fre- quency from those found in the later levels. The fact that several SWK LEG/C sherds, the index type of sherds found in sites from the Shuiwaku Period, are also found in level 6 suggests that the site was first occupied in Shuiwaku times and then again later in Tamalin times (see TML LEG/C and LBG/M, Appendix B). The majority of the 4,555 sherds recovered from this Site date to the Tamalin Period. In fact, only 27 sherds from other sites are classified as sherds of the classes found at Ta-ma-lin. In broad terms, the gray ceramics found at Ta-ma-lin are similar to the gray wares found in other areas of central Taiwan (Chang 1969:206) and P'uli Basin. More Specific ceramic traits, however, tend to isolate site 21 at Ta-ma-lin from the other 79 Sites in the basin. This isolation was probably due to a temporal difference. The full range of stone artifacts recovered in the overall survey were also recovered at Site 21. Stone hoes accounted for only 13 percent of the lithics recovered, but trapezoid, spatula, rectangular, oval and waisted examples were found. K'o halberds made both from sandstone and Slate were uncovered. Adzes, net Sinkers (both grooved and notched) and rectangular reaping knives were discovered. Awls made from slate, nepherite and shist were located. Also of interest were Sixteen quartz flakes found scattered through the levels of pits C and D. The most striking find from this Site, however, was 49 projectile points and 13 partially completed blanks or cut Slate fragments. Twenty- six percent of the lithics recovered were points or point fragments. Willow-leafed, triangular and tanged examples are all present at this site (Figs. 38, 39). Unique to this site was the find of three "boot-shaped" pieces of polished slate (Fig. 41). Their function is unknown. Inter- views with local residents revealed two large collections made by amateurs. One collection from this area, made by Hsieh Tien-hua, consists of close to a thousand artifacts. A long-time resident of the area, Mr. Hsieh claims to have recovered artifacts from most of the construction sites in the last fifty years. He estimates that eighty stone cists 80 have been destroyed on the southeast end of the Ai-lan terrace. Another neighbor, a competing collector, claims to have cleared approximately thirty cists from his and surrounding prOperty. Both men agreed that most of the cists were empty, although occasionally one would contain pottery or worked stone. They also confirmed the general east-west orientation. Hsieh claimed that the cists were wider on the east end and narrower on the weSt end, to accommodate extended burials with the feet to the west and the wider shoulders to the east. Site 21 was undoubtedly a large village where a variety of subsistence and community activities took place. The level land around the site with its thick, fertile soil cover is ideal for agriculture. Streams and a Spring provide water and fishing resources. This site is also located at a natural crossroads where the Nan-chiang stream flowing east to west is joined by the T'ao-mi stream which flows north from the mountain basins located to the south. Ta-ma-lin was the largest and most productive site located in the P'uli Basin. Another Site of extreme interest is number 30, located on the northeast corner of the Ai-lan terrace (Fig. 16). The Site is located in a modern cemetery. Workers excavating dirt to pile over the t0ps of modern Chinese graves had encountered three Slate, stone burial 81 I fi‘ . —X k A4005" Wi 4 r J .4 (.3 *;~~rawv ‘, "W-“gfi'r J; ”M I. ‘8 ~ W“‘ {,7 v. , . v I" v. ‘7. ‘ 114’... ”5” TbéT ARE-45 —”‘—x.__ P cc ’9‘" Hot/5&5 I Fig. 16 P.L. 30 O\‘——-—-———_ -h“\O/// \‘ 1 5:990 G 6157 \ 6’57. Fir A \. F 6 > ‘V 0 I, \ I “ \‘ / , e \ . / \ ,>\ ‘ [A \z/ ..\ 2, ‘\ - Lgégo . #— lo ’ ‘ ICAAWOCVMUQGMJlSAAdflbl-flNH1/Zoaj ', in“:- Fig. 17 Site P.L. 30, profile drawing of intrusive pit between burial cists A and F. 82 cists. A closer search and test excavations uncovered five rectangular cists, one intrusive pit and two areas with concentrated horizontal Slate slabs. The maximum cist measurement in cms were: cm below Feature (cist) Height Width Length surface 30A 36 51 179 72 BOB 25 54 189 65 30 E-l (disturbed) 20? 36? 150? 50 30 E-3 37 43 170 22 30 F (disturbed) 28 45 ? 72 All of the cists were oriented east and west and made of slate Slabs with both covers and bottoms. Cist 30B had a 4 cm hole in the center of the bottom. Liu (1956:83) illustrates an example from Ta-ma-lin with a hole in the cover. All of the cists had filled with fine soil that washed in through the joints. No bones or grave goods were located in any of the cists. An intrusive pit 46 cm wide and 56 cm deep was located between cists 30A and 30F. The top of the pit was on a level equal to the cultural materials found above cists 30A and 30F (Fig. 17). From this pit several waste flakes, one hoe, one polished slate k'o halberd, an adze, four slate fragments and five river pebbles were recovered. A radiocarbon sample from this pit was dated at 3207 i 96 (NTU 203). The associated stone cists were at 83 least this old or older. Two other radiocarbon dates came from areas with concentrations of horizontally-laid slabs. It was thought that these slabs may have been the remains of house floors or living areas. Structural remains, however, could not be located. The floor in 30C-2 was dated to 3282 t 98 BP (NTU 201). This was the earliest date obtained from the survey. Another possible floor (PL 30 D-2) was dated to 2994 i 90 (NTU 202). The ceramics from site 30 are of the SWK LBG/C and swx YO/C class, both of which are characteristic of the Shuiwaku Period. The stone artifacts suggest that both agriculture and fishing were practiced. The importance of this site is that it not only provides several more examples of burial cists but also provides dates for the ceramics found in the Shuiwaku style. The Nan-chia stream cuts through the 450 m terrace isolating the Ai-lan portion in the center of the basin from a section that merges with the lepes on the south rim of the basin. Located on this southern terrace is site 2. The site is level, with thick soil cover. Water resources are located to the north and east from the year- round streams--the Nan-chiang and the T'ao-mi. The recovery of net Sinkers (21 percent) suggests that fishing was important. Trapezoid, spatula, rectangular and oval shaped hoes constituted 42 percent of the stone tools 84 recovered. One grooved pebble, one worn quartz crystal and a rectangular reaping knife were also found. A small stone cist was found. The base slab was 27 cm wide and 60 cm long. There was no cover stone and the sides leaned in, touching at the center to form a triangular cross section. Vertical Slabs were placed at both ends. The orientation was east and west. Three rectangular hoes and seven SWK LBG/C sherds were found in the fill around the cist. South and east of site 2 is a mountain Spur that juts north into the basin. On the east side of this Spur was located a small flat area with artifacts on the surface. The area was designated Site 34. The surface collection produced two Spatula hoes, three flakes, one Slate k'o halberd and three grooved net Sinkers. A l by 1 m test pit near the center of the Site contained a rectangular cist 95 cm below the surface. The cist was on an east-west orientation. The length was 200 cm, the width 65 cm and the depth 32 cm. Eight H LBG/C sherds were recovered. The thin scattering of artifacts suggests only a short occupation time for the Site. Immediately south from the previous site is a small hill which is being eroded by a tributary of the Nan-chiang stream. Protruding from the face of the cliff was half of a stone cist. The area was designated Site 31. A surface 85 survey in the area recovered seven hoes, two net Sinkers (one grooved and one notched) and one k'o halberd. No ceramics were recovered. The local farmer said that in clearing the land he had thrown many hand-sized pieces of worked stone over the cliff. The stone cist measured 30 cm high, 46 cm wide and 104 cm long. It may well have been longer, but the eastern end is now missing. The area was probably occupied as a living Site at least briefly. The gentle hill could easily have been farmed and the nearby streams fished. §2pg_5. This sampling zone is found mainly on the eastern half of the basin. It is made up of old flood plains or river terraces with no laterite soil cover. The general SlOpe is from east to west. In this sampling zone only two sites were located. Site 23 was located south and east of the town of P'uli in a brickyard. Excavations for clay had uncovered a "stone box' which contained what the owners described as a Semi-lunar knife, a polished adze, two bracelet cores and a piece of worked stone. The description of the stone box fits the rectangular stone cists found elsewhere in the basin. The cist was estimated to have been 180 cm long. This Site has since been totally destroyed by the mining of clay for bricks. Site 46 is located on a very rocky, terraced area in the southeasternmost corner of the basin. Only two 86 artifacts were found: a small stone hoe and a worked slate fragment. These artifacts may well have been dis- carded by peoples just passing through the area, for no concentration of artifacts or soil color change was noted. ggpe 6. This is the lowest zone in the basin: it includes the recent flood plains and is lower than zone 5 by 2-5 m. There is no laterite soil cover and the slope is from east to west. Only one Site was located in this zone. Site 45 was a stone cist in a Korean apple orchard. The orchard is on t0p of an alluvial fan formed by a stream which descends from between two steep mountains. The cist was roughly square, 50 x 42 cm. The depth of the slabs (there was no bottom) was 80 cm. A local farmer who had dug others in the same area said they had been made by aborigines at the turn of the century. He claimed to have witnessed the burial of an aborigine in a similar cist. The body was placed in the cist in an upright, flexed position. Summary of Supvey Results The total survey located 28 sites or artifact con- centrations. The most heavily occupied area was the river terrace in zone 4, where ten sites were located. Figure 18 lists the sites found in each zone. No correlation could be shown for sites from different time periods being located 87 Zone Sites by Identifying Number Total 1 24, 27, 36, 37, 38, 39 and 40 7 2 - O 3 25, 28, 35, 41 and 44, 19, 32. 33 8 4 29, 42, 43, 22, 26, 21, 3O 2. 34, and 31 10 23 and 46 2 45 l Fig. 18. Sites located in each of the sampling zones. in any particular sampling zone. The preferred site locations were near streams or sources of water, on relatively level ground, in an area with good soil cover. Chronology gf’thg_glpli_B§§i3.Archaeological Remains. Fourteen charcoal samples were recovered and submitted for radio carbon dating (Fig. 19). The resultant dates combine with the ceramic analysis (Appendix A) in distinguishing three major temporal periods of occupation for the P'uli Basin. On the basis of stylistic similarity, the following periods were isolated: Shuiwaku Period (1332-431 B.C.), Tamalin Period (24? B.C.-A.D. 167), and Modern Period (A.D. 1500-present) (Fig. 20). These periods generally correspond with Chang's (1974:272) Tamalin Phase (1500 B.C.-A.D. 700) and Historic Phase (A.D. 1650 to present). The Shuiwaku and Tamalin Periods designated for 88 swmmm Has.m one Some mmpmc Comancowumm ma .wflm auoooz Hus .mH Has.a somusez csoooz 0 .sm has.a moausez cameos m .em Haa.a smflusez mm H sod .o.< mm H mmea Sue .Hm has.a NaH-sez mm H mas .o.< mm H smma sum .HN “Ha.a maausez mm H sea .a.< mm H papa 3-0 .HN aaa.a aa~-sez .o.m no H awe no H deem m:o .Hm Has.a maausez .o.m so H new so H seam m-o .Hm has.a saa-saz .o.m as H an: an H Hmmm m-m .am has.a oomusez .o.m cm H SSOH oa.H seam mum .om aaa.a momusez .o.m om H smma as H aomm o .om Has.a momnsez .o.m as H mmma mm H mmmm m-o .om Has.a Hom-aez Aommm u omflqumamxv Aosmm u .o.< poo .o.m oaHH:MHomv .a.m seaflooog soossz Ams~.ssoa moors seesaw sesame Hospaoopoo3_ca wooded Houseaao om 89 use Hapmmoo \ \\ \ \\ \ \. \ s s r . x... \\ x. t a \ a \\ x... \. \ \ 1' ... {solidi-g. 9: . .I. .1: 4:140:10! .lsflm‘).N 5‘“...le 79.1%). v. 2 £331.11 ”x . x \ ..V 1 dill-.0: ! t It!!! All-It'll x. \ . s l . x \. _ .c c .e. r s a . ;-knlvlsir.»-n.sar:m;+as«.x:m:+ ..4 i.. w x .. \ \ , v .\ , x. r... . . \ \ .r x \ . x r J :rx (5.? .. r .. \r , J 4\.. x. L . r u . H L!!! ( 71.5.4313. 1 .~.\.\ \\ \\ .\\ . . . ._ . .\ . c. .\.s l x_ a a . A. k. M i r a \ .1 V . .1 N .3 .... w r. 1:111:11“. . _ \ l4 s . . ..\ . \ .\ \ ex x W t a. \ . x \ .x K x \ \ v. \ \ fi. . .\ \ x. \\ \ . \x .\ r \\ \\ . xx \ \\ . . . .c .. x .\ .. .\ x .. \ \ \ . .. r. x r . ,. x .L.L...~ \r\\ihlw .x. ...|.i i K it, acaopcsoz capsohnwh .Has.m moomuuoe moomuuoa nooam chom Hmuvcoo dogma omumoo ocHnmusvmm cam wchHm ocoarwam uo>am oases: Hoeoooo .wam om comm chasm csvsoflmnom.me om ooom chasm SNSSSH.:ome om oomH ommnm Swansea oomaloom D< cauopmdnopoum owed o< oauoeoam 90 the P'uli Basin are grouped together by Chang as the Tamalin Phase. Shuiwaku Period. This period is chiefly charac- terized by the presence of SWK LEG/C pottery. Examples of this pottery type are found at sites numbered PL 19, PL 21C-6, PL 28, PL 29, PL 30 and PL 38. The earliest date for this period comes from PL 300-2 and dates to 3282 i 98 BP (NTU 201). The sample was recovered from what appeared to be an occupation level h2-55 cm from the surface in good association with horizontal slate slabs, SWK LEG/C pottery. flaked and polished tool fragments and burial cists (PL 30A and F). Another sample from an intrusive pit between the burial cists (PL 30A and F) yielded a slightly more modern date of 3207 i 96 (NTU 203). A third date of 2994 t 90 (NTU 202) from PL BOD-2 is also found in association with the same pottery and tool types. The site seems to have been occupied over at least a 288-year period. although it was not determined if the occupation was continuous or reoc- curring. A similar pottery and lithic assemblage is also found at PL 29. The single radiocarbon charcoal from this site (PL 293-3) dates to 2381 i 71 BP (NTU 200). From that level and the level below came several polished adzes, polished slate arrow points, reaping knife fragments, flaked hoes, double-grooved net sinkers and polished slate 91 k'o halberd fragments. The suggestion is that the site had been occupied for several years before the dated level. A surface collection and text excavation at PL 38 produced only SWK LEG/C sherds. Although the stra- tigraphy showed two different stratigraphic levels, the artifact inventory was the same. Even though stone hoes predominated, polished k'o halberds, adzes, reaping knives and a net sinker were also found. Of special interest were five flaked hoes which were much smaller (average 9 cm x h.5 cm x l.h cm) than most of the others found in the basin. Another important site from this period was PL 28, located on the s10pe behind (and probably under) the Chieh Ling Temple in Hsiao-p'u-she. Although only 65 potsherds were recovered. they represent four different types, all characteristic of the Shuiwaku Period. This site was most important, however, because it yielded 157 lithic artifacts which help us to show the range of variation in the lithic tool types for this time period. At PL 21 several swx LEG/b sherds were found scattered through the various levels, suggesting that this style of ceramics continued in a very minor way after the Shuiwaku Period. The majority of this type, however, came from the lower levels; 78 percent of those recovered from PL 210 test pit came from the lowest level. 92 Tamalin Period. Possibly the largest and unquestionably the most productive site in the basin in Ta-ma-lin (PL 21). In the 1972 excavations, three test excavations were made across the street to the southwest from Hsieh, Tien-hua's drugstore. Test pit A-B was on the level ground, pits C and D were on the t0p of a 2 m mound. The earliest date for the site is 2197 i 66 BP (NTU 196) from test pit D level 3, 70 cm from the surface. This date is somewhat out of sequence with two other dates from test pit C immediately to the east that came from levels # and 5 at 125-150 cm and 150-180 cm, reSpectively. The level n date is 1846 i 55 BP (NTU 194) and the level 5 date is 2104 i 63 BP (NTU 195). No particular distur- bance was noted in direct association with any of the samples, although ants were noted as far down as 110 cm and roots at 115 cm. Perhaps the D3 sample was disturbed by the later inhabitants of the site. The date itself seems to fit fairly well with the others for the site in general, but it seems to be out of sequence for that profile. Two additional dates were derived from a shallow test pit (PL ZlA-B) west of the mound. The date from the east end of the trench was 1783 i 53 BP (NTU 192); the date from the west end was 1837 i 55 BP (NTU 193). Previous to 1972 the remains from Ta-ma-lin have been dated by comparisons with the site of Ying-p'u #0 km 93 downstream, which yields similar pottery. The dates from Ying-p'u are 2970 1' 80 BP, 2810 1' BP and 2250 i 60 BP (CIhang 1969.206). The dates recovered in 1972-73 from flki-ma-lin are later in time than those from Ying-p'u. It unist be remembered, however, that these dates did not come from the earliest levels, suggesting an earlier beginning -oe.1415 I... . .0. I. u . . u...ll...h ‘. ...-J ...: a... .J . PL 21 0-3 PL’21~o-3 PL 21 0-3 Id. . (I la Wank .bauM mflk.w “fine“? .a.. ...?!th IV".{ (-f PL 21 0-3 mmgszr..c. .....turx. ,Muwcmmnv ...... PL 21 0-3 a '1‘“. o.“ ~ PL 21 0—3 PL 21 043 o a» . 4| .4... ..zzzus «£22223» ~az2223’ PL 21 o_3 Fig. 38 uzaa . Examples of triangular, tanged and serrated arrow polnts. 161 ... 0....”fi.kflu.fl Won”! ‘rdf ....«1..1.I.u......h ..1.\.. ...-NW. Y4...1JW.4.1NJ\°..I~ Jh-NI- .. LL. r... |w\ <22» PL 21 0-2 PL 21 0-3 PL 21 C-1 PL 21 0-2 .4! lib—{r 42 PL 29 8 PL 21 C-4 -6 PL 29 B-3 PL 21 C 1-1 Scales I .l .. a .I |l .. o . ....1\ (0.1... 4 ......fl. .. . 1 I . . .. I ... - . .. .. ........ “.... .6. ..rwwflv, . . .1”... ./..n... (so-\vl FJ./n.f1..\d... Ell 111‘.Q.Lu*...~. . . W‘%0! 11' a I Ns—DOV-Iu.v.0.'¢01.o. 0.0-0 1 . 311232....» (.- .. -@ W‘ W PL 21 C-3 PL 21 0-2 PL 21 0—} Examples of willow leaf and Triangular shaped arrow poinfs. Fig. 39 PL 21 C-2 162 Scale: 1-1 ..".:‘goo'-10 0» 1' i Q I . . - fi". 1 manta... 3". ’ - ‘k'°"."!.£ 03 ‘\ ' _ ‘. °.'; . . %.;m."’ .1 u v I 0‘ . ‘11, ‘52 . .‘I ~ J PL 21 c_4 PL 21 A_2 ‘.,.:..,,,,..,,. S '. 25.11211: a: 19.-{- M PL 28 TT-1 Fig. h0* Examples of recfangular knives. Scale: l-l PL 38 TT-1 PL 28 s—2 ' PL 21 0-5 PL 21 0-2 Fig. #1 Examples of stone artifacts: knives - A - rectangular; B, C - crescent shaped; D - semi-lunar: unknown - E, F - boot-shaped objects. Appendix C: Summary of Sites This appendix brings together in one place the summary of data for all of the sites located during the survey. For convenience in retrieval of information, each site was given a numeric designation as it was located. A file was kept on each site that included the initial survey sheet, maps, photos, drawings of test pit profiles, arti- facts, features and any additional information. All of the stone artifacts were drawn and photographed. Drawings and photographs were filed within the folder for the site from which they came. This information is available from the author for the use of future researchers. In the fifty-year interval between 1895 and 1945 when Taiwan was part of the Japanese Empire, several Japanese officials made collections of prehistoric arti- facts. Liu (1956:55-57) lists nineteen villages and neighborhoods which were reported by the Japanese officials and local collectors to have produced artifacts. An attempt to locate sites in these designated areas was disappointing , and only three sites previously mentioned could be relocated. Therefore, sites listed by Liu (1956:55-57) but not relo- cated in 1972-73 kept Liu's original numbers (i.e., 1-l9), while new sites were assigned numbers 20-u6 as they were located. 16h 165 A site, for purposes of the survey, was defined as any location where evidence of prehistoric human activity could be found. Evidence for activities of prehistoric peoples was usually seen by the presence of stone tools, ceramics, worked slate fragments or Specific features--most often stone burial cists. Sites were recorded in the notes with an arabic numeral preceded by a PL for P'uli. The following abbreviations were also used in the notes and to mark the artifacts: S = surface collection (may be divided into S-l, S-Z, S-3, etc. to designate subdivisions over a large area): TT = test trench; A, B. C = labels for test pits: Roman numerals were used to designate stratigraphic levels in the test pits, with l closest to the surface. P.L. Sites located in the 1972-73 Survey 2 Sampling Zone: 4 Type of collection: 100% surface collection and one test pit Ceramics: sherd count = 7: types represented = one Lithics: total number = 33: types represented = trapezoid, Spatula, rectangular, ovaloid hoes: sandstone fragments: grooved and notched net sinkers: rectangular reaping knife and a grooved pebble. Other features: small burial cist Time periods represented: Shuiwaku Period Possible function: small hamlet? field? P.L. 12 Sampling Zone: 3 Type of collection: 50% surface collection and five test pits (A-E) Ceramics: sherd count Lithics: total number 161: types represented rectangular, ovaloid and waisted hoes: adzes: grooved net sinkers: cut slate fragments: one grooved pebble and 136 sandstone fragments. 166 915: types represented = six u, v If 167 Other features: small square burial cist, a small metal pipe (Chinese trade ware) Time periods represented: Modern Period Possible function: small hamlet, fields P.L. 21 Sampling Zone: 4 Type of collection: 100% surface collection, test pits Ceramics: sherd count = 4,688: types represented = ten Lithics: total number 185: types represented = quartz flakes: awls: trapezoid, spatula, rectangular, ovaloid and waisted hoes: sandstone wasteflakes and fragments: slate and sandstone k'o halberds: chisel: adzes: grooved and notched net sinkers: slate fragments and rectangular reaping knives: projectile points: disc shaped waste plugs, boot- shaped artifacts and slate ornaments. Other features: clay spindle whorls Time periods represented: Shuiwaku and Tamalin Periods Possible function: village site *fiw - P.L. 22 Sampling Zone: 4 Type of collection: 100% surface collection, examined native excavations Ceramics: sherd count = 0, types represented = 0 168 Lithics: total number = 8, types represented = Spatula and rectangular hoes, k'o halberds and adzes. Other features: rectangular burial cist Time periods represented: Shuiwaku or Tamalin Periods Possible function: field? small hamlet? P.L. 22 Sampling Zone: 5 Type of collection: surface collection, interviewed .15 local inhabitants about collections they had made Ceramics: sherd count = 0, types represented = 0 Lithics: total number = 4, types represented = adzes, semi-lunar reaping knife and disc-Shaped waste plug Other features: rectangular burial cist Time periods represented: ? Possible function: ? 24 P.L. Sampling Zone: 1 _"’ WT ‘ Type of collection: 100% surface collection Ceramics: sherd count = 0, types represented = 0 Lithics: total number = 10, types represented = rectangular hoe, sandstone fragments, adzes, notched net sinkers and reaping knives Other features: none Time periods represented: ? Possible function: field 169 P.L. 25 Sampling Zone: 3 Type of collection: 100% surface collection, test pit two Ceramics: sherd count 77. types represented Lithics: total number 71, types represented Spatula, rectangular and ovaloid hoes: sandstone fragments: adzes: notched net Sinkers: slate fragments and a rectangular reaping knife Other features: none Time periods represented: Shuiwaku Period Possible function: small hamlet? field? P.L. 26 Sampling Zone: 4 Type of collection: 100% surface collection Ceramics: sherd count 0, types represented = 0 Lithics: total number = 16, types represented = rectangular and ovaloid hoes: sandstone fragments: slate k'o halberd: notched net sinkers and semi- lunar reaping knife 9 Other features: none Time periods represented: ? Possible function: field? 170 P.L. 22 Sampling Zone: 3 Type of collection: 100% surface collection H O Ceramics: sherd count = 0, types represented Lithics: total number = 4, types represented trapezoid and ovaloid hoes and slate fragment Other features: none Time periods represented: ? Possible function: field? P.L. 28 Sampling Zone: 3 Type of collection: 100% surface collection, test pit Ceramics: sherd count 65, types represented = 4 Lithics: total number 151, types represented = trapezoid. rectangular, ovaloid, pointed and irregular hoes: sandstone fragments: slate k'o halberds: adzes: notched net Sinkers: slate fragments: reaping knives and two pointed fragments Other features: none 4 Time periods represented: Shuiwaku Period Possible function: small hamlet : P.L. 22 Sampling Zone: 4 Type of collection: representative surface collection (c. 50%), test pits 171 1,572, types represented = 10 Ceramics: sherd count 251, types represented = Lithics: total number spatula, rectangular, ovaloid, waisted, pointed and irregular hoes: sandstone waste flakes and fragments: slate k'o halberds: a chisel: adzes: grooved net Sinkers (one notched sinker): slate fragments: reaping knives: projectile point frag- ments and a grooved pebble Other features: Burial cists had been found by farmers but were not in evidence in 1972-73. Time periods represented: Shuiwaku Period Possible function: hamlet or village, fields? Sampling Zone: 4 Type of collection: Test pits Ceramics: sherd count 242, types represented = 2 Lithics: total number 45, types represented = Spatula and rectangular hoes, sandstone waste flakes and fragments, Slate k'o halberd, adze, two grooved and one notched net Sinker and Slate fragments Other features: five rectangular burial cists, two areas with horizontal Slate slabs and one intrusive pit Time periods represented: Shuiwaku Period Possible function: hamlet? 172 P.L. 21 Sampling Zone: 4 Type of collection: 100% surface collection Ceramics: sherd count 0, types represented = O Lithics: total number 10, types represented = trapezoid, ovaloid and irregular hoes: one chipped rectangular slate k'o halberd and net sinkers--one each, grooved and notched Other features: rectangular burial cist Time periods represented: ? Possible function: small hamlet? field P.L. :2 Sampling Zone: 3 Type of collection: surface examination Ceramics: sherd count = 0, types represented = O O Lithics: total number 0, types represented Other features: A stone burial cist had been destroyed by modern farmers and the larger Slate fragments reused in the construction of a water cistern. r Time period: ? Possible function: burial P.L. 33 Sampling Zone: 3 Type of collection: 100% surface collection Ceramics: sherd count = 3, types represented = l 173 Lithics: total number = 0, types = O Other features: none Time periods represented: Modern Period? Possible function: field? path? P.L. 34 Sampling Zone: 4 Type of collection: 1.00% surface collection, test pit Ceramics: sherd count = 8, types represented l Lithics: total number = ll, types represented Spatula, hoes, sandstone fragments, Slate k'o halberd, grooved net sinkers and slate fragments Other features: a rectangular burial cist Time periods represented: Modern Period? Possible function: burial? field? P.L. 25 Sampling Zone: 4 Type of collection: 100% surface collection Ceramics: sherd count = 0, types represented = O Lithics: total number = 4, types represented rectangular and ovaloid hoes, sandstone fragment and a notched net Sinker Other features: none Time periods represented: ? Possible function: field? P.L. 6 174 Sampling Zone: 1‘ Type of collection: 100% surface collection Ceramics: Lithics: 0 sherd count 0, types represented total number = 0, types represented = 0 Other features: Although no ceramics or tools were found, a large quantity of slate fragments, which are not found naturally in this area, were located. Time periods represented: Modern Period? Possible function: building site? P.L. 37 Sampling Zone: 1 Type of collection: 100% surface collection, test pit Ceramics: Lithics: ll 0 sherd count = 0, types represented 2, types represented total number ovaloid hoes Other features: One short rectangular burial cist and one dish-Shaped depression Time periods represented: Modern Period Possible function: small hamlet? P.L. 88 Sampling Zone: 1 Type of collection: 100% surface collection, teSt pits Ceramics: sherd count = 138, types represented = l V q“ 175 Lithics: total number = 36, types represented = trapezoid, rectangular. ovaloid, pointed and irregular hoes: sandstone fragments: slate k'o halberds: one grooved net sinker and a slate fragment Other features: several rectangular burial cists had been destroyed by the present owner Time periods represented: Shuiwaku Period Possible function: hamlet, fields P.L. 32 Sampling Zone: 1 Type of collection: 100% surface collection, test pit Ceramics: sherd count = 3, types represented = 1 Lithics: total number = 13, types represented = trapezoid, rectangular and ovaloid hoes: sandstone fragments: grooved net sinker and two disc-Shaped waste plugs Other features: none Time periods represented: Shuiwaku Period Possible function: field, small hamlet? P.L. 40 Sampling Zone: 1 Type of collection: 100% surface collection Ceramics: sherd count = 1, types represented = l WWR‘JL-r- tan- ...- .- 176 Lithics: total number * unknown. This collection was made by Mr. Hwang and Mr. Liu, and the artifacts were returned to the Geology Department at National Taiwan University. The author did examine the Site and the collection, and it contains hoes, adzes, a small Spear point and slate fragments. Other features: none Time periods represented: ? Possible function: field? P.L. 41 Sampling Zone: 3 Type of collection: 100% surface collection Ceramics: sherd count = 4, types represented = 1 Lithics: total number = 17, types represented = spatula, rectangular and ovaloid hoes: sandstone fragments plus grooved and notched net sinkers Other features: none Time periods represented: Modern Period? Possible function: field? P.L. 42 Sampling Zone: 4 Type of collection: 100% surface collection Ceramics: sherd count 0, types represented Lithics: total number = 9, types represented rectangular and ovaloid hoes and sandstone = 0 fragments ! a 177 Other features: rectangular burial cists Time periods represented: ? Possible function: hamlet? burial P.L. 43 Sampling Zone: 4 Type of collection: 100% surface collection, test pit l Ceramics: sherd count 19, types represented Lithics: total number = 22, types represented trapezoid, rectangular:irregular,spatula and ovaloid hoes and sandstone fragments Other features: a rectangular burial cist Time periods represented: Modern Period? Possible function: burial, field P.L. 44 Sampling Zone: 3 Type of collection: 100% surface collection Ceramics: sherd count 0, types represented = 0 Lithics: total number = 16, types represented = trapezoid, rectangular, ovaloid and waisted hoes and sandstone fragments Other features: none Time periods represented: ? Possible function: field? I l 178 P.L. 45 Sampling Zone: 6 Type of collection: 100% surface collection, test pit Ceramics: sherd count = 4, types represented = 2 Lithics: total number = 0, types represented = 0 Other features: a square burial cist Time periods represented: Modern Period Possible function: burial P.L. 46 Sampling Zone: 5 Type of collection: 100% surface collection II 0 Ceramics: sherd count = 0, types represented Lithics: total number = 2, types represented rectangular hoe and slate fragment Other features: none Time periods represented: ? Possible function: field? path? Brown, James A. 1966 Chai, C. K. 1967 Chang, C. Y. 1953 Bibliography Dimensions of status in the burials at Spiro. Paper prepared for the 65th annual meeting of the American AnthrOpological Association, Pittsburgh. Taiwan Aborigines. Harvard University Press, Cambridge. Outline History of Taiwan. Taipei. Chang, Kwa ~chih 19 1956 1959 1962 1964A 1964B 1964C Prehistoric Remains in Taiwan. Taiwan fen -tu, Supplement to Kung-lun-pao, 161- 168, Taipei. A Brief Survey of the Archaeology of Formosa. Southwestern Journal of Anthropology 12: 371’386 o A Working Hypothesis for the Early Cultural History of South China. Bulletin of the Institute of Ethnology, Academia Sinica 7: 43—103. Major Problems in the Culture History of Southeast Asia. Bulletin of thejlnstitute of Ethnology, Academia Sinica 13: 1:23. Introduction: Special Taiwan Section. Asian Perspectives 7: l95-202. Prehistoric Ceramic Horizons in Southeastern China and their Extension into Formosa. Asian Perspectives 7: 243-250. Prehistoric and Early Historic Cultural Horizons and Traditions in South China. Current Anthropology 5: 359, 368-375. 179 'i‘m 180 Chang, Kwang-chih (continued) 1968A The Archaeology of Ancient China. Yale, New Haven. 1968B Archaeology of Ancient China. Science 162: 519-526. 1969 Fengpitou, Tapenkeng, and the Prehistory of Taiwan. Yale University Publications’in AnthrOpology. No. 73. New Haven. 1970A The beginning of A riculture in the Far East. Antiquity 4: 175-185. 1970B Prehistoric Archaeology of Taiwan. Asian Perspectives 13: 59-77. 1970, MS National Science Foundation Research Proposal. 1971, MS Anthropological and Environmental Investigations of the Choshui and Tatu River Valleys in Central Taiwan. National Science Foundation Research PrOposal. 1974 Man and Land in Central Taiwan: The First Two Years of an Interdisciplinary Project. Journal of Field Archaeology. Vol. 1, No. 3/‘h Chang, K. C. and Collaborators 1974 Man in the Choshui and Tatu River Valleys in Central Taiwan: Preliminary Report of an Interdisciplinary Project, 1972-1973 Season. Asian Perspectives. Vol. XVII (1). Chang, Kwang-chih and Minze Struiver 1966 Proceedings of the National Academy of Science. 55: 539-542- Chartkoff, Joseph L. 1975 Great Questions and Opaque Hypotheses: Testing Inadequate Alternate EXplanations for the Origin of Food Production. Paper presented to the Society for American ArchaeologY. Dallas, Texas. Ch'en, Che -hsiang 196 The Geography of the Mountainous Region of Taiwan. Economy of the Mountainous Region of Taiwan 109-136. 181 Ch'en, Johnson T. F. 1969 A Synopsis of Vertebrates of Taiwan. Taipei. Chen, Chi-lu 1958 A Cultural Configuration of the Island of Formosa. Bulletin of the Ethnological Society of Cfiina 2: 1-14. 1968 Material Culture of_the Formosan Aborigines. The Taiwan Museum, Taipei. Chung-kuo K'o-hsfieh Yflan K'ao-ku Yen-Shiu Suo Shih-yen-Shih 1972 A Fang-she-hsing t'an su ts'e-ting nien-tai pao-kao (l) K'ao-ku 1: 52-56. 1972 B Fang-she-hsing t'an su ts'e-ting nien-tai pao-kao (2) K'ao-ku 5: 56-58. Chung, Thein-fook and Tseng-chieng Huang 1972 Paleoecological Stud of Taipei Basin. Taiwania 17: 117-1 1. Chung, Thien-fook, Tseng-Chien Huang and Richard B. Stamps 1973 Paleoecological Study of Taiwan (3) The P'uli Basin. Taiwania 18: 179-193. Davidson, James W. 1903 The Island of Formosa. Ferrell, Raleigh 66 19 The Formosan Tribes: A Preliminary Linguistic, Archaeological and Cultural Synthesis. Bulletin of the Institute of Ethnology, Academia Sinica 21: 97-130. 1969 Taiwan Aboriginal Grou S: Problems in CulturaI and Linguis 1c 8881 1ca ion. Institute of Ethnology Academia Sinica Monograph No. 17, Nankang. Gorman, Chester F. 1969 Hoabinhien: A Pebble-tool Complex with Plant Associations in Southeast Asia. Science 163: 671-673. 1970 Excavations at Spirit Cave, North Thailand. Asian PerSpectives. Vol. XIII. I ‘ 182 Heine-Geldern, Robert von 1932 Urheimat und Frfiheste Wanderungen der Austronesier. Anthrgpos XXVII: 543-619. Hsieh, Chiao-min 1964 Taiwan-Ilha Formosa, A Geography in Perspec- tive. Washington. Kamaki, Yoshimasa, editor 1965 Nihon no Kokogahu, II Jomon Jidai. Tokyo, KaWde o Kanaseki, Takeo 1943 Taiwan Senshijidai ni Okeru Hokuho Bunka no Eikyo. Taiwan Bunka Ronso 1 Taihoku. Kanaseki, Takeo and Kobubu Naoichi 1949 T'ai-chung Hsien Ying-p'u Yi-chih Tiao-ch'a Yd Pao. Taiwan Wen-hua 5: 29-34. 1950 T'ai-wan K'ao-Ku-hsueh Yen-chiu Chien Shih. T'ai-wan Wen-hua 6: 9-15. 1953 On Recent Researches in the Prehistoric Archaeology of Formosa. The Ja anese Journal of Ethnology 18( : 7- O. Kano, Tadao 1952 The Place of Formosa in the Prehistory of Southeast Asia. Idem 2: 89-186. Kano, Tadao (Sung Wen-hsfin, translator) 1955 An Outline Review of Taiwan Archaeology. Provincial CommiSsion f3? Historic Research, Taipei o Kokubu, Naoichi 1943 Yuken sekifu Yudan sekifu o obi Kokuto-bunka. Taiwan Bunka Ronso 1: 17- 5. 1956 Note on the Burial Customs in Prehistoric Formosa. Proceediggs of the FourthFar Eastern Preh sto Con ress Part I: Second Fascicle, Section I, 308-318, Quezon City. 1959 Stone Knives of Prehistoric Formosa--on the Japanese 'isibo chyo,‘ stone Sickle, and stone knife with a handle. The Japanese Journal of Ethnology 23 (4): 1-38. 183 Kotani, Yoshinoku 1969 Upper Pleistocene and Holocene Environ- mental Conditions in Japan. Arctic Anthropology V (2): 133-158. Lin, Ch'ao-ch'i 1964 1966 Geology and Ecology of Taiwan Prehistory. Asian Perspectives 7: 203-213. An Outline of Taiwan's Quaternary Geo- history with Special Discussion of the Relation between Natural History and Cultural History in Taiwan. Bulletin of the Department of Archaeologyand AnthrOpology N.T.U. 28: 7-44. Ling, Shun-sheng 1956 1958 1959 Liu, Chih-wan 1956 1956 1960 Liu, Pin-hsiung 1954 1955 Formosan Sea-going Raft and Its Origin in Ancient China. Bulletin of the Institute of Ethnology, Academia Sinica 1: 81-54. A Comparative Study of Kava-Drinking in Pacific Regions. Bulletin of the Institute of Ethnology, Academia Sinica 5: 45-86. Origin of the Ancestral Temple in China. Bulletin of the Institute of Ethnology, Academia SiniEa 7: 141-185. An Outline of the Archaeology of Nan-t'ou. Nan-t'ou Wen-hsien Ts'ung-chi 4: 7-89. The Prehistoric Site on She-liao Terrace on the Southern Bank of the Cho-Shui River. Nan-t'ou Wen-hsien Ts'ung-chi 4: 91: 91- 108. Report of the Investigations at the Chfin-kung- liao Site in Nan-t'ou County. Taiwan Wen-hsien 11 (3): 1-67. A Preliminary Report on the Excavation at the Tung-chiao Site. Nan-t'ou Wen-hsien 1: 3-4. T'ai-chung Hsien Ch'ing-shui Chen Niu-ma-t'ou Yi-chih Tiao-ch'a Pao-kao. Taiwan Wen-hsien 63 69-830 184 Liu, Pin-hsiung (continued) 1964 Excavations and Discoveries at Ta-p'en-k'eng and Other Prehistoric Sites of Pali Dis- trict. Asian Perspectives 7: 214-223, Liu, Pin-hsiung and Liu, Chih-wan 1957 Report on the Archaeological Work at Jih- yueh T' an, or Sun-Moon Lake. Nan-t' ou Wen-hsien TS' ung-chi 5: 64. Liu, T'ang-jui and Liu, Chih-wan 1959 Botany. Nan-T'ou Wen-hsien Ts'ung-chi. VOlo 30 Nan'T .Quo Miyamoto, Nobuto 1956 A Study of the Relation Between the Existing Formosan Aborigines and Stone Age Remains ‘in Formosa. Proceedings of the Fourth Far Eastern Prehistory Congress Part 1, Second Fascicle, Sec. 1, 329-334. Pearson, Richard 1972 Archaeological Survey in Southeastern Taiwan. Bulletin of the Institute:g£ Ethnology, Academia Sinica 30: 317-330. Radiocarbon 1973 NTU-125 Chengpin 4. Radiocarbon 15: 2. Sauer, C. O. 1952 Agricultural Origins and Dispersals. Bowman Memorial Lectures, Series 2, AmeriEan Geo- graphic Society, New York. Saxe, Arthur A. 1968 EXplaining variability in disposal of the dead, social relationship factors. Paper presented at the 1966 meeting of the Society for American Archaeology, Santa Fe. Shih, Chang-ju 1953 T'aiwan Ta-ma-lin Yi-chih Fa-chfleh Chien Pao (A Brief Re ort on the Findings from Ta-ma- lin, Taiwan . Bulletin of the Department of Archaeology andpAnthropology No. 1, Taipei. 1956 Prehistoric Cultures in Nan-t’ou. Nan-t'ou Wen-hsien Ts'ung-chi 4: 1-6. 185 Shih, Chang-ju and Sung Wen-hsfln 1956 Minor Excavations at T'ieh-chan-Shan. Bulletin of the Department of Archaeology and AnthroEology 8: 35;50. Solheim II, W. G. 1968 Early Bronze in Northeastern Thailand. Current Anthropology 9: 59-62. Stamps, Richard B. 1972, MS Preliminary Notes on the P'uli Basin Archaeological Survey. Paper presented to the Choshui and Tatu River Valley Project Interim Meeting, Taipei. 1975, MS An Archaeological Survey of the P'uli Basin, West Central Taiwan. Report sub- mitted to the Choshui and Tatu River Valley Project. Sung, Wen-hsfin 1962 T'ai-chung-hsien Fan-tzu-yflan Pei-chung Chih.Mu-tsang. Bulletin of the Department of Archaeology and AnthrOpoiogy 19/20: $90. 1965 T'aiwan HSi-pu Shih-ch'ien Wen-hua te Nien-Tai. Taiwan Wen-hsien 16: 144-155. 1964 The Stone, Bone, and Horn Industries of the Yfian-shan Shellmound. Bulletin of the China Council for East Asian Stadies 3: 98-99. Sung, Wen-hsfln and Chang Kwang-chih 1954 Digging of Prehistoric Sites by the Banks of the Shu-wei River, Tai-chung Prefecture. Bulletin of the De artment of Archaeology and Anthropology 3: 26-58. Tomita, Yoshiro 1951 Physio raphic Development of the Hori P'uli Basin Group of Central Taiwan Formosa). Annals of the Tohobu Geo- gpaphical Association 3: -7. Torii, Ryuzo 1897 Enzan Kaizuka ni Kansuru Tsushin. Jinruizaku Zasshi 13: 116-118. Treistman, Judith M. 1970 Problems in Contemporary Asian Archaeology. Journal of Asian Studies XXIX: 363-371. 1972 Prehistory of the Formosan Uplands. Science 175. 186 Tsuboi, Kiyotari 1956 Fengpit'ou--A Prehistoric Site in Southern Formosa that Yielded Painted and Black Pottery. Proceedings of the Fourth Far Eastern Prehistory and Anthropological Division of the Eighth Pacific Science Congresses Combined Part 1, Fascicle 2, Sec. 1 277-302. Tsukada, Matsuo 1966 Late Pleistocene Vegetation and Climate in Taiwan. Proceedings of the National Academy of Sciences 55: 1967 Vegetation in SubtrOpical Formosa During the Pleistocene Glaciations and the Holcene. Palaeogeographyp Palaeoclimatology, Palaeoecology. Watson, Patty Jo and Steven A. LeBlanc, Charles L. Redman 1971 Explanation in Archaeology: An Explicitly Scientific Approach. Columbia University Press, New York. Watson, William 1971 Cultural Frontiers in Ancient China. University Press, Edinburgh. Fig. 3 Sites located in the P'uli Basin Archaeological Survey. . o = Modern Period D = Tamalin Period A = Shuiwaku Period . = Unknown Period .- ,_ ..__.,,.___,._, h. l 1 l 5 _: 6 175 32