ANTHROPOLOGICAL RECORDS 18:2 ARCHAEOLOGICAL EXCAVATIONS IN YAP BY E. W. and D. S. GIFFORD UNIVERSITY OF CALIFORNIA PRESS BERKELEY AND LOS ANGELES 1959 ARCHAEOLOGICAL EXCAVATIONS IN YAP BY E. W. AND D. S. GIFFORD ANTHROPOLOGICAL RECORDS Vol. i8, No. 2 UNIVERSITY OF CALIFORNIA ANTHROPOLOGICAL RECORDS Editors (Berkeley): J. H. Rowe, R. F. Heizer, R. F. Murphy, E. Norbeck Volume 18, No. 2, pp. 149-224, plates 24-41, 5 figures in text, map Submitted by editors July 30, 1958 Issued February 15, 1960 Price, $1.50 University of California Press Berkeley and Los Angeles California Cambridge University Press London, England Manufactured in the United States of America CONTENTS Page Introduction ...........................,.,,................. 149 Archaeological sites .................. , , . ..... 152 Walgom .......................... 152 Ruuway ..... , , , , , , . ......... 154 Penin ............................................ , . . 155 Boldanig and Wolom .,,,,, 156 Pemrang .................1..................................... , 157 Composition of sites ............ ........... 159 Zoological specimens ...................,,,,,............. 162 Unworked bone ............. 162 Decapods .......... 164 Echinoderms ........... , , , ... 164 Mollusks ....... 164 Coral ........... , ,.. 175 Artifacts ....,,,,,,,,,,,,,,,,,,,,,................. 179 Pottery .................,,,,,,,,,....................... 179 Laminated ware .............. 180 Unlaminated ware ......................................,,...... . 181 Potsherd rims and lips ........ 181 Drilled potsherds ............... ...,,,... 183 Gambreled shoulders .................. 184 Incised ware .184 Painted ware ........... , , , , . ...... . 184 Petrography ....................................................... 184 Shell artifacts ............... 185 Adzes .185 Straight knives ............... 188 Taro peelers .190 Paring knives ............ . . . ......,,... 190 Scrapers ................................. 191 Bracelets and rings ...........,. , , . .... 191 Disks ........................... 191 Disk beads ................... .. . . . .. 191 Perforated univalves ..................,,,,,... 192 Perforated bivalves ..................,.,.,.. 192 Conus cap ........ , , . ............,,..... 192 Pestles ...... , ...........,., . 192 Pottery smoothing tools ............................................ 192 Money . 193 Trumpets ......... 193 Bone artifact ............,.,,............,,.... 193 Stone artifacts ........................ 193 Adze ......,......... 193 Disk ...........................,. , . ........... 193 Stone money ..... .. . .. . .. .. 193 "Toy stone money" ..... . . . . . . .. . 194 Whetstone, pestle, and hammerstones ............................... 194 Quartz strike-a-light .194 Mineral pigments .............................,.,.,.,.,.194 Radiocarbon dates . .,.,.,.195 Local distributions in space and time .196 Foreign cultural relationships .200 Literature cited .201 Explanation of plates .204 MAP Villages of Yap ............ facing page 149 ULUOL\ ~ . f RUNAUNG R MECUCOL ~~FA /,ORU ECIJEL 2 > { " /N~~~~UJ.UL _ t t w S ) $ "@ /~~~~~~UNCD 'I FAN IF~~~~~~~~~~~~~~~~~~I DALIFtBINAU tU ouoD ) KANIFAY>t,^ / MAAM.A"F 0 0 m GALIMAN ,'ANo~~~~~~~~~~~~~~~~~,MLAA ,GORRFR) ^, . DUL UKAt / ISCTHAL A P P RORE 20N70.5 r y v / CH@EFS OTHER D4STRICT sSsS~~~~~~~~~~~~~~~~~~~~~~~~(;4'0 z p^iat^#K106 / / ARE 14 PARENTHESILANG I ~ ~ ~~~~~~~~~~~~~~~~~~~~~~ I U ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ MU t :z~~~~~~~~~~~~~~~UU ARCHAEOLOGICAL EXCAVATIONS IN YAP BY E. W. AND D. S. GIFFORD INTRODUCTION The position of Yap near the western edge of Micro- sia, in the longitude of Tokyo and about 750 miles st of the northern tip of Mindanao, makes it a criti- Lgroup for study in reconstructing the prehistory outer Oceania, for it lies in the path of the east- d movement of peoples from Malaysia into Micro- ia and Polynesia. For this reason, it was selected the third region for archaeological excavations in e University's program of investigations in Oceanian ehistory. The first region was Viti Levu, Fiji (1947), ! second New Caledonia (1952). The time spent in was January 27 to June 2, 1956. The authors of report, Edward W. Gifford and Delila S. Gifford, ustituted the personnel of the expedition from the versity of California. Yapese workers assisted ectively in the field. The culture of Yap has produced a unique feature, stone cartwheel forms of aragonite and calcite ed for money, which have earned for this tiny chipelago of 39 square miles the appellation of land of Stone Money.' Outside Colonia, the administrative center of Yap, e island confirmed cultural impressions we had ined from the publications of forty, fifty, and sixty ars ago-from the writings of Kubary (1895), Chris- a (1899), Furness (1910), and Muller (1917, 1918). e visible or material culture is largely unaltered, manifested by the costumes of both sexes, the one money, the stone platforms, pavements, quays, d fish traps, the dwellings and men's clubhouses, the taboo places. ACKNOWLEDGMENTS Leave of absence from teaching duties, granted by Regents of the University of California, is hereby knowledged with appreciation. The financial support r our expedition came primarily from the Wenner- 'en Foundation for Anthropological Research, sup- mented by research grants from the University of fornia. Permission to make excavations in Yap kindly granted by Mr. Delmas H. Nucker, Acting Commissioner (now High Commissioner) of the st Territory of the Pacific Islands. In the field, three consecutive staff anthropologists the Trust Territory of the Pacific Islands made angements for the work in Yap: Professor Saul H. *senberg, Allan H. Smith, and John E. de Young. Yap itself, the district anthropologist, Mr. Shigeru eshiro, well liked by the Yapese, laid the ground- k for smooth operations by discussing the pro- sed excavations with the magistrates and chiefs. In Yap District, the consecutive district administra- tors, D. Donald Heron and Robert Halvorsen, made the necessary arrangements for housing, transporta- tion, and workmen. Mr. Halvorsen's staff effectively implemented his cooperation during the four months of field work. For land transportation Mr. Halvorsen arranged for the use of a Dodge power wagon owned by Mr. Alexander A. Tretnoff, who either drove it or provided a Yapese driver. Sea transportation was mostly supplied by the Yap Trading Company. We are especially indebted to Mr. Duane H. Kipp, at the time residing in Guam. He undertook to ship from Guam important items of equipment, thus re- lieving us of the necessity of sending these from Cali- fornia. This equipment included picks, shovels, ham- mers, hatchets, axes, sledge hammer, mattocks, boxes for packing specimens, nails, buckets, machetes, files, whetstone, and surveyor's stakes, all of which were waiting for us on our arrival. To the magistrates and chiefs of the five munici- palities in which systematic excavations were con- ducted we owe hearty thanks for their cooperation, which included guidance to the ancient sites. The res- pective magistrates were Ranganibay of Rull, Finiginam of Dalipebinau, Gilinug of Kanifay, Kenrod of Galiman (Giliman), and A. J. Roboman of Tomil. Francisco Luktun of Meloj (Weloi) also had us conducted to a small hill site, which we tested but did not dig sys- tematically. These gentlemen also helped us to obtain Yapese workers. As in Fiji and New Caledonia, the native workers became keenly interested in the exca- vations and carefully watched for pertinent materials, notably charcoal for radiocarbon dating. Joseph Sogon served as interpreter throughout our four months' work in Yap. Through the kind offices of Mr. Clayton W. Whitcomb, educational officer in Yap, two schoolteachers, Francis Nuuan and Vincent Unta- man, helped us during their two months' vacation. To various scholars we owe a debt of gratitude for the identification of specimens excavated in the several sites. David H. Johnson, United States National Museum: mammal. Dean Amadon, American Museum of Natural History; Alexander Wetmore, Smithsonian Institution: bird. Joseph R. Slevin and Alan E. Leviton, California Academy of Sciences: reptile. Henry W. Fowler, Academy of Natural Sciences, Philadelphia: fish. Tune Sakai, Yokohama National University, Kama- kura, Japan, and R. Imaizumi, Tohoku Univer- sity, Sandai, Japan: decapod. J. Wyatt Durham, University of California: echino- derm. [ 149 1 ANTHROPOLOGICAL RECORDS Leo G. Hertlein, California Academy of Sciences: mollusk. J. Wyatt Durham, University of California: coral. Howel Williams, University of California; C. G. Johnson, United States Geological Survey: stone. G. H. Curtis, University of California: potsherd constituents. Radiocarbon measurements of charcoal samples were made by Professor H. R. Crane, of the Harri- son M. Randall Laboratory of Physics, University of Michigan. Geological and geographical information has been kindly supplied by Dr. Charles G. Johnson, United States Geological Survey. Two pictures (pl. 26, a, b) were taken for us by Mr. Porter E. Ward, of the United States Geological Survey. We are obliged to Dr. Harold J. Coolidge, execu- tive director of the Pacific Science Board of the National Research Council, for aid in obtaining the necessary permit to enter Guam, issued by the Chief of Naval Operations, Washington, D. C. Our work in Yap had the approval of the Pacific Science Board. Upon our return to the University of California we received full aid and co8peration from the staff of the Museum of Anthropology in unpacking and cata- loguing the material shipped from Yap. The extensive collection, comprised of many thousands of individual sherds, shells, etc., was catalogued in 5,011 entries. Professor George M. Foster, Acting Director of the Museum, kindly assigned members of his staff, nota- bly Miss Rachel K. Eckman and Miss Shirley R. Gud- mundsen, to assist us. The Institute of Social Science of the University granted funds for the hiring of re- search assistants, Dr. Sheilagh T. Brooks, Dwight T. Wallace, and Sylvia M. Broadbent. Wallace assisted especially on pottery, Broadbent on shell artifacts, and Brooks on bone identification. Mrs. Mary Anne Whipple, editor of the publications of the Department of Anthropology, has much improved this paper by her ministrations. GEOGRAPHY Our good friend, Dr. Charles G. Johnson, supplied us on March 7, 1957, with the following memorandum (see map): There are no "official" figures for the area of Yap. I am using the area computed by E. H. Bryan of the Bishop Museum in 1946. He used a U. S. Hydrographic Chart for Yap in his computa- tion. His figures are in an unpublished report: E. H. Bryan, "A Geographic Summary of Micronesia and Notes on the Climate of Micronesia:' Vol. 2, pt. 1, U. S. Commercial Company, Economic Sur- vey, Honolulu, 1946. Bryan computed the total area of Yap, including the reef, to be 48.5 square miles, and the land area of Yap to be 38.5 square miles. Mr. Richard S. Baker, of the United States Weather Bureau, Yap, made available the records of rainfall at the Weather Bureau station in Colonia (lat. 9 31' N., long. 138' 8' E., ground elevation 53 ft.) for the five months we were in Yap: January, 7.78 in.; February, 3.63 in.; March, 8.73 in.; April, 18.15 in.; May, 12.17in, Colonia, also called Yaptown, situated on Tomil Harbor, has been the administrative center for the four foreign nations that have consecutively ruled Yap since 1884: Spain, Germany, Japan (for the League of Nationu and the United States of America (for the United Nation An American political innovation was the establish- ment of elected magistrates, usually district or village chiefs, for the nine districts, now called municipalities. Each municipality includes a number of villages or ham-| lets. There are 130 of these, as shown in the map in this paper. The author of this map is unknown, but the map is believed to be of Japanese origin. It is reproduced by permission of the Trust Territory administration. Be- fore we started on our expedition, Professor David M. Schneider, of the University of California Department of Anthropology, kindly gave us a photographic copy of it. For the purposes of this paper we have accepted its orthography, since no two ethnologists seem to agree just how the language of Yap should be written. On the map, j is used for i, as in Malaj, pronounced "Malai"; z is used for th, as in Anoz, pronounced "Anoth." The archipelago of Yap lies between longitude 138' 3' E and 1380 13' E and between latitude 90 26' N and 90 38' N. It originally comprised three main islan separated by narrow, shallow, mangrove-fringed chan- nels: Yap, the largest, Map, the second in size, and Runmung, the smallest and northernmost. In 1901, undet German rule, the excavation of the Tageren Canal through a narrow isthmus in the Tomil district of Yap Island created a fourth island, made up of the Gagil district and the major part of the Tomil district. It is now customary to disregard this artificial origin and speak of four main islands of the archipelago, the fourth being called Gagil-Tomil. The rolling hills of the archipelago, with their forests and grasslands, form one of its pleasant aspects. The highest hill, 585 ft., is on Yap Island. The following summary of the geology of the archi- pelago was kindly supplied by Dr. Charles G. Johnson, of the United States Geological Survey, who was work- ing in Yap during the early part of our stay. The basement rocks of Yap are metamorphic, dominantly chlorite and hornblende schists and massive amphibolite, intruded by irregular masses and dikes and sills of peridotite and gabbro which have been altered to serpentinite. The metamorphic rocks underlie most of Yap Island and Rumung and the western part of Gagil-Tomil. The massive ser-- pentine intrusions occupy eastern Gagil-Tomil at Gatjapar, Thol, and Maki, and numerous small silhi and dikes are scattered about on Yap Island. The basement rocks have been severely brecciated by thrust or normal faulting; a great mass of breccia underlies most of Map, southeastern Rumung, and eastern Gagil-Tomil. Conglomerate and sandstone, derived from erosion of the breccia, are intimately0 associated with the breccia in sheetlike deposits aWd channel fillings. Foraminifera found in the sandstOn channel fillings are Miocene (Tertiary fl) in age. Andesitic and basaltic volcanic breccia, tuff, and lava flows overlie the basement rock and breccia in southern and northeastern Yap Island, central Gagil-Tomil, and scattered patches on Map. The volcanic rocks are deeply weathered to red and 150 GIFFORD: EXCAVATIONS IN YAP yellow kaolinitic clay, which erodes easily to bare slopes and deep gullies typical of badland topogra- phy. Raised reef limestone exists only as small remnants, such as Garim Island and two small isolated mushroom rocks on the fringing reef that surrounds Yap. The limestone is considered to be of Pleistocene age. Terraces, reflecting several changes of sea level, are prominent on Yap; these are at levels approximately 10, 20, and 40 meters above sea level. There is no evidence of the 5-foot stand of the sea which is so well preserved in the Mariana Islands and elsewhere in the Pacific. A submerged terrace, at a depth of about 20 meters below sea level, is preserved along the outer reef margin near the entrance to Tomil Harbor. Additional sub- marine terraces are probably present but have not been observed. The sequence of changes in sea level is not known, but the net result has been a partial inundation of Yap, as is evident in the drowned valley that comprises Tomil Harbor and similar drowned valleys that are deep reentrants in Map and Gagil-Tomil. EXCAVATIONS each rectangle was dug. Rectangles not numbered were not excavated. Letters followed by numbers in the dia- grams indicate the corners of the rectangles, as marked on surveyor's stakes used for laying out the area to be excavated. All compass readings given in the text and in the diagrams are based on magnetic north. The workers screened 1,809 cubic feet of archaeo- logical deposit. The following tabulation gives the lowest depth of the excavation at each site, and the total number of cubic feet of deposit screened in each. Municipality, Village, and Site Tomil, Merur village, Walgom site Rull, Balebatt village, Ruuway site Dalipebinau, Kanif village, Penin site Kanifay, Malaj village, Boldanig-Wolom site Galiman, Giror village, Pemrang site Depth Cu. ft. (in.) screened 30 99 30 153 30 171 90 765 90 621 Systematic excavations were conducted in five of the municipalities of Yap: Tomil, Rull, Dalipebinau, Kanifay, and Galiman. The municipalities of Map, Rumung, Ueloj, Fanif, and Gagil were visited, but no promising sites were found, hence no excavations were undertaken in these places. The grid areas for our excavations were staked out in rectangles 6 ft. by 3 ft. The deposits were removed in 6-in. layers, each newly exposed surface being leveled. From certain pit walls earth samples were taken at 6-in. intervals, beginning at 3 in. be- low the surface, for analysis of the components at the various levels. All the excavated deposit was sifted through a screen of half-inch-square mesh, and also, whenever advisable, through a screen of quarter-inch mesh. These methods are the same as those employed in Fiji and New Caledonia on earlier expeditions, so results in the three regions are statis- tically comparable. In diagrams 1 to 5 the numbers entered in the rec- tangles indicate the lowest depth in inches to which Where excavation was on sloping ground, the actual amount screened is slightly in excess of the figures given, because the top triangle of material is in- cluded with the first level block of 0-6 in. depth. Each 6-in. block, 3 ft. by 6 ft. in horizontal dimen- sions, contains 9 cu. ft.; thus, for example, when a diagram indicates a depth of 30 in., it means that 45 cu. ft. of deposit were excavated and screened. Specimen numbers cited are in catalogue 11 (Oceania) of the Museum of Anthropology, University of Cali- fornia. In all but one of the tables depth is expressed by a single figure, which represents the lowest depth of a 6-in. block: 6 means surface to 6 in.; 90 means 84-90 in., and so on. Table 1, giving composition of sites, is an exception; in this, the single figure for depth is the depth at which the sample of midden material was taken. In most tables Wolom data are included under the heading "Boldanig," since the two spots are only 142 ft. apart. 151 ARCHAEOLOGICAL SITES Practically every Yapese village has a certain num- ber of unused or abandoned house sites, which are, in a sense, archaeological. The date of abandonment may be from one generation ago to many generations. Ex- cept for Ruuway, the five excavated sites were all within the limits of present-day villages. Ruuway, lying between the villages of Orowoo and Balebatt in the municipality of Rull, is called the chiefs' area, and is claimed by two rival chiefs. The claim has not been settled but both chiefs gave us permission to dig there. According to the natives, whoever lived at Ruuway automatically became chief of all Yap. Yap has a tremendous amount of stone work. Most houses and all men's clubhouses are erected on stone platforms. The banks of sloughs and creeks near set- tlements have retaining walls of stone. Stone cause- ways connect clubhouses on tiny islets with the main- land. Quays of stone, built out into water deep enough to allow canoes to lie alongside, are common along the shore lines. Coral heads are extensively used for house platforms, retaining walls, causeways, and piers. Sometimes the coral is trimmed with steel tools to brick shape (MUller, 1917, pl. 52, fig. 2), thus making more compact walls. Paths and plazas are paved with stone, usually with slabs of schist but slabs of iron-rock (probably limonite) or other stone were also used when available. Both Kubary and MiUer give full accounts of house construction. The following description and measure- ments of stone platforms and dwellings are by Kubary (1895, p.36). Die Wohnhauser liegen immer auf einer steiner- nen C6bok und haben die ubliche sechseckige Form. Die Grosse kann verschieden sein, die Dimensionen eines kleineren Wohnhauses betragen: Hohe des steinernen Fundamentes........ Lange desselben in der Mitte............ Linge desselben auf den Seiten.......... Breite desselben....................... Lange des Hauses ...................... Breite desselben....................... Wandh8he desselben.................... Ganze H8he desselben.................. 0.90 m. 18. 00 m. 13. 00 m. 7.25 m. 11.00 m. 5.50 m. 1.50 m. 6.00 m. As in Fiji and New Caledonia the soil of the upper layers of the sites excavated was black or grayish, but the color became lighter and more like the sur- rounding terrain as we dug deeper. At Walgom the underlying soil was reddish brown and clayey; at Ruu- way and Penin greenish gray, clayey, and stoney; at Boldanig-Wolom and Pemrang, sandy. The soil of Rugog's grave, like that at Walgom, was reddish brown. Only the Boldanig and near-by Wolom sites in Malaj village are on such low ground that water was encountered in their lower levels. As in similar sites in New Caledonia (Gifford and Shutler, 1956, p. 16) lime-cemented material was also encountered at levels seasonally reached by ground water. In the sandy sites of Boldanig and Pemrang we were plagued by the activities of land crabs and hermit crabs, which frequently knocked objects from the surface into the open pits or themselves fell into the trenches. We were not permitted to dig any cemeteries, since the chiefs as well as the people were strongly opposed to excavation there. We were invited, however, by Magistrate Roboman, of Tomil, to dig the grave of Rugog, a mythological chief who is supposed to have survived the flood that all but submerged Yap (Walleser, 1913, p.620). Roboman's interest was not only in the bones; he wished also to learn, through radiocarbon analysis, the date of the burial. Christian (1899, pp. 295-296, 302) describes graves on the northern end of Yap Island, which we did not visit. They resemble in style the one observed at the clay pit near Gitam, mentioned in our description of pottery making. Fragments of iron, glass, crockery, and other ob- jects of modern introduction were occasionally found on the surface and in the top layers of sites. WALGOM This site is in the village of Merur, Tomil munici- pality, on the inland edge of the shore and on an escarp- ment behind the narrow coastal plain. It is a mound, at about 15 or 20 ft. elevation, with a three-terrace house platform on it. The two lower terraces are rec- tangular, the top one hexagonal like other Yapese house platforms. There is no house there now, but the former dwelling, called Walgom, served as the retreat for the priest of the iguana god, Magrogoy. MUller (1917, pp. 320, 344) gives an account of this god (dlmon), whose name he spells "Magarogoi." On March 30 Magistrate Roboman, of Tomil, took us to this site. We were accompanied by three mem- bers of our regular crew, Joseph Sogon, Francis Nuuan, and Vincent Untaman. Pitug also went with us. He knew the site well, since, if he had not become a Roman Catholic, he would have been the god's priest. The site is at the intersection of two paved walks which, running north-south and east-west respectively, meet at a right angle. The mound on which the three terraces stand has a red, clayey soil. Its slopes con- tain many shells of mollusks, chiefly Chama, evidently thrown there after the priest had eaten the flesh. The priest's retreat lasts one hundred days, during which time the flesh of vertebrate fishes is taboo to him. The top surface of the hexagonal house platform is 57 in. above the level of the paved path that runs in an east-west direction 20 ft. south. A slender piece of schist, about 4 ft. long, lay on the middle terrace against the southeast side of the hexagonal platform, facing this path. More or less triangular in transverse cross section and tapering at both ends, it bore more than a hundred transverse grooves on one rounded angle (fig. 1). These grooves were said to be the tally of the days the priest spent in retreat. Excavations were carried on at this site from May [ 152 1 GIFFORD: EXCAVATIONS IN YAP 7 to May 14. The first rectangle (Al-2 Bl-2) was dug in the paved middle terrace just south of the hexagonal house platform. This terrace, referred to as a sitting platform, served as the front porch of the house, from which both of the intersecting paths were visible. The rectangle was staked in this middle terrace, the 6-ft. sides running north and south, the 3-ft. sides east and west. The paving stones of the terrace were temporar- ily removed. The rectangle was so unproductive-not a trace of sherd or shell-that we dug only six inches be- fore replacing the paving. On the paved surface just east of the grid were a dozen or so Hippopus valves. The second rectangle (Cl-2 Dl-2) was dug in the hexagonal house platform and was carried down to a depth of 30 in. The third rectangle (E1-2 Fl-2) was laid out on the sloping northern end of the mound and was dug down to 24 in. at datum stake F2. (See diag. 1.) From Walgom we visited five other sacred spots, including three places mentioned in the myth of the world flood. These five places are all within the vil- lage precincts of Teb and were later examined care- fully on May 15 and 16. Our first stop was at the place where the iguana god, Magrogoy, was fed, about a quarter-mile north of Walgom. It is a large flat slab of hard gray stone To Merur hi1 0 0 c 0 E Paved Pat h I mound slope 0 X. *0 U, a- a. mound slope 1 15" F2 Fl I, To Magrogoy s shrine in Teb Diagram 1. Plan of excavations in Walgom, Merur village, Tomil municipality. N Fig. 1. Day-count stone at Walgom. 153 ANTHROPOLOGICAL RECORDS (basalt ?), with diameters about 40 and 50 in. The edge is irregular, and the slab has been cracked by roots. Near by, a small stone pavement, with two leaning stones imnbedded in it as a back rest, indi- cates where the audience sat. We next visited Magrogoy's shrine, called Bulne- man, about 50 ft. east of the place where the god was fed. The long axis of the shrine runs east and west about 8 ft. Six feet beyond the east end a slab of stone money, about 3 ft. in diameter, has toppled over and leans at an angle of 15 degrees to the ground. The shrine formerly consisted of four low corner- stones supporting a thatched shelter, about 4 ft. wide, over a small canoe whose ends rested on two blocks of coral (p1. 24, a). The priest placed in the canoe the offerings handed to him-offerings from all over Yap. We were told that Japanese soldiers had re- moved the canoe and dumped the contents on the ground beneath. These offerings included two unmodi- fied Tridacna and Hippopus valves, which we did not collect. According to Roboman, there were also origi- nally nine Charonia shell trumpets. One of these, the best, he had taken, inserting in a hole drilled near the edge a string to suspend it by. He presented this shell to the expedition; it is shown in plate 36, a. Two of these Charonia shells, which were in good condition, we collected. Most of the offerings were pieces of pearl-shell (Pinctada margaritifera) money, drilled at the umbo. A number of undrilled valves were also found. We collected the specimens at the shrine on May 15, but did not excavate the site. Be- fore any of the offerings were removed, we cleared the vegetation from the shrine and took photographs. We then visited the places connected with the mythology of the "world flood": the grave of the Rat Woman, whose death is supposed to have caused the flood; the place of refuge from the rising waters; and last, the grave of Rugog, the Rat Woman's son-in- law, who survived the flood and might be called the Yapese Noah. The grave of the Rat Woman was a low mound covered. with small, flat stones. The flood myth re- lates how the chief, Rugog, set a trap in his field to catch a rat which had been eating his sugarcane. Once it was caught, his wife declared he had killed her mother, who could transform herself into rat form. As a consequence there followed a flood that submerged most of Yap. (A version recorded by Walleser, 1913, p. 620, relates that it was the chief's father-in-law who assumed rat form.) The place of refuge from the flood waters, called Cheen, is within the limits of Teb village, a short distance, perhaps 100 yds., from the Rat Woman's grave. Here Rugog took refuge with his wife and children. At Cheen there is a depressed area of probably 30 ft. by 100 ft. with a dozen heavy pillars of stone set in two rows about 20 or 30 ft. apart. Most of them were lying flat on the ground. The big- gest, which measured 9 ft. 7 in. from the ground to its top, was still standing, though leaning slightly (pl. 24, b). According to Mr. Shigeru Kaneshiro, district anthropologist of Yap, these stones are said to have been brought from Rumung Island. In front of the tall leaning pillar, a basalt grind- stone (diams. 374 in. and 29 in.) lay flat on the ground (pl. 24, c). The many broad grooves in its face were apparently caused by the grinding of adze blades. On February 22, we had seen a similar stone align- ment in the jungle near Gaychapar village, Gagil muni- cipality, in a paved area about 25 to 30 ft. long and 12 ft. wide. There a row of slab pillars, 6 ft. high, were set up, six on a side, along the long sides of the paved platform, with a short upright in the middle at each end of the platform. This was a taboo place, no longer in use. On May 16 we excavated the grave of Rugog (pl. 24, d), which is near the men's clubhouse in Teb, its axis lying east-west, with the head to the west. Excavations here were carried down 32 in., potsherds and charcoal being found down to 31 in. At 20 in. skull fragments were encountered; all other bones had crum- bled to meal. Slightly below the level of the bones we found sherds (wt. 27.5 oz.) of a flat-lipped, laminated pottery bowl of a late type, apparently buried in the grave. Sogon, one of our crew, explained that it was customary to bury a whole vessel at the head of a grave. Unlaminated, tempered body sherds (wt. 3 oz.) and charcoal were recovered at a depth of 31 in. and may be older than the burial, which would appear to be contemporary with the laminated pot. If so, the un- laminated potsherds are "antediluvian," the charcoal likewise. The charcoal, from a depth of 20 in., yields a radiocarbon date of A.D. 1756, presumably the date of the burial (see table 22). RUUWAY The Ruuway area is in dense jungle, at about 50 ft. elevation and a quarter-mile from the shore of Tomil harbor. To reach it one proceeds southward along the shore road after crossing the causeway leading from Yaptown, or Colonia, to Orowoo and Balebatt villages in Rull. Upon arriving at the Money Bank, a stretch of road bordered by Yapese stone money, one turns westward up a stone-paved path that ascends a gentle slope, passing dwellings, taro and yam beds, and clumps of bananas. The paving stones are mostly slabs of greenish gray schist, with an occasional slab of iron- stone. About a quarter-mile up the stone path one leavei the path and turns right into the jungle, reaching the site about 80 yds. beyond. On February 4 Chief Ruepong, one of the claimants of Ruuway, and Chief Moorou, representing the other claimant, Chief Ranganibay, took Mr. Kaneshiro and Fig. 2. Calendar stone on "spirit" mound in Ruuway. me to see several features in Ruuway west of the trail leading through the jungle in- to our site. Lying upslope from south to north, are, first, the platform of a for- mer men's clubhouse; second, an extensive, depressed danc- ing area; and third, a small rounded mound, 2-3 ft. high, called the "spirit" mound, which we were forbidden to dig. At the north, or upslope, edge of this mound is imbed- ded a vertical stone, 14 in. high, with many parallel, vertical grooves which were said to serve as mnemonics for raconteurs of mythi- cal events, the grooves at the right referring to the emergence of the southern part of Yap from the sea. This engraved "calendar" stone is shown in figure 2. At Ruuway we dug in three spots, Ruuway 1, 2, and 154 GIFFORD: EXCAVATIONS IN YAP 3. Ruuway 1 and 3 were in the chiefs' area. Ruuway 1 is a house platform said to have been abandoned before the advent of Europeans, that is, before 1884. Ruuway 2 is a small but extremely productive cook- house site within the limits of Balebatt village and just south of the Ruuway boundary. It belongs to Ramangan, a Balebatt man, who kindly allowed us to excavate it. he told us that the site was called Fite- awat, "fireplace"; awat means "charcoal." Ruuway 3, about 100 ft. slightly east of north of Ruuway 1, is the platform of a chief's abandoned house, said to date from Spanish times. The platform had been badly disturbed-by Japanese soldiers digging foxholes, it was said. Its retaining wall, 1.5 ft. high, was built of coral heads carried up from the shore. We dug here only one day, March 1, and then decided that the site's disturbed condition and the small yield of artifacts did not justify further effort. Ruuway 1 and 3 show how unproductive house plat- forms may sometimes be, since they are usually built in a short time and yield few artifacts. Thus they offer no time sequence, except when they are built upon an earlier site, as, for instance, at Boldanig, a priest's house mound in Malaj village, Kanifay municipality, where our excavation uncovered ash lenses, apparently indicating earlier fireplaces and consequently earlier surfaces. On February 9 we cleared the house site, Ruuway 1, and dug rectangle A2-3 B2-3 to a depth of 12 in. The soil was damp and clayey, black near the sur- face, becoming gray a few inches down. It was diffi- cult to sift, and we had to finger the lumps of clay to determine whether they contained any hard object. Consequently, the work went very slowly. The yield in the first six inches consisted of a few potsherds, bits of coral, a small amount of shell, one small fragment of quartz, and a tiny one of whitish stone. One of the men, Untaman, tested these two bits of stone on an iron file and drew sparks from the quartz, but not from the other stone. The site's yield in the 6-12 in. block was less than in the upper 6 in. We decided to quit Ruuway 1 and undertake excavations in Ruuway 2, the cookhouse mound, which had a great quantity of potsherds on its surface. The northern edge of Ruuway 2 is about 36 ft. south of the southern edge of Ruuway 1. The distance of 50 ft., recorded in diagram 2 between the grids in Ruuway 1 and Ruuway 2, is not the actual distance between their borders. We cleared most of the surface of Ruuway 2. The site is 40 ft. long from north to south and about 20 ft. wide from east to west. The ground on three sides is level. The highest part of the mound is 2 ft. above the level ground on the east side and 5 or 6 ft. south of the mound's center. The distance from this highest B point to a small tree, left standing in the northern part of the mound, is 14 ft., the base of the tree being 1 ft. 2 in. below the mound's summit. North of the tree, the mound slopes to ground level in about 6 ft. On the west, the ground falls away to a depression, perhaps former- ly used as a taro patch. The potsherds found on the surface of Ruuway 2 were gathered and washed on February 13, 14, and 15. (See pl. 25, a.) Their weight is 4,245 oz. All are of the characteristic, laminated type, a late type which three women of Gitam village still know how to make, although pottery is no longer made in Yap. We began excavations in Ruuway 2 about noon on February 15, starting with rectangle B3-4 C3-4 at the highest part of the mound. We uncovered a stone about 20 in. long, lying flat in the ground. Moorou, an aged subchief of Balebatt who was one of the crew, said it was probably one of three hearthstones used to support a cooking pot. On February 24, in digging the 18-24 in. block of rectangle B2-3 C2-3 we encountered four coral blocks (Pontis sp.) which Mooru thought were part of the origi- nal mound of the cookhouse. These are shown in plate 25, c, which also shows wall B2-C2, viewed from the north side of the rectangle. Wall B3-C3 shows only as a 6-in. step with one coral block overhanging it. Rec- tangle B3-4 C3-4 was dug to a depth of 30 in. Plate 25, b shows the 30-in. wall B4-C4, viewed from the south. The 18 cubic feet screened in the house platform Ruuway 1 yielded only 15.5 oz. of potsherds, the 9 cubic feet in the house platform Ruuway 3 yielded 3.5 oz., while the 126 cubic feet screened in the cookhouse mound Ruuway 2 yielded 2,672 oz. PENIN The Penin site lies at an elevation of 10 to 15 ft. on a low slope above the waterfront of Kanif village in Dalipebinau municipality. We visited it first on February 26, parking the car at the Kanif schoolhouse, about 150 yds. to the south. It is an old site, no longer occupied, and belongs to Finiginam, chief of the village and magis- trate of the municipality, who gave us permission to ex- cavate. It is upslope from the waterfront of Kanif village and about 250 ft. inland from the head of a slough in the mangrove swamp. Its western edge dips down to a taro swamp, with a paved walk running northeast-southwest along the far edge of the swamp. Three excavations, a short distance apart, were made on sloping ground-one at the garbage dump, two at spots where cookhouses once stood (diag. 3). The gar- bage dump is on the steepest slope, the ground dropping 53 in. in a distance of 21 ft. between stakes G8 and Gl. 30" 30' 24" |: ! 4jj2It 4Y i Y2 I l 2 3 4 5 6 B7 A2 A3 XI X2 Ruuway 2 Ruuway I Ruuway 3 Diagram 2. Plan of excavations in Ruuway, at border of Balebatt and Orouro villages, Rull muniicipality. 155 ANTHROPOLOGICAL RECORDS G 18 -1 50'o 2" 6" E3 7 6 4 Cl BI Al Zi Cookhouse I 3 6'' 2 CCookhouse 2 G age Hi JI Diagram 3. Plan of excavations in Penin, Kanif village, Dalipebinau municipality. Our informant said that cookhouse 1 was much the oldest of the three spots excavated. The disintegration of the shell excavated at Penin, especially at cookhouse 1, suggests a leaching action in the drainage of the site, according to Professor F. F. Harradine, Soil Technologist of the University of California. Gilrow, an old Kanif man, said that the fragments of glass and china found in the garbage dump were pieces of objects brought by voyagers from Ulithi, who stopped at Kanif on their way back from the Philippines. Two sacred objects in the jungle near an old house platform in Penin were noted but were not disturbed. One was a small circular stone pavement about 18 in. in diameter, in the center of which was imbedded a roughly cylindrical stone (pl. 26, a), 15 in. high and about 4 in. in diameter. About 15 ft. away and far- ther back in the jungle was a cubical cairn made of slabs of stone (pl. 26, b). In the box thus formed was a large, whole Hippopus hippopus shell, which con- tained pearl-shell (Pinctada margaritifera) money. Removing the front slab of stone and lifting the top valve of the shell, we could see the money. On February 25, when we visited Amin village in the hills on the west side of Map Island, we were taken to a somewhat similar sacred spot in the bush near by. It is called Tinimeau and is owned by an Amin man named Fhlabizin. Here, on a low knoll, is a stone pavement, 6 ft. square, bordered by stones set on edge; its use is unknown. Near it are an old house platform and a "rubbish" dump, surmounted by a circle of shells which includes several Tridacna and Hippopus shells and Charonia trumpets. BOLDANIG AND WOLOM Boldanig and Wolom are really parts of one site in Malaj village, in the municipality of Kanifay on the west coast. This is one of the neatest villages of Yap. The main road runs north and south through the town parallel to the shore and within sight of the Philippine Sea. The road is bordered on its inland side by a stone wall 4 or 5 ft. high, cut at intervals by lanes leading to the inland (eastern) part of the village. Boldanig and Wolom are dedicated to Wuthrei, god- dess of childbirth, who is supposed to have descended from the sky on a spider web. Wolom is the shrine of the goddess; there is another well known shrine to this goddess at the village of Alog, Ueloj municipality. Boldanig is the mound (pl. 26, c) of the house where the priests of Wuthrei lived and is 142 ft. northeast of the shrine in the jungle. (Measurements are taken from stake E1 at Boldanig to stake A2 at Wolom; see diag. 4.) There were a few pieces of stone money in the vicinity of the house mound. A man about seventy years old told us that the house at Boldanig had last been inhabited when he was about eighteen, that is, about the beginning of the twentieth century. Another old man, Gilsowuth, gave us an account of the last priest who lived at Boldanig. Gilsowuth and his wife Ruutingen offer an example of the levirate, since Ruutingen was originally the wife of Gilsowuth's older brother, now dead. Gilsowuth, who is now about eighty years old, had attended the mission school at Boldanig when he was seven or eight; this was in Spanish times, that is, after 1881. Gilsowuth knew Wayag, the last priest at Boldanig and said that Wayag, older than he, never married. Wayag knew no "magic," but performed ceremonies for sick people. He died in Japanese times, but not at Boldanig, which was no longer occupied at that time. The Japanese wrecked the walls of the house at Boldanig, when they were hunting for land crabs. No Yapese ever went there to hunt, since Boldanig was a taboo place. Nevertheless, our excavation crew did not hesitate to catch and cook (on the motor block of our car) the crabs they found there. The lane leading to the Boldanig house mound is directly in line with the men's clubhouse on a rocky knoll at the shore. Going eastward on this lane from the main highway, one passes on the right a Japanese concrete cistern, 76 ft. from the junction, which is F 0 c 8 0 44 7. 6. 3 0 I 5F3 F 2 f Garb, 156 GIFFORD: EXCAVATIONS IN YAP T ;72is841 90" 84" ZI Al B I C I DI 90" El ~~~Fl LlI L2 Surface fire boxg Ml M2 Diagram 4. Plan of excavations at Boldanig house site, Malaj. no longer used because of a great opening in its northern wall. Ninety- six feet farther on, one turns from the lane diagonally to the right at an angle of 45. The Boldanig house site is 88 ft. beyond this point. Boldanig is 52 ft. long (east-west) and 19 ft. wide (north-south). On the surface, near the middle, was a stone fire box, formed by four slabs set on edge, which was removed in the course of excavation. The photograph, plate 26, d, was taken looking into the fire box from the south, at a distance of about 5 ft. The north stone of the fire box was the highest point of the mound and was about 37 in. above the level ground on the north. Our first excavation, rectangle Al-2 Bl-2, was made alongside the west stone of the fire box; wall Bl-2, 6 ft. long, was close to the stone slab. Our crew explained that fire boxes of this kind were for heating the house, not for cooking. We saw similar ones in the men's clubhouse at Malaj, where the floor was covered with flattened strips of areca-palm trunk. Ash lenses and black soil, apparently indicating earlier fireplaces, were found below the surface of the Boldanig house mound. Several were encountered in various rectangles, down to a depth of 58 in. These heatths below the surface seem to indicate that the platform was not built to its full height in kone operation. If this had always been the residence of the priests, the construction was probably the work of successive generations. The mound was composed largely of sandy beach debris, which had been used as fill within the re- taining walls of the house platform. Most of the stone .fragments in the fill were coral, with only an occa- sional bit of schist. Lumps of lime-cemented sand began to appear be- tween 36 and 42 in. and continued below this. At a tdepth of 72 in. we began to find objects with sand cemented to them with lime, indicating our approach to the water table. Cemented sand had evidently en- Lased roots, which subsequently rotted away, leaving sand tubes. These continued to occur to a depth of 84 and 90 in., where the ground water prevented fur- tlier excavation. The water was brackish. (See pl. 27, a, b.) The site of the shrine Wolom, 127 ft. east of the man road, is level. The lane leading to it is 129 ft. south of the lane to Boldanig. Of the shrine itself (pl. 27, c), only four cornerstones and two taller stones remain. The two tall stones, set one in the middle of each end, are 32 in. high; the distance between them, marking the east-west length of the shrine, is 8 ft. 4 in. The width of the shrine is 5 ft. 8 in. The two sides, measured between cornerstones, are of unequal length; the north side 7 ft., the south 7 ft. 5 in. Thus the structure is apparently hexagonal, a miniature of the typical Yapese dwelling. The two tall stones for- merly supported a ridgepole for the low, thatched roof. Wolom resembles the shrine of the god Magro- goy, described earlier under "Walgom." The Wolom shrine, according to our informant Gilsowuth, was formerly a tiny hut, with a roof thatched with coconut leaves. It was so small it could be entered only by crawling. Offerings to the goddess, shell money and coconuts, were taken to the priest, Wayag, who then prayed to the goddess. Gilsowuth said he had never been inside the shrine. Ground water was encountered nearer the surface in Wolom than in the mound at Boldanig: at a depth of 60 in. in Wolom, at 84 in. in Boldanig. At both places, lime-consolidated material first occurred a couple of feet above ground water. No hearths were found at Wolom and there was no evidence that the shrine had ever been used as a habitation. Its small size, of course, would prevent its' being so used. Beach gravel was encountered at a depth between 54 and 60 in. The rectangle was dug within the shrine precincts, the walls of the excavation (6 ft. east-west, 3 ft. north-south), paralleling the house walls. PEMRANG The Pemrang site is a garbage dump in the inland part of Giror village, Galiman municipality. It is said to have been used originally as a cookhouse and garbage heap by the hereditary priests, whose duty it was to prevent typhoons from striking and washing away the southern part of the island. This is a low- lying, sandy coastal plain, which, according to myth- ology, was originally built up by a typhoon. The first priest, a very big man, began to function after the land had been built up. His grave, now marked by a great tree, is not far from Pemrang. The mound contains countless thousands of shells. Fanoway, one of Giror's hereditary chiefs, asserted that the mollusks deposited here were not harvested in the Galiman region but were collected by people in certain villages on the island of Map and were brought here by canoe as offerings. A chief from Map accompanied the offerings. When the mollusks were delivered to the priest at Pemrang, the animals were dead and rotten and wholly unfit for food. The priest therefore deposited them on the garbage heap. Fanoway was quite old at the time of our interview on April 16. He said that, even in his youth, he had never seen the last priest who officiated at Pemrang. The Pemrang mound is 85 ft. long from northeast to southwest and 58 ft. wide near its center. At its top, near its center, it is 41 in. high. The slope is so gradual that it is impossible to measure the two diameters with absolute accuracy. Several large trees stand on the site, and there is much smaller growth. The surface of the mound is littered with thousands of shells and hundreds of potsherds. The most abun- dant shell species was Strombus luhuanus. A few sake 157 ANTHROPOLOGICAL RECORDS bottles, fragments of iron pots, and a piece of a Spanish pot with the letter "C" impressed on it were found on the surface. Hundreds of potsherds were collected before the grid was laid out for excavation. As shown in table 1, shell constituted from 30 to 57 per cent of the mound material in the four sam- ples taken at Pemrang from depths ranging from 3 to 21 in. We found nothing to approach this quantity in any other site. At 27 in., the percentage was 1; at 21 in. it was 30; and below 21 in. the highest per- centage was 7. Could this mean that the abundance of shell first evident at 21 in. represents the begin- ning of the offerings from Map? Our crew member, Nuuan, thought so. Two conflicting radiocarbon dates in table 22 suggest that the date of the beginning of the shell accumulation at 21 in. may have been about A.D. 1800. In the course of excavation we found, besides coral, occasional fragments of stone foreign to this locale, which probably originated in the higher parts of Yap. Ash lenses, indicating fires, were found at various levels down to 42 in. At Pemrang we put three members of the crew in charge of three of the rectangles dug. Each was given assistants, and they worked carefully and kept excellent notes in English. Francis Nuuan was respon- sible for rectangles A1-2 B1-2; Vincent Untaman for A3-4 B3-4; Joseph Sogon for A7-8 B7-8. The Giffords took charge of A5-6 B5-6. 8l 2 3 4 5 6 90' i 84!' 84" 84"T Al 2 3 4 5 6 z ! a 9 Do Al1 30" 9 l 0 A II 42" 8 ZIO ZIl Diagram 5. Plan of excavations, Pemrang garbage site, Giror. 158 COMPOSITION OF SITES No selection was exercised in taking samples of midden deposit from pit walls in the excavated sites. Whatever came-stone, shell, sand, earth, coral, potsherds, etc.-was bagged for the samples. These ranged in weight from 280 gm. to 932 gm. Beginning at 3 in. below the surface, samples were taken at 6-in. intervals. The midden samples have been analyzed by Dwight T. Wallace, who screened them through a wire screen with meshes 3 mm. square. The materials intercepted by the screen were then segregated: stone, pumice, coral, charcoal, mollusk, decapod, echinoderm, bone, potsherds. The residue consists of fine fragments of these materials, plus soil, sand, and ash which passed through the screen. When segregated, the various constituents, and also the residue, were weighed. Table 1 lists each constituent by its percentage of the total weight. The analyses may be compared with those published for Fiji (Gifford, 1951, pp. 202- 205) and for New Caledonia (Gifford and Shutler, 1956, pp. 16-27). Pumice and coral were not segre- gated in the Fijian samples. TABLE 1 Composition of Sites: Percentage of Weight (x, less than 2 of 1 per cent) Site E e (depth in inches)* 0 0 o o 0 _ )C ' CU ,-d CU 0* cu- 0 U0 0 0 0 C) Walgom, Wall C1-2 3 in ............... 5 95 374 34797 9 in ............... 7 x 93 551 34798 15 in ............... 5 95 470 34799 21 in ............... 3 97 502 34800 27 in ............... 2 98 518 34801 Ruuway 2, Wall B2-C2 3 in ................ 15 x x 85 577 32049 9 in ............... 28 72 574 32050 15 in ............... 37 x x x x 63 666 32051 21 in ............... 22 x 78 538 32052 27 in ............... 16 x x 84 672 32053 Penin, Wall Al-Bl 3 in ................ 32 x 68 419 32436 9 in ............... 42 58 375 32437 15 in ............... 58 42 379 32438 21 in ............... 61 39 603 32439 Penin, Wall F5-G5 3 in ................ 33 x 2 65 490 32440 9 in ............... 28 72 327 32441 15 in ............... 30 70 491 32442 21 in ............... 35 65 536 32443 27 in ............... 38 62 373 32444 Penin, Wall H2-J2 3 in . ........ 53 47 685 32445 9 in . ........ 39 1 4 56 657 32446 15 in . ........ 28 x 4 68 636 32447 21 in - 49 51 616 32448 27 in ...... ....... 38 62 692 32449 *Figures for depth indicate the depth at which the sample was taken. [ 159 ] ANTHROPOLOGICAL RECORDS TABLE 1 (continued) Site CU d' ~a (depth in inches) 0 C- o 04 0 ) y w S F-4 Cd CU a) 0 CU 0 0 a)0 0 a) 0 Boldanig, Wall Al-Bl 3 in ............... 2 x 6 2 x 90 280 33856 9 in ............... x x 2 2 x 96 458 33857 15 in ............... x x x 3 x x 97 491 33858 21 in ............... x x x 2 x x 98 595 33859 27 in ............... 44 x 1 55 824 33860 33 in ............... x 1 2 x x x 97 441 33861 39 in ............... x x x 1 x x x 98 573 33862 45 in ............... x x x 1 x 99 524 33863 51 in ............... x x x 1 x x 98 496 33864 57 in ............... x 2 1 x x x 97 654 33865 63 in . .x x 1 x 99 459 33866 69 in x 1 99 435 33867 Pemrang, Wall A4-B4 3 in ................ 1 5 50 x 7 37 519 34783 9 in ............... 1 x 1 57 x x x 40 540 34784 15 in ............... 13 x 5 36 x 3 43 520 34785 I 21 in ............... 1 x x 30 x 69 692 34786 27 in ............... 3 1 x x 96 699 34787 33 in ... x x x 1 x x 99 932 347881 39 in ... x x x 5 x x x 94 913 34789 45 in ... x x 4 x x 95 669 34790 51 in ... x x 1 7 x x 3 89 846 34791 57 in ... x x x 1 x x x 99 953 34792 63 in ... x x 1 x 99 565 34793 69 in ... x x 3 x x 96 408 34794 75 in x x 100 670 34795 81 in x x x x 100 722 34796 The character of each of the five sites may influ- ence its composition. Ruuway 1 and 3 are house plat- forms, Ruuway 2, a cookhouse mound, with a quantity of potsherds on the surface. The excavated part of Penin comprises two cookhouse sites, of different periods, and a garbage heap, all close together on the slope of a hill. Boldanig and Walgom, once house mounds of priesthoods, have retaining walls of stones. Pemrang, once the garbage mound of the typhoon priesthood, has thousands of molluscan shells on the surface. Stone.-The dominant stone in the three northern sites of Walgom, Ruuway 2, and Penin is schist. At Ruuway 1, stone constituted 26 per cent of the total weight (341 gm.) of our sample (32047) from Wall A2-B2, depth 3 in. At Ruuway 2 at 3 in. stone was 15 per cent of the total weight (577 gm.) of the sample (32049) from Wall B2-C2. At both depths char- coal was less than 2 of 1 per cent. In the southern sites, Boldanig-Wolom and Pem- rang, stone is generally scarce; sandy soil, sand, and fine beach debris make up the bulk of most of the samples analyzed. These two sites are on a flat coastal plain, whereas the three northern ones are on low hill slopes. Pumice.-Pumice occurs in only one of the 52 samples of mound material analyzed, this from a depth of 63 in. in Boldanig. In the course of excava- tion 44 samples of pumice were collected from screen- ings: 9 from depths of 6 to 78 in. at Boldanig, 34 from 0 to 84 in. at Pemrang, and 1 from Rugog's grave at Teb in Tomil municipality. No pumice was found at Ruuway, Penin, or Walgom. Professor Howel Williams, of the University of California, has the following to say about the 44 pumice samples. The dark pumice fragments are presumably andesi- tic, and the white fragments (almost all of the samples) are probably dacite pumice. But whether andesitic or dacitic, all of the pumice is extremely poor in crystals or devoid of them. This being the case, correlations are impossible. The presumptio is that all of the pumice was derived from distant volcanoes; how much was airborne, and how much was waterborne could only be determined by obser- vations of field-occurrence. However, considering the size of many of the fragments, it seems likely that most and perhaps all were waterborne. We saw occasional pieces of pumice on sandy beac in Galiman, Gagil, and Rumung municipalities. Evide it is often washed ashore, but the places of its origin are unknown. The possible sources suggested by Pro- 160 GIFFORD: EXCAVATIONS IN YAP fessor Williams are the Philippines, New Guinea, and Indonesia, including the famous volcano Krakatoa, which erupted with unprecedented violence in 1883. Charcoal.-Charcoal was found in all five sites and at various levels, but always in small quantities which never amounted to as much as half of one per cent of the weight of the midden sample. Only a few samples from Fiji (Gifford, 1951, p. 203) and New Caledonia (Gifford and Shutler, 1956, p. 17) yield 1 per cent or more of charcoal. Coral.-Coral, found as a fringing reef everywhere around the shores of Yap, is used for retaining walls of all sorts. In our samples, however, it is not con- spicuous. Indeed, it is absent in those taken from Walgom and Penin, two sites on slopes above the coastal plain, although at both of these blocks of coral are used. Mollusks.-Although molluscan shell was found at all sites, many midden samples show none. There is none in the five from Walgom, although there were mollusks at the site. The largest quantity of shell is found in the samples from Pemrang; shell constitutes 57 per cent of the total weight of one sample from a depth of 9 in. However, as we have pointed out, great quantities of mollusks were brought from Map as offerings to the shrine at Pemrang. Midden samples from Boldanig, which, like Pemrang, is situated on the coastal plain in southern Yap, yield much less shell. A comparison of the occurrence of molluscan shell in samples of mound material from California, Fiji, New Caledonia, and Yap is interesting. The averages for the four regions, in terms of percentage of total weight, are as follows: California 52 per cent (Gif- ford, 1916, p. 15); Fiji 13 per cent (Gifford, 1949, pp.223-224); New Caledonia 9 per cent (Gifford and Shutler, 1956, p. 30); Yap 4 per cent. The fact that in Yap women do not usually gather mollusks may possibly account for the reduced percentage there. Decapods.-Although, as in other Pacific islands, decapods are used as food, midden samples from Walgom, Ruuway, and Penin yield none. In Boldanig and Pemrang samples, the percentages are all less than half of one per cent. Echinoderms.-Echinoderm fragments, of sand dol- lars and sea urchins, are found only in samples from Boldanig and Pemrang. Bone.-Most fragments of bone seem to be from fishes or turtles. Ruuway, Boldanig, and Pemrang samples yield these. No bone is found in the samples from Walgom and Penin. Potsherds.-Potsherds occur in varying quantities in our samples. There is none from Walgom, and only a trace from Ruuway, in spite of the abundance of potsherds at this site. Penin samples from most levels yield no potsherds, but samples from two levels have 4 per cent each and a third has 2 per cent. Six samples from Pemrang, from depths be- tween 3 and 51 in., yield potsherds: 7 per cent at 3 in., 3 per cent at 15 and 51 in., less than half of 1 per cent at the three other levels. The potsherds in four samples from Boldanig, from depths between 39 and 63 in., constitute less than half of 1 per cent of each sample. Residue.-The residue in the samples reflects the nature of the terrain of each site. Walgom residue is mostly red clay of volcanic origin, Ruuway and Penin residues are greenish gray clay, apparently related to schist bedrock of similar color. Boldanig and Pemrang on the southern coastal plain yield sandy residue which often contains finely comminuted beach debris. In all sites the darker material in the upper few inches is due to fires and to decaying vegetable material, humus. 161 ZOOLOGICAL SPECIMENS The zoological specimens collected throw some light on the protein portion of the food of the ancient Yapese. No specimens of plant origin, except charcoal, were excavated. The only domestic animal before the coming of Europeans seems to have been the chicken, which is also present as a feral or an indigenous wild species. Pigs and dogs were introduced by Europeans and the former are now abundant enough, but there are few dogs. Pig bones and teeth were found in excavations but no dog bones. During our stay in Yap we saw a few cattle but no horses or donkeys. UNWORKED BONE Man.-Human teeth and bones, apparently not burials, were found at Boldanig and Pemrang: Boldanig, molar tooth at 6-12 in.; Pemrang, 2 molar teeth, 24-30 in.; Pemrang, bones, 6-12 in. and 42-60 in. A piece of a skull from Pemrang (48-60 in. level) is charred, but no other bone fragments show signs of burning. Hence there is not sufficient evidence to suggest cannibalism. From Rugog' s grave, in Teb village, Tomil municipality, fragments of skull were recovered at 20 in. Mammal.-In aboriginal times rats and flying foxes were apparently the only mammals in Yap. No rem- nants of these, however, were found in the screenings. Pig bones and teeth are presumably all postcontact. Dr. David H. Johnson kindly identified the fragments of mammal bone excavated: all but one were pig, one was whale. The whale specimen (32337) was the lateral process of a vertebra, from a whale about the size of the blackfish (Globicephala). It came from a depth of 6-12 in. at Ruuway 2. There were five lots of pig bone from the Penin garbage dump (rectangle Fl-8 Gl-8). This dump was on a slope, whose descent was 53 in. in 21 ft. All the fragments were found in the 0-6 in. level, except three fragments of vertebra (32733) from a depth of 12-18 in. well down the slope (in rectangle F4-5 G4-5). The depth at which these three were found does not necessarily indicate that they were older, since the ground in this area has slipped. The bones in the 0-6 in. level comprise fragments of humerus and ulna (32533), 2 cheek teeth (32534), deciduous premolar and part of maxilla (32582), unworn crown of molar (32582), and the distal end of a humerus (32586). Dr. Johnson comments in his letter of October 29, 1957, on the surprising absence of dog and rat. It seems remarkable also that we found no pig bones in the other four sites. Perhaps this indicates a late introduction of the pig. Bird.-Eight lots of bird bones were recovered, which have been identified by Dr. Dean Amadon, of the American Museum of Natural History. The bones include femora, tibiae, radii, and one humerus. Seven lots from Pemrang are chicken (Gallus), and one from Wolom (12-18 in.) is probably chicken. The seven lots are from depths of 6 to 60 in. Reptile.-There were many more reptile bones than mammal or bird, although none were found in Walgom. Joseph R. Slevin and Alan E. Leviton, of the California Academy of Sciences, have kindly identified the species. All but three appear to be turtle, probably the green turtle (Chelonia mydas Linne). These three (35613, 35655, 35921) are lizard vertebrae (Varanus sp., probably V. indi- cus). Two vertebrae (35588), from a depth of 24-30 in. at Pemrang, are of the monitor lizard (V. indicus Daudin). The reptile bones were distributed as follows: Ruuway 2, 0-6 in., 2 lots, 6-12 in., 1 lot, 12-18 in., 3 lots; Penin, 0-6 in., 2 lots; Boldanig, 0-6 in., 2 lots, 6-12 in., 3 lots, 12-18 in., 2 lots, 18-24 in., 3 lots, 24-30 in., 1 lot, 30- 36 in., 1 lot, 42-48 in., 2 lots, 48-54 in., 1 lot, 54-60 in., 1 lot, 66-72 in., 1 lot; Wolom, 0-6 in., 1 lot, 60-66 in., 1 lot; Pemrang, 0-6 in., 5 lots, 6-12 in., 5 lots, 12-18 in., 2 lots, 18-24 in., 2 lots, 24-30 in., 3 lots, 30-36 in., 5 lots, 36-42 in., 1 lot, 42-48 in., 3 lots, 48-54 in., 1 lot, 54-60 in., 1 lot, 60-66 in., 2 lots, 66-72 in., 1 lot. Fish.-Fish bones were the most abundant of all verte- brate remains in the archaeological deposits on Yap. None, however, were found in the Walgom house mound, where fish were taboo as food for the resident priest. The 1,809 cubic feet of deposit yielded 4,053.3 gm. of fish remains, averaging more than 2 gm. per cubic foot. A comparison of these figures with the data from Viti Levu (slightly more than 1 gm. of fish remains per cu. ft. of deposit) and New Caledonia (about 3/4 gm. per cu. ft.) indicates that the Yapese depended more than the Fijians and New Caledon- ians on fish for food. The fish bones were distributed as follows in the sites sampled: Walgom, 99 cu. ft., none; Ruuway, 153 cu. ft., 2,076.8 gm.; Penin, 171 cu. ft., 123.5 gm.; Boldanig- Wolom, 765 cu. ft., 191.3 gm.; Pemrang, 621 cu. ft., 1,661.7 gm. TABLE 2 Distribution of Fish Bones (wt. in gm.) Depth Site (in.) Ruuway Penin Boldanig-Wolom | Pemrang 6 221. 2 114.5 12.5 504. 0 12 906. 0 9.0 29. 4 253. 3 18 834.3 34.7 365.0 24 95.3 8.5 150.0 30 20.0 7.2 34.7 36 60.8 68.2 42 11.0 24.6 48 13.5 44.0 54 4.5 57.2 60 0.7 77.8 66 54.0 72 8.5 17.4 78 11.5 [ 162 1 GIFFORD: EXCAVATIONS IN YAP Perhaps the proximity of Ruuway to both Tomil Harbor and the outer coast made it a favorite cooking spot, and thus accounts for the abundance of fish remains at this site. The weight of the fish bones found at different depths in the four sites is given in table 2. In general, upper levels yield more than lower levels. All fish remains collected in Yap were studied by the ichthyologist, Dr. Henry W. Fowler, of the Aca- demy of Natural Sciences of Philadelphia, to whom we are indebted for the following identifications (see also table 3). TABLE 3 Distribution of Fishes Depth (in.) Species 6 | 12 | 18 | 24 | 30 | 36 | 42 | 48 | 54 | 60 | 66 | 72 | 78 . . . . .~~~~i Ruuway Balistes .......................... Cheilinus undulatus (Rueppell)...... Diodon hystrix (Linne) ............ Labroid ........................... Lethrinus ......................... Lutjanid .......................... Lutjanus gibbus (Forskael)......... Monotaxis grandoculis (Forskael)... Naso ............................. Naso lituratus (Forster)............ Novaculichthys .................... Percoid ........................... Scarus ............................ Serranid .......................... Serranus .......................... Penin Diodon hystrix (Linne)............. Halichoeres ....................... Labroid ........................... Monotaxis grandoculis (Rueppell).... Scarus ............................ Boldanig- Wolom Cheilinus undulatus (Rueppell) ...... Labroid ........................... Monotaxis grandoculis (Forskael)... Scarus ............................ Scomberomorus ?.................. Pemrang Aetobatus narinari (Euphrasen)..... Balistes.......................... Cheilinus undulatus (Rueppell)...... Diodon hystrix (Linne)............. Labroid ........................... Lutjanus gibbus (Forskael)......... Monotaxis grandoculis (Forskael)... Naso ............................. Percoid ........................... Scarus ............................ Serranus .......................... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 163 ANTHROPOLOGICAL RECORDS Determinable species Aetobatis narinari (Euphressen) .... .... 1 Lutjanus gibbus (Forskael) ............. 4 Monotaxis grandoculis (Forskael) ....... 21 Cheilinus undulatus (Rueppell) .......... 16 Naso lituratus (Forster) ..... .......... 1 Diodon hystrix (Linne) ................. 20 Generic determinations Lutjanus .............................. 1 Lethrinus .......... ................... 1 Halichoeres ........................... 1 Novaculichthys ....... ................. 1 Scarus ................................ 40 Naso . ................................. 4 Balistes ........... ................... 11 Scomberomorus? ....... ............... 1 General determinations Percoid ............................... Serranid .............................. Labroid ............................... 7 6 16 Dr. Fowler comments (Feb. 28, 1957): ECHINODERMS Echinoderm remains are scarce in our sites. They comprise sea-urchin and sand-dollar fragments, which have been identified by Professor G. Wyatt Durham, of the Department of Paleontology of the University of California. Two pencil-like spines of an unidentified species of Heterocentrotus (36836 and 35681) were excavated in Pemrang at depths of 12-18 in. and 48-54 in. respec- tively. Fifteen lots of sand-dollar fragments of the species Jacksonaster (Laganum) tongaense (Quoy and Gaimard) were excavated. Two of these lots (36003 and 35589) are from Pemrang at levels of 18-24 in. and 24-30 in., respectively. One (32649) is from Penin, 0-6 in. Twelve lots are from Boldanig, at levels ranging from 0-6 in. to 36-42 in., as follows: 0-6 in. 6-12 in. 12- 18 in. 18-24 in. 36-42 in. 32913, 33329, 33619, 33645, 33109, 33285, 33874 33904 33935 33964 34306, 34338 Although these determined species are few, they are very interesting, inasmuch as only a very few fishes have ever been recorded from the island of Yap. DECAPODS Crabs and lobsters, represented in the screenings by fragments, were found in two sites, Boldanig- Wolom and Pemrang. Their distribution by depth is given below. None were excavated in Walgom, Ruuway, or Penin. The species were identified by Professor Tune Sakai, Yokahama University, Kama- kura, Japan, and Dr. R. Imaizumi, Tohoku Univer- sity, Sendai, Japan. Boldanig-Wolom (1 spec. from Wolom) ? Atergatis (family Xanthidae), 20-24 in. Brigus (probably B. latro), 0-6 in., 6-12 in., 12-18 in., 42-48 in. Coenibita sp., 36-42 in. (Wolom), 30-36 in., 72- 78 in. Caldorphia (family Perthenopidae), 6-12 in. Geograpsus (family Grapsidae), 6-12 in. Geograpsus or Sesarma (family Grapsidae), 42-48 in. Pemrang Brs (?), 72-78 in. Carpilius sp. (family Xanthidae), 6-12 in., 12- 18 in. Coenobita sp., 54-60 in., 60-66 in. Scylla serrata (family Portunidae), 60-66 in. Apparently no regional significance attaches to the absence of decapod fragments in the three northern sites. Because our specimens are so few we are inclined to invoke chance as the ex- planation of both the geographic and stratigraphic distributions. MOLLUSKS Mollusks, represented by the molluscan shell found in the deposit, were the most abundant life forms in our sites, except coral, which was used for build- ing purposes. Identification of molluscan species was kindly undertaken by Dr. Leo G. Hertlein, of the California Academy of Sciences. A total of 167 species have been identified, of which 52 are bi- valves and 115 univalves. The univalves include Nautilus. Table 4 presents the distribution of molluscan species by sites; tables 5 to 9 give the distribu- tion by depth in each site. The five sites exca- vated produced the following number of species: Walgom Ruuway Penin Boldanig Pemrang 13 67 37 106 114 According to the Yapese, many of the shells at Pemrang are from Map Island and were brought to Pemrang as offerings. The antiquity of this practice is not known. The analysis of samples of the Pemrang deposit (table 1) suggests that it may have begun at the time represented by the depth of 21 in. in the deposit. From that depth up to the surface shell is much more abundant than below. Ninety-three species of the total of 114 found at Pemrang come from the upper 24 in. and may be chiefly from Map. Obviously, the num- ber of different species at this site may be due to a mixture of local, Galiman, and Map species. The Yap archipelago is so small that there is prob- ably little local difference in molluscan fauna except on an ecological or environmental basis. Nothing about the stratigraphic distribution of species in the five sites suggests any change in food habits of the Yapese. I 164 GIFFORD: EXCAVATIONS IN YAP 165 TABLE 4 Distribution of Molluscan Species Sites Species } - R P B Peman iWalgom i Ruuway I Penin I Boldanig I Pemrang Bivalves Arca fusca Bruguiere ............. x Arca scapha Chemnitz ............. x x x x x Arca velata Sowerby x x x Asaphis dichotoma Anton x x Cardita variegata Bruguiere ........ x Cardium alternatum Sowerby x x x Cardium fragum Linne .............. x Cardium mindanense Sowerby x x x x Cardium unedo Linne x x x x Chama brassica Reeve x x x x Chama lazarus Linne ............... x x x x Chama obliquata Reeve x Codakia punctata Linne x x Codakia tigerina Linne x Corbis fimbriata Linne x Ctena divergens Philippi x x Gafrarium pectinatum Linn n x x x x Gafrarium tumidum R6ding x x x x Hippopus hippopus Linne ............ x x x Isognomon ephippium Linne x Isognomon isognomum Linne x Lioconcha castrensis Linne .......... x Lucina stearnsiana Oyama x x x Macoma awajiensis Sowerby x Mactra maculata Gmelin x x x Mactra mera Deshayes ............ x Mesodesma striatum Gmelin ....... x Modiolus metcalfei Hanley ......... x Ostrea crista-galli Linne . . ......... x Ostrea echinata Quoy and Gaimard x Ostrea mordax Gould x Ostrea plicatula Gmelin x Ostrea sinensis Gmelin ............ x Pecten radula Linne x x Pecten tigris Lamarck x Periglypta puerpera Linne x x x x Periglypta reticulata Linne ......... x Pinctada margaritifera Linn e x x x Pitar striata Gray x x x Pitar subpellucidus Sowerby x Spondylus ducalis Roding in Bolten. . x Spondylus radians Lamarck x Spondylus zonalis Lamarek ......... x Tapes literata Linne x x x Tapes variegata Sowerby x x Tellina corbis Sowerby x Tellina discus Hanley x Tellina palatam Iredale x x x x Tellina scobinata Linne x x x Tridacna crocea Lamarck x x x x Tridacna maxima Bolten ........... x x x x x Tridacna noae Bolten x x x x Univalves Amphiperas ovum Linne x Atys cylindrica Helbling x Atys naucum Linne x Bursa bufonia Gmelin x 166 ANTHROPOLOGICAL RECORDS TABLE 4 (continued) Sites Species I I Walgom j Ruuway Penin I Boldanig Pemrang Univalves (continued) Bursa lampas Linne ................ x x Cantharus proteus Reeve ........... x x x Cassis cornuta Linne ............... x x x Cassis vibex Linne ................. x Cerithium alternatum Sowerby ...... x x x Cerithium aluco Linne ............ x x x x Cerithium asperum Linne .......... x Cerithium columna Sowerby ...... x Cerithium nodulosum Bruguiere ..... x x x x Cerithium traillii Sowerby x Cerithium vertagus Linne x Charonia tritonis Linne x x Cheilea hipponiciformis Reeve x Conus capitaneus Linne ............. x Conus catus Hwass x Conus coronalis Rbding in Bolten x Conus distans Hwass x Conus ebraeus Linne x Conus fuscatus Lamarck x Conus literatus Linne x x Conus magus Linne ................ x x Conus marmoreus Linne ............ x x x x Conus miles Linne x x Conus pulicarius Bruguiere x Conus quercinus Bruguiere x Conus spectrum Gmelin ............ x Conus sponsalis Bruguiere x Conus striatus Linne x x Conus textile Linn n x Conus vexillum Gmelin ............ x Conus vitulinus Hwass in Bruguiere. x x x Cymatium aquatile Reeve x Cymatium gemmatum Reeve ........ x Cymatium pileare Lamarck ........ x x x x Cymatium tuberosum Lamarck ..... x x x x Cypraea annulus Linne x Cypraea arabica Linne x x Cypraea carneola Linne ............ x x Cypraea chinensis Gmelin .......... x x Cypraea errones Linne x Cypraea felina Gmelin x x Cypraea lynx Linn n x x Cypraea mauritiana Linn n x x Cypraea moneta Linne x x x Cypraea talpa Linn. x Cypraea tigris Linne x x x Cypraea vitellus Linne x Drupa margariticola Broderip ...... x x Fasciolaria filamentosa Lamarck x x Harpa amouretta Bolten ............ x x Hawaiia minuscula Binney .......... x Hipponix acutus Quoy and Gaimard.. x x lopas sertum Bruguiere x x Lambis chiragra Linne x Lambis lambis Linne ............... x x x x x Littorina scabra Linne ............. x x x Melania gouldiana Reeve ........... x x Melanoides loebbecki Brot x Mitra episcopalis Linne. x - - -I GIFFORD: EXCAVATIONS IN YAP 167 TABLE 4 (continued) Site Species _ Walgom Ruuway | Penin Boldanig Pemrang Univalves (continued) Mitra litterata Lamarck ........... x Mitra mitra Linne x Mitra plicaria Linne x Mitra tayloriana Sowerby x Murex capucinus Lamarck x Murex torrefactus Sowerby x Nassarius bellula A. Adams x x Nassarius caelatus A. Adams x x Nassarius crematus Hinds x Nassarius deshayesii Hombron and Jacquinot x Natica onca Reeve x Natica rufilarbis Reeve x Nautilus pompilius Linn n x Nerita albicilla Linne ............... x x x Nerita chamaeleon Linne ............ x x x Nerita plicata Linne x x Nerita polita Linne ................. x x x Nerita undata Linne . ............... x x x x Neritina coromandeliana Sowerby x x Neritina mertoniana Recluz x Neritina plumbea Sowerby x x Neritina turrita Gmelin x x Opeas gracilis Hutton .............. x Polinices flemingiana Recluz x Polinices opacus Recluz x x x Polinices pyriformis Recluz ..... x x x Pythia scarabeus Linne x Sistrum concatenatum Lamarck ..... x Strombus bulbulus Sowerby x Strombus gibberulus Linn n x x Strombus labiatus Bolten ........... x x Strombus lentiginosus Lin ne x x Strombus luhuanus Linn n x x x Strombus mutabilis Swainson x x Strombus urceus Linne x x Strombus ustulatus Schumacher ..... x x x x Terebra dimidiata Linn n x Terebra maculata Linne x x Terebralia sulcata Born .... ....... x x x x x Thais pica Blainville .............. x x x Tonna perdix Linne x x Trochus incrassatus Lamarck x x Trochus maculatus Linn n x Trochus niloticus Linne ............. x x x x x Trochus pyramis Born ............. x x x x Turbo argyrostomus Linne .......... x x x Turbo chrysostonus Linne ........... x x Turbo nicobaricus Gmelin .......... x x Turbo ticaonicus Reeve ............ x x x x x Vasum ceramicum Linn n x Vasum cornigerum Lamarck ....... x x x Vasum turbinellum Line n x x ANTHROPOLOGICAL RECORDS TABLE 5 Molluscan Species in Walgom: Distribution by Depth Depth (in.) Species - - 6 12 18 24 30 Bivalves Arca scapha Chemnitz .x x x x Chama lazarus Linn .x x x x Hippopus hippopus Linne. x Ostrea crista-galli Linne x Ostrea sinensis Gmelin .x x Spondylus zonalis Lamarck. x Univalves Cymatium pileare Lamarck x Lambis lambis Linne x Strombus labiatus Bolten. x Terebralia sulcata Born x Trochus niloticus Linn. x Trochus pyramis Born x Turbo ticaonicus Reeve. x TABLE 6 Moiluscan Species in Ruuway 2: Distribution by Depth Depth (in.) Species | 6 | 12 18 | 24 30 Bivalves Arc& scapha Chemnitz .x x x x x Arca velata Sowerby x Asaphis dichotoma Anton x Cardium alternatiim Sowerby x x x x Cardium mindanense Sowerby x x x x Cardiu.m unedo Linn .x x Chama brassica Reve .x x x x Chama lazarus Line .x x x x Gafrarium pectinatum Linne x x x x x Gafrariujm tumidum Bolten x x x x Isognomon ephippium Linne x x Isognomon isognomuim Linne x Lioconcha castrensis Linne x x x Lucina stearnsiana Oyama x x x x x Macoma awajiensis Sowerby x Mactra maculata Gmelin x Ostrea echinata Quoy and Gaimard. x x Ostrea mordax Gould. x Pecten radula Linn. x Periglypta puerpera Linne .x x x x x Pinctada margaritifera Linn e x x x Tapes literata Linn .x x x x x Tapes variegata Sowerby .x x Tellina palatam Iredale .x x x x x Tellina scobinata Linn. x Tridacna crocea Lamarck x Tridacna maxima Bolten x Tridacna noae Bolten. x 168 GIFFORD: EXCAVATIONS IN YAP 169 TABLE 6 (continued) Depth (in.) Species -1 j 24 3 6 1 12 1 18 24 30 Univalves Cantharus proteus Reeve ........... x Cassis cornuta Linne ............... x x Cerithium alternatum Sowerby ...... x x Cerithium aluco Linne .. ........... x Cerithium asperum Linne .......... x Cerithium columna Sowerby ........ x Cerithium modulosum Bruguibre x Conus capitaneus Linne x Conus marmoreus Linn n x x Conus spectrum Gmelin x Conus vexillum Gmelin ..... ....... x Cymatium gemmatum Reeve ........ x Cymatium tuberosum Lamarck x Cypraea carneola Linne x Cypraea chinensis Gmelin .......... x Drupa margariticola Broderip x Harpa amouretta Bolten ............ x Hawaiia minuscula Binney .......... x Hipponix acutus Quoy and Gaimard.. x Lambis lambis Linne ............... x x x x x Littorina scabra Linne .............. x Melania gouldiana Reeve ........... x Mitra litterata Lamarck x Nerita albicilla Linne ............... x x Nerita chamaeleon Sowerby ........ x x Nerita polita Linne ................. x Nerita undata Linne . . .............. x x Opeas gracilis Hutton .............. x Polinices pyriformis Recluz ........ x Sistrum concatenatum Lamarck ..... x x Strombus ustulatus Schumacher x x Terebralia sulcata Born .... ....... x x Thais pica Blainville x Trochus niloticus Linne'.. ........... x x x Turbo argyrostomus Linne ......... x x Turbo chrysostomus Linne .......... x Turbo nicobaricus Gmelin .......... x Turbo ticaonicus Reeve ............ x x x x Vasum cornigerum Lamarck x 170 ANTHROPOLOGICAL RECORDS TABLE 7 Molluscan Species in Penin: Distribution by Depth Species Depth (in.) 6 1 12 1 18 Bivalves Arca scapha Chemnitz ..... ....... x x Cardium mindanense Sowerby ...... x x Cardium unedo Linne x Chama brassica Reeve ............ x Chama lazarus Linne ............... x x Gafrarium pectinatum Linne ........ x x x Gafrarium tumidum Bolten ........ x x x Hippopus hippopus Linne ............ x Periglypta puerpera Linne .......... x Pitar striata Gray ................ x Spondylus ducalis R6ding in Bolten. x Tellina palatam Iredale .... ....... x x Tridacna crocea Lamarck ......... x x Tridacna maxima Bolten .... ...... x x Tridacna noae Bolten ..... ........ x x Univalves Bursa lampas Linne . . ............. x Cerithium alternatum Sowerby x x Cerithium aluco Linne .............. x x Cerithium nodulosum Bruguiere .... x x Conus marmoreus Linne . .......... x x Conus vitulinus Linne ...... ........ x Cymatium pileare Lamarck x Cymatium tuberosum Lamarck ..... x Cypraea carneola Linne x Cypraea moneta Linne x Cypraea tigris Linne ............... x x Lambis lambis Linne . ............. x x x Nerita undata Linne ............... x Polinices opacus Recluz x Strombus bulbulus Sowerby ........ x Strombus luhuanus Linne ........... x Strombus ustulatus Schumacher .... x x x Terebralia sulcata Born ........... x x x Trochus niloticus Linne ...... ...... x x Trochus pyramis Born ............ x x Turbo ticaonicus Reeve ..... ...... x x Vasum turbinellum Linne .......... . x I 171 GIFFORD: EXCAVATIONS IN YAP TABLE 8 Molluscan Species in Boldanig: Distribution by Depth Depth (in.) Species_________ _|__6 |12 |18| 241 301 361421 481 541 ,60 661 721 781841 90 Bivalves Arca divaricata Sowerby.......... Arca scapha Chemnitz............ Arca velata Sowerby ............. Cardium alternatum Sowerby...... Cardium fragum Linne............ Cardium mindanense Sowerby. Cardium unedo Linne............. Chama brassica Reeve ........... Chama lazarus Linne............. Chama obliquata Reeve........... Codakia punctata Linne ........... Codakia tigerina Linne ........... Corbis fimbriata Linne........... Ctena divergens Philippi.......... Gafrarium pectinatum Linne....... Gafrarium tumidum Bolten........ Lucina stearnsiana Oyama........ Mactra maculata Gmelin.......... Ostrea plicatula Gmelin .......... Pecten radula Linne.............. Pecten tigris Lamarck ........... Periglypta puerpera Linne ........ Pinctada margaritifera Linne. Pitar striata Gray .............. Pitar subpellucidus Sowerby ...... Spondylus radians Lamarck....... Tapes literata Linne ............. Tellina corbis Sowerby........... Tellina discus Hanley ............ Tellina palatam Iredale........... Tellina scobinata Linne........... Tridacna crocea Lamarck ........ Tridacna maxima Bolten.......... Tridacna noae Bolten............. Univalves Atys cylindrica Helbling.......... Atys naucum Linne............... Cantharus proteus Reeve ......... Cassis cornuta Linne............. Cerithium alternatum Sowerby .... Cerithium aluco Linne............ Cerithium nodulosum Bruguiere... Cerithium trailii Sowerby......... Cerithium vertagus Linne......... Charonia tritonis Linne........... Cheilea hipponiciformis Reeve .... Conus catus Hwass............... Conus literatus Linne ............ Conus magus Linne .............. Conus marmoreus Linne.......... Conus miles Linne............... Conus pulicarius Bruguibre....... Conus quercinus Bruguiere ....... Conus sponsalis Bruguiere........ Conus striatus Linne............. Conus vitulinus Bruguiere ........ Cymatium pileare Linne.......... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x ANTHROPOLOGICAL RECORDS TABLE 8 (continued) Depth (in.) Species |_6 | 12 18 | 24 | 30 | 36 | 42 | 48 | 54 | 60 | 66 | 72 | 78 | 84 | 90 Univalves (continued) Cymatium tuberosum Lamarck... Cypraea annulus Linne ........... Cypraea arabica Linne .......... Cypraea errones Linne........... Cypraea felina Gmelin............ Cypraea lynx Linne .............. Cypraea mauritiana Linne ........ Cypraea moneta Linne............ Cypraea tigris Linne ............. Drupa margariticola Broderip..... Fasciolaria filamentosa Lamarck.. Iopas sertum Bruguiere .......... Lambis lambis Linne............. Littorina scabra Linne . ......... Melanoides loebbecki Brot........ Mitra plicaria Linne ............. Mitra tayloriana Sowerby......... Nassarius beliula A. Adams ...... Nassarius caelatus A. Adams ..... Natica rufilarbis Reeve........... Natica onca Bolten............... Nerita albicilla Linne ............ Nerita chamaeleon Linne ......... Nerita plicata Linne.............. Nerita polita Linne............... Nerita undata Gmelin............. Neritina coromandeliana Sowerby.. Neritina mertoniana Recluz ....... Neritina plumbea Sowerby ........ Neritina turrita Gmelin........... Polinices opacus Recluz.......... Polinices pyriformis Recluz....... Strombus gibberulus Linne........ Strombus labiatus Bolten ......... Strombus lentiginosus Linne ...... Strombus luhuanus Linne.......... Strombus mutabilis Swainson...... Strombus urceus Linne ........... Strombus ustulatus Schumacher.... Terebra maculata Linne.......... Terebralia sulcata Born.......... Thais pica Blainville............. Tonna perdix Linne .............. Trochus incrassatus Lamarck..... Trochus niloticus Linne .......... Trochus pyramis Born ........... Turbo argyrostomus Linne........ Turbo chrysostomus Linne........ Turbo ticaonicus Reeve........... Vasum cornigerum Lamarck...... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 172 GIFFORD: EXCAVATIONS IN YAP TABLE 9 Molluscan Species in Pemrang: Distribution by Depth Species 61 Depth (in.) 12 1 18 1 24 1 30 1 36 1 42 1 48 54 60|66 |72 78 84 90 Bivalves Arca fusca Bruguiere.............. Arca scapha Chemnitz ............. Arca velata Sowerby.............. Asaphis dichotoma Anton........... Cardita variegata Bruguiere ....... Cardium alternatum Sowerby....... Cardium mindanense Sowerby ...... Cardium unedo Linne .............. Chama brassica Reeve............. Codakia punctata Linne............. Ctena divergens Philippi........... Gafrarium pectinatum Linne........ Gafrarium tumidum Bolten......... Hippopus hippopus Linne........... Lucina stearnsiana Oyama ......... Mactra maculata Gmelin........... Mactra mera Deshayes ............ Mesodesma striatum Gmelin ....... Modiolus metcalfei Hanley ........ Periglypta puerpera Linne ......... Periglypta reticulata Linne......... Pinctada margaritifera Linne....... Pitar crocea Reeve................ Pitar striata Gray................. Tapes literata Linne............... Tapes variegata Sowerby........... Tellina palatam Iredale ............ Tellina scobinata Linne ............ Tridacna crocea Lamarck.......... Tridacna maxima Bolten........... Tridacna noae Bolten.............. Univalves Amphiperas ovum Linne ........... Bursa bufonia Gmelin.............. Bursa lampas Linne............... Cantharus proteus Reeve........... Cassis cornuta Linne.............. Cassis vibex Linne ................ Cerithium aluco Linne ............. Cerithium nodulosum Bruguiere..... Charonia tritonis Linne............ Conus coronalis Roding in Bolten... Conus distans Hwass............... Conus ebraeus Linne............... Conus fuscatus Lamarck........... Conus literatus Linne.............. Conus magus Linne................ Conus marmoreus Linne........... Conus miles Linne ................ Conus striatus Linne .............. Conus textile Linne................ Conus vitulinus Bruguiere.......... Cymatium aquatile Reeve .......... Cymatium pileare Lamarck ........ Cymatium tuberosum Lamarck ..... Cypraea annulus Linne............. Cypraea arabica Linne............. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 173 ANTHROPOLOGICAL RECORDS TABLE 9 (continued) Species [ 6 12 18 24 30 36 Depth (in.) Species_|_ 6 | 12 | 18 | 24 | 30 | 36 | 42 | 48 | 54 6 60 | 66 1 72 1 78 1 84 1 90 Univalves (continued) Cypraea chinensis Gmelin.......... Cypraea felina Linne .............. Cypraea lynx Linne............... Cypraea mauritiana Linne.......... Cypraea moneta Linne............. Cypraea talpa Linne............... Cypraea tigris Linne .............. Cypraea vitellus Linne............. Fasciolaria filamentosa Lamarck ... Harpa amouretta Bolten............ Iopas sertum Bruguiere............ Lambis chiragra Linne ............ Lambis lambis Linne.............. Littorina scabra Linne............. Melania gouldiana Reeve........... Mitra episcopalis Linne............ Mitra mitra Linne................. Murex capucinus Lamarck.......... Murex torrefactus Sowerby......... Nassarius bellula A. Adams........ Nassarius caelatus A. Adams....... Nassarius crematus Hinds ......... Nassarius deshayesii Hombron and Jacquinot ...................... Nautilus pompilius Linne........... Nerita albicilla Linne.............. Nerita chamaeleon Linne........... Nerita plicata Linne .............. Nerita polita Linne................ Nerita undata Linne ............... Neritina coromandeliana Sowerby ... Neritina plumbea Sowerby.......... Neritina turrita Gmelin............ Plectotropis cathcartae Reeve ...... Polinices flemingiana Recluz....... Polinices opacus Recluz ........... Polinices pyriformis Recluz........ Pythia scarabeus Linne' .. ......... Strombus gibberulus Linne......... Strombus lentiginosus Lamarck..... Strombus luhuanus Linne........... Strombus mutabilis Swainson....... Strombus urceus Linne ............ Strombus ustulatus Schumacher..... Terebra dimidiata Linne........... Terebra maculata Linne ........... Terebralia sulcata Born ........... Thais pica Blainville .............. Tonna perdix Linne................ Trochus incrassatus Lamarck...... Trochus maculatus Linne........... Trochus niloticus Linne............ Trochus pyramis Born............. Turbo argyrostomus Linne......... Turbo nicobaricus Gmelin.......... Turbo ticaonicus Reeve............ Vasum ceramicum Linne........... Vasum cornigerum Lamarck........ Vasum turbinellum Linne........... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x I 174 GIFFORD: EXCAVATIONS IN YAP CORAL The 245 samples of coral taken from screenings have been examined by Professor J. Wyatt Durham. Tables 10 and 11 present the distribution of the 25 coral species identified from our five sites. Some specimens could be identified by genera only. In Walgom, Ruuway, and Penin the coral specimens were found on hill slopes and were evidently brought there by man. In Boldanig-Wolom and Pemrang, on the low southern part of the island of Yap, some of the corals may be beach debris, since this area has been frequently inundated by the sea. Indeed, legend says that this region owes its existence to a great typhoon, which caused the sea to deposit the low shelf forming this part of the island. At a depth of 18-24 in. in the cookhouse mound (Ruuway 2) we encountered four coral heads of the genus Porites. These, our informant Moorou said, were probably from the original mound of the cook- house. Coral blocks are commonly used everywhere in Yap for the retaining walls of house platforms. Nowadays many are trimmed into brick form with steel hatchets. According to Muller (1917, p. 117), calcined coral was used for white pigment. TABLE 1 0 Distribution of Corals Site Species Walgom Ruuway Penin Boldanig- Pemrang .Wolom Acropora sp. Alveopora sp. Astreopora myriophthalma Lamarck Astreopora sp. Coralline algae . Cyphastrea seralia Forskael. Cyphastrea sp. Diplora sp. Echinopora lamellosa Esper. Euphyllia glabrescens Chamisso and Eysenhardt . Favia pallida Dana. Favia sp. Favites complanata Ehrenberg. Favites favorsa Ellis and Solander .... Favites sp. Fungia echinata Pallas. Fungia fungites Linnn. Fungia rapanda Dana. Fungia sp. Goniastrea retiformis Lamarck. Heliopora sp. Hydrophora microconos Lamarck. Leptoria gracius Dana. Leptoria sp. Madrocis sp. Millepora dichotoma Forskael. Millepora exaesa Forskael. Millepora sp. Montipora sp. Montipora ? . Pavona frondifera Lamarck. Platygyra lamellina Ehrenberg. Platygyra rustica Dana. Platygyra sp. Plesiastrea versiposa Lamarck. Pocillopora damicormis Linne. Pocillopora verrucosa Ellis and Solander . Pocillopora sp. Porites sp. x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 175 ANTHROPOLOGICAL RECORDS TABLE 1 0 (continued) Site SpeciesBodng Walgom Ruuway Penin Bolodanig Pemrang Psammacora sp x x Seriatopora angulata Kluzinger x Seriatopora hystrix Dana x Seriatopora sp .x x Stylophora murdax Dana .............. x Stylophora pistillata Esper x Stylophora sp x x Tubinaria sp x TABLE I 1 Distribution of Corals by Depth Depth (in.) Site_____ _________ ___ _|_ _ 6 12 18 1 24 | 30 | 36 | 42 1 48 | 54 | 60 1 66 | 72 1 78 | 84 90 Walgom Acropora ......................... Coralline algae................... Heliopora sp...................... Montipora (?) sp.................. Porites sp........................ Stylophora murdax Dana........... Ruuway Acropora sp...................... Cyphastrea seralia Forskael....... Montipora (?) .................... Montipora sp..................... Pocilopora damicornis Linne....... Porites sp........................ Seriatopora sp.................... Penin Acropora sp...................... Cyphastrea seralia Forskael....... Favia pallida Dana................ Favites complanata Ehrenberg..... Favites sp........................ Heliopora sp...................... Leptoria sp....................... Leptoria gracilis Dana............ Millepora sp...................... Montipora sp..................... Montipora (?) .................... Platygyra sp...................... Platygyra rustica Dana............ Pocilipora damicornis Linne....... Porites sp........................ Boldanig-Wolom Acropora sp...................... Astreopora myriophthalma Lamarck Coralline algae................... Cyphastrea sp.................... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x I 176 GIFFORD: EXCAVATIONS IN YAP TABLE 1 1 (continued) Site } 6 121181 24 ~ Depth (in.) Site_____|__ 6 12 |18 24 30 | 36 1 42 | 48 1 54 | 60| 66 1 72 78 84 1 90 Boldanig-Wolom (continued) Cyphastrea seralia Forskael...... Euphyllia glabrescens Chamisso and Eysenhardt ............... Favia pallida Dana .............. Favites complanata Ehrenberg .... Fungia sp........................ Fungia sp. (possibly of rapanda group) ........................ Fungia echinata Pallas ........... Fungia fungites Linne ............ Leptoria sp...................... Madrocis sp..................... Millepora sp..................... Montipora sp.................... Platygyra sp..................... Platygyra lamellina Ehrenberg.... Plesiastrea versipora Lamarck... Pocillopora sp................... Pocillopora damicornis Linne..... Porites sp....................... Psammocora sp.................. Seriatopora sp................... Seriatopora angulata Klunzinger... Seriatopora hystrix Dana......... Stylophora pistillata Esper........ Stylophora sp.................... Pemrang Acropora sp..................... Alveopora sp.................... Astreopora sp................... Coralline algae .................. Cyphastrea seralia Forskael...... Diplora sp....................... Echinopora lamellosa Esper...... Favia pallida Dana.............. Favites complanata Ehrenberg .... Favites favorsa Ellis and Solander Fungia echinata Pallas ........... Fungia fungites Linne ............ Fungia rapanda Dana............. Fungia sp....................... Goniastrea retiformis Lamarck... Heliopora sp..................... Hydrophora microconos Lamarck.. Leptoria gracilis Dana........... Leptoria sp...................... Millepora dichotoma Forskael..... Millepora exaesa Forskael ....... Millepora sp..................... Montipora sp.................... Montipora ( ................... Pavona frondifera Lamarck....... Platygyra lamellina Ehrenberg.... Platygyra rustica Dana........... Platygyra sp..................... Plesiastrea versipora Lamarck ... Pocillopora damicornis Linne..... x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x 177 178 ~~~~~~ANTHROPOLOGICAL RECORDS TABLE 11 (continued) Site ~~~~~~~~~~~Depth (in.) Site j~~~~ 12 18 124 130 1361 42 148 154 60 66 72 78 184 1-90 Pemrang (continued) Pocillopora verrucosa Ellis and Solander . . .. .... . ...x x Pocillopora sp ...........x x x x x x x x Porites sp .............x x x x x x x x x x x x x Psammocora sp.......... x Stylophora sp . .... ......x Tubinaria sp............ x 178 ARTIFACTS POTTERY Two types of pottery are found in Yap. The com- moner and later type is laminated; the scarcer and earlier is unlaminated. The brown laminated ware, with no tempering added intentionally to the clay, we found at every site visited. It probably occurs at all sites, since it is a late type. It fractures like rich pastry, in thin, ragged layers with projecting edges, and is of poor quality (see pl. 32, a). The second and earlier type of ware was found chiefly in our excava- tions in southern Yap, at Boldanig and Pemrang. It is unlaminated, tempered, and makes a relatively clean break (see pl. 32, b, c). Both types yield a hard- ness range of 2 to 3 by University of Michigan stand- ards. The relationship of the unlaminated and laminated tpes is not clear. Two possibilities suggest them- selves. Laminated ware may represent a degeneration in technique from the earlier unlaminated; we must confess that some sherds are difficult to assign to one or the other type. Alternatively, laminated ware may have been introduced by later immigrants. The two types occur in our five sites as follows (percentage of total weight of sherds): Walgom, 99 per cent laminated, 1 per cent unlaminated; Ruuway, 99 per cent laminated, less than 0.5 per cent unlami- nated; Penin, 100 per cent laminated; Boldanig-Wolom, 70 per cent laminated, 30 per cent unlaminated; Pem- rang, 72 per cent laminated, 28 per cent unlaminated. Total weight of potsherds collected is 20,669 oz. The average yield in Yap is 11 oz. to the cubic foot; in New Caledonia, 20 oz.; in Viti Levu, 12 oz. Table 12 gives the distribution of potsherds of both laminated and unlaminated wares in Walgom, Boldanig- Wolom, and Pemrang, expressed in percentage of total weight of potsherds found. The Penin sherds and all but two from Ruuway were laminated, hence we have nt included these sites in the table. Penin yielded 1,915 oz., Ruuway, 6,917 oz. Table 12 shows the transition in dominance from aminated ware in the upper levels to unlaminated vare in the lower levels. This distribution parallels tat of incised and relief wares at two sites excavated EnViti Levu (Gifford, 1951, tables 20 and 25). Some othe laminated sherds recorded at low levels may ot rightfully belong there, for in the sandy soil of loldanig and Pemrang we were plagued by the falling objects from higher levels or from the surface oges of the pits. Land crabs and hermit crabs were ostantly moving about and falling into the pits at * ht. They probably knocked in some laminated sherds m upper levels; these, when detected, were thrown y. Sherds from the surface could often be recog- * ed by spots of white fungus growth. The suggested time difference between the two types pottery is corroborated by radiocarbon dates, which of respectable antiquity at Pemrang, A.D. 176, and ldanig, A.D. 847. These dates suggest that unlami- ed, tempered pottery, found in the lower levels at se two sites, is not only the older of the two types, [ 1 but is actually very ancient. Samples of the unlami- nated ware were unhesitatingly identified by Dr. Alex- ander Spoehr as the same type as his Marianas Plain ware from Saipan and Tinian. TABLE 12 Distribution of Potsherds at Walgom, Boldanig, and Pemrang: Percentage of Total Weight (x, less than 2 of 1 per cent) Site Laminated Unlaminated Weight (depth in inches) (oz.) Walgom 6 in......... 12 in......... 18 in......... 24 in......... 30 in......... Total. Boldanig 6 in......... 12 in......... 18 in......... 24 in......... 30 in......... 36 in......... 42 in......... 48 in......... 54 in......... 60 in......... 66 in......... 72 in......... 78 in......... 84 in......... 90 in......... Total. Pemrang Surface ...... 6 in......... 12 in......... 18 in......... 24 in......... 30 in......... 36 in......... 42 in......... 48 in......... 54 in......... 60 in......... 66 in......... 72 in......... 78 in......... 84 in......... 90 in......... Total ...... 100 93 100 90 50 99 97 94 90 83 90 65 86 86 67 39 8 x 12 0 79 84 82 70 55 31 8 6 1 1 7 5 4 22 20 7 . . . 10 50 1 3 6 10 17 10 35 14 14 33 61 92 100 88 100 21 16 18 30 45 69 92 94 99 99 93 95 96 78 80 ________________________I _________________I_______ 208. 40 21. 20 1. 80 5. 10 2.40 238. 90 49. 20 69. 95 49. 25 29. 58 48. 57 29. 90 25. 50 20. 05 18. 70 20. 40 20. 35 23. 40 18. 82 26. 00 26. 35 476. 02 2, 352. 00 3, 456. 90 2, 096. 10 1, 590. 80 498. 65 304. 00 230. 20 172. 50 135. 25 113. 00 73. 00 79. 30 13. 75 4.50 2.50 11, 122.45 L79 I ANTHROPOLOGICAL RECORDS Pottery making in Yap is limited at present to the low-class village of Gitam. Unfortunately, neither the present-day potters there, nor Gaak, the chief of the village, knew of any mound or rubbish heap in the vicinity, which we might excavate. The story of the origin of fire-making and of pottery-making was re- corded for us in English by Vincent Untaman from the lips of Finek, an aged potter, who related the story in Yapese. Long, long ago the Yapese did not have fire. Food was eaten raw after being dried in the sun- shine. A woman from Dinay village [Dinau on map], named Ruwlayan, was [once] drying food on a stone platform. While so doing, she heard thunder. The thunder was a goddess named Dirra' ["thunder"] who fell from the sky on to a kind of tree, called rachaloy. She called on Ruwlayan to take her down, so Ruwlayan helped her get out of the tree. The goddess asked Ruwlayan what she was doing. Ruwlayan told the goddess that she was there dry- ing food. Then the goddess asked Ruwlayan to bring her some clay and some sticks. Ruwlayan brought her the clay and the sticks. The goddess rubbed the sticks together by the drill method and produced fire. Then she taught Ruwlayan how to make pots. She told Ruwlayan to bring clay. Ruwlayan got it at Dinay, where the goddess showed her what kind of clay to get. The goddess asked Ruwlayan whether she would like the pot to break into pieces or not. Ruwlayan told her she would like to have pots break. Then the goddess told her that it was better for the pots to get broken into pieces, so Ruwlayan could earn payment for making new pots. The goddess asked Ruwlayan to help her to get on top of a pandanus tree [choy]. Once she was on top of the tree, the rain fell in torrents and the goddess disappeared. After that Ruwlayan taught the people of Yap how to make fire and pots. The people brought Ruwlayan taro, yams, fruits, shell money, stone money, and other things to trade for her pots. Wilhelm Muller presents two versions of this story from one informant (Muller, 1918, pp. 604-607). Chris- tian (1899, p.320) presents another version. Laminated ware.-The preceding myth evidently re- fers to the laminated ware, which is apparently pecul- iar to the Yap archipelago, being found at all the sites in Yap, Map, and Rumung islands. One reason for the laminated structure is, we think, the absence of inten- tional tempering, although a few sherds show some tempering material apparently present in the clay. The potters, however, declared that temper is never added; indeed we saw one of them, Finek, pick out a small stone fragment from the soft clay. A few laminated sherds (32069, wt. 3.5 oz.), from a depth of 0-6 in. at Ruuway 2, appear to have fine sand in them, but if the potters' statement is true, this must be an ac- cident rather than an intentional addition. A second, and perhaps more important, reason for the peculiar type of fracture may be the way of mold- ing the clay. This is not first formed into cylindrical rolls for coiling, nor are the paddle and anvil used to compact or compress it. As we saw the process, the clay was compressed, shaped, and smoothed with the hands, but without any great pressure. Small bits were added from time to time with the fingers and smoothed into place. Three women of Gitam village still know how to make this characteristic laminated type of pottery. The youngest, whom we did not meet, taught pottery-making in school in the early days of the American occupation of Yap. Ruwaureng, who had married into Gitam from Benik village, and Finek demonstrated the method for us. The two little old women allowed us to take photo- graphs of the process. Mrs. Gifford used kodachrome negative, Gifford black-and-white (see pls. 28-31). Four photographs of pottery-making at Gitam have also been published by Wilhelm Muller (Miller, 1917, pp. 123-125, pl. 35). On March 17 and 24 the two potters demonstrated their craft. Each day, they worked less than two hours. On March 17 Finek shaped the vessel, but the clay was still too soft for trimming, so this process had to be postponed. When we arrived on the seventeenth, the potters had already gotten a mass of yellow clay from a small pit, a few yards from their homes, at the junction of the Kanif road with the main road to the south end of Yap. The pit was at the south edge of a terraced grav Although we arrived at an agreed nine o'clock, the potters did not put in an appearance until ten. Then they showed us the pit. The cylindrical mass of clay, a foot high and eight inches in diameter, was wrapped in banana leaves aM carried to the village in a coconut-leaf basket, suspe from one shoulder by a braided hibiscus-fiber rope attached at each end. The two women seated themsel on the ground under a small tree that gave a little shade. Finek placed the cylindrical mass of clay up- right on a board, made a slight depression in the top with her right fist, and began to form a rudimentary rim, which she continued to enlarge as she deepened the depression. From time to time she turned the clay mass-not the board. Once we saw her use her right fist against the inner wall while she patted the outside with her left hand, but not vigorously or for any length of time, nor did she do this completely around the vessel. Finek did all the work of shaping the pot. Ruwaureng sat by watching, and once she fetched a coconut cup full of water for Finek to moi her fingers as she smoothed the vessel. After an hour the clay cylinder, with the expanded bowl formed in its top, was transferred to a piece of areca spathe. At this time no attempt was made to separate the bowl from its clay matrix. The only tool used in shaping the bowl was a praea tigris shell of medium size, which had had the cap removed so it could be readily held by the colu- mella (pl. 28, b-d). We obtained this tool as a speci- men (31967). Its convex sides were used to smooth the inner surface of the bowl and to shape the in- curved rim. The outer surface was smoothed with thl palm of the hand and the fingers. On March 24, we again watched the potters at w The pots had been set to dry under a one-sided lean- to shelter, open at both ends, and by this time the clay was dry enough to trim. The women had made three vessels, the one we saw shaped on the seven- teenth and two smaller ones made from the surplus clay after the first pot had been cut from the base ol matrix. The potters said that they had rubbed the inside and out with a cowry each day. When the yes 180 GIFFORD: EXCAVATIONS IN YAP were brought out, they had giant taro leaves sticking to their flat bases. The two women sat cross-legged or with one leg tucked under. We could not see the exact position be- cause of their voluminous "grass" skirts. They spread giant taro leaves on their laps and each put a pot up- side down on the leaf. A slender bamboo stick, as thick as one's finger, was split in half lengthwise, each woman taking a half. With the sharp edges the potter pared away the flat base and rounded the bot- tom of the pot, using the stick as a drawing knife, or scraper, pulling it toward her with both hands. Apparently they did not appreciate the advantages of vessels with flat bottoms, for they said these were never made. Plate 30 shows this trimming process. The parings that accumulated on the taro leaves were saved to repair any cracks that might develop during the long period of drying. They told us the vessels would be dried further in the shelter and that they would let us know when it was time to fire them. We inquired twice about this in the following weeks but were told that it was not yet time. Indeed, we left Yap on June 2, without having seen the firing. The vessels we saw made were finally fired on July 25, 1957. At that time Mr. Kaneshiro and his Yapese assistant, Defngin, took photographs of the process. Plate 31 reproduces these photographs, show- ing three steps in the firing. Mr. Kaneshiro's account follows. Defngin (my assistant) and I arrived at the firing site in Gitam, Rull, at 9:40 a. m. Ruwaureng and Finek had already started the fire. While the fire was kept going and more fuel added to build up the coals, Finek moistened and smoothed the small pot with coconut husk and water. At 10:20 Finek placed the pot inverted over the coals and added more fuel (areca palm bark and spathe sections). The fire was kept going for twenty-five minutes during which time Finek moistened and smoothed the other (larger) pot and placed a few stones on the ground upon which to place the fired pot. At 10:45 she picked up the pot with a pole and placed it, right side up, on the stones to cool. (I was in the midst of putting another roll of film in the camera, but Finek, bless her heart, couldn't wait.) More fuel was added to the fire. At 10:55 Finek placed the other pot on the fire, covered it with more fuel (coconut husks and shells and wood) and fired it until about 11:20, then picked it up and set it on stones as with the first. According to Finek and Ruwaureng, pots are not made within the village but in the grasslands (teyid). The site for this firing is actually part of the teyid, although a family has built dwellings on the site in recent years. I asked if women usually, or at least in the past, did not wear grass skirts while making pottery as I'd read somewhere. The women replied that they did and the grass skirts could be worn back to and in the village area. One restriction, at least for Gitam, is that the women may not eat fish prior to firing pots lest the pots crack or peel. It is all right to eat fish after the firing, however. This restriction was. decided a long time ago, they said, and gave no other explanation for it. Other Yapese villages that have a tradition of pottery making are Wenifara', Murru', Deboc, Goocol, Muyub and Numunug, all of which are low-caste (milingay) villages. Ruwaureng (as you know) is originally from Benik, Rull (also a low-caste village), is married to a Gitam man and learned pottery making from Tinag Niga' (now deceased) and Finek. Unlaminated ware.-This ware, identified by Dr. Spoehr as Marianas Plain ware, occurs mostly in southern Yap. In the lower levels at both Boldanig and Pemrang more unlaminated than laminated pottery is found. (See table 12.) Of the two unlaminated sherds found at Ruuway, one (31976), from the 0-6 in. block at Ruuway 1, weighs 2 oz.; the other (pl. 32, e), from the 24-30 in. block at Ruuway 2, weighs 1 oz. Their combined weight is a very small percentage of the 6,917 oz. of potsherds collected at Ruuway. The scant occurrence of this ware at Walgom and at Rugog's grave is indi- cated in the general discussion of pottery and in table 12. The possibility that Pemrang may have been the center of greatest development of unlaminated ware is suggested by the fact that a larger variety of lip types of this ware is found at Pemrang than at Boldanig. (See table 15.) Potsherd rims and lips.-The rim is the upper por- tion of a vessel wall, ending in the lip, which joins the inner and outer walls at the mouth of the vessel. There are three types of rims: straight (or vertical), outcurved, and incurved. The rim varies in thickness; it may thin noticeably as it approaches the lip; it may thicken noticeably; or it may remain the same as the body. a b c d Pfl??i e f 9 h i Fig. 3. Lip types of potsherds. Exterior of pot- sherd is to right. a. Rim curved in, lip thickened, inrolled (32009). b. Rim curved in, lip rounded, not thickened (32010 and 36381). c. Rim curved in, lip flat, not thickened (35004). d. Rim curved in, lip flat, not thickened, with sharp inner edge (36469). e. Rim curved in, with outer edge of lip reflexed, not thickened (35289). f. Rim curved in slightly, lip rounded, not thickened (36818). g. Rim curved in slightly, lip flat, not thickened (35211). h. Rim curved in, lip beveled, thickened (36854). i. Rim curved in slightly, lip thickened, rounded (36563). As in Fiji and New Caledonia, our excavations pro- duced no complete vessel; only sherds were found. Consequently, little can be deduced about ancient pot- tery shapes. Most of the rims we collected are in- curved; some unlaminated rim sherds suggest the 181 ANTHROPOLOGICAL RECORDS straight, or vertical, but no rims of either ware curve out. Muller (1917, fig. 173, p. 124) pictures a vertical- walled bowl, but we do not know whether it is lami- nated or unlaminated. Nine lip types are shown in cross section in figure 3. Five of these (fig. 3, a-e) occur in both laminated and unlaminated pottery; four (fig. 3, f-i) are found only in unlaminated ware (Marianas Plain) from Pem- rang. The five lip types occurring in both laminated and unlaminated are as follows: a, thickened, inrolled; b, not thickened, rounded; c, not thickened, flat; d, not thickened, flat, with sharp inner edge; e, not thickened, with outer edge of lip reflexed. All are in- curved rims. TABLE 13 Distribution of Lip Types of Laminated Rim Sherds: Percentage of Total Weight Types Total Site weight Walgom ....... ... 53 47 ... ... 27 Ruuway 2 ..... 80 8 4 5 3 2,188 Penin ......... 33 61 5 1 ... 470 Boldanig ...... 24 62 7 6 ... 75 Pemrang ...... 44 42 10 2 1 2, 542 These five lip types of laminated sherds (a-e) occur in our five sites in varying frequencies, as table 13 shows. Ruuway 2 is the center of abundance of type a. Types a and b occur in about equal quantities at Pem- rang, where together they constitute about 86 per cent of the total weight of all laminated rims. At Penin the dominant type is b, constituting 61 per cent. Types c, d, and e occur in small quantities, 10 per cent or less, except at Walgom where c is 47 per cent of the total weight of all rims, but it must be noted that only 2 7 oz. of rims were collected at that site. Obviously, Ruuway 2 and Pemrang, each yielding more than 2,000 oz. of laminated rim sherds, provide the most signifi- cant data. Table 14 presents the distribution by depth of five lip types (a-e) of the laminated rim sherds found in the five excavated sites. This distribution is expressed in percentage of total weight; x indicates a negligible quantity. The few sherds collected at Ruuway 1 and Ruuway 3 are omitted. The sample of laminated lip types from Walgom is so small that the percentages of types may not be too accurate. Types a, d, and e are absent at this site; types b and c are about evenly divided. In Ruuway 2, type a is the dominant lip type of the laminated ware (1,747 oz.), occurring in greater quantities in the upper levels and decreasing with depth; it is lacking in the 24-30 in. block. Types c and d are present in all subsurface levels. Type d is not found on the surface and its greatest quantity occurs in the 12-18 in. level. Type e is apparently a late style at Ruuway and Pemrang. In Penin, in contrast to Ruuway 2, lip type b (lami- nated ware) is most abundant (287 oz.). Type e is absent. TABLE 14 Distribution of Lip Types a-e of Laminated Rim Sherds: Percentage of Total Weight (x = trace only) Site Types Weight (depth in inches) a j b I c I d I e (oz.) . X i Walgom 6 in........ 12 in........ 18 in........ 24 in........ 30 in........ Total ..... Ruuway 2 Surface..... 6 in........ 12 in........ 18 in........ 24 in........ 30 in........ Total .... Penin 6 12 18 24 30 in........ in........ in........ in........ in........ Total. Boldanig 6 in........ 12 in........ 18 in........ 24 in........ 30 in........ 36 in........ 42 in........ 48 in........ 54 in........ 60 in........ 84 in........ Total ..... Pemrang Surface 6 in........ 12 in........ 18 in........ 24 in........ 30 in........ 36 in........ 42 in. 48 in........ 54 in........ 60 in........ 66 in........ Total... 42 * .. . . . 100 100 88 61 50 2 50 25.5 28 60 65 100 42 31 . . . 39 . . . * . . * . . 51 42 45 38 22 6 100 . . . * . . 5 23 4 x 25 68.5 62 40 33 46 65 77 100 49 67 100 43 100 17 100 42 44 42 35 48 94 20 100 * . . * . . * . . 58 * . . . . . * . . 3 6 10 13 x 50 5 9 * . . 12 * . . 21 * . . 8 12 * . . 9 * . . * . . * . . 4 10 12 21 23 . 2 5 25 * . . . . . . . . 4 1 . . . * . . * . . * . . . . . 8 36 85 25 50 1 1 2 4 2 . . . 4 21 48 . . . 83 1 3 1 6 8 . . . 55 1 0 . * . . 100 22.20 4.20 1.00 27.40 1, 685.00 386.00 52.00 59.00 4.00 2.00 2, 188.00 270.00 119.50 48.00 27.50 5.00 470.00 13.00 24.00 9.15 3.50 12.67 4.85 1.00 2.30 2.00 1.20 1.00 74.67 752.00 911.40 514.60 270.40 73.75 12.70 2. 00 3.40 0.50 1.00 2, 541.75 1 1 1 * . . * . . * . . . . . I . _ i 182 GIFFORD: EXCAVATIONS IN YAP TABLE 15 Distribution of Lip Types of Unlaminated Rim Sherds: Percentage of Total Weight (x, less than 2 of 1 per cent) Site Types Weight (depth in inches) a lb | ci d | e f I g I h J i (oz.) Boldanig 6 in .............. ... 12 in .100 0. 25 18 in .............. ... 24 in .100 0. 25 30 in .50 50 2.00 36 in .............. ... 42 in .............. 48 in .............. ... 54 in .100 0. 50 60 in .100 1.30 66 in . .25 75 4.00 72 in .100 9.00 78 in .40 60 2.50 84 in .............. 90 in 100 0. 35 Total 20.15 Pemrang Surface .......... 23 64 12 1 95.00 6 in . ............. x 41 13 3 42 107.50 12 in . ............. 48 13 39 52.50 18 in . ............. 6 54 8 31 1 76.40 24 in . . ....... 16 37 41 6 24.60 30 in . ............. 14 47 6 27 6 26.10 36 in . ............. 66 9 25 21.10 42 in . ............. 30 39 6 26 10.10 48 in . ............. 4 16 80 6.25 54 in .............. 60 in . .50 17 33 3.00 66 in . .3 58 38 6.00 Total 428.55 In Boldanig, lip type b (laminated ware) is dominant. We question whether the sample recorded from the 78- 84 in. block really belongs there; it may have fallen in from a higher level. Type e is absent. At Wolom only a small quantity of laminated body sherds was recovered and no rim sherds whatsoever were found. In Pemrang all five types of laminated lips occur, but type e is very scantily represented. Types a and b are dominant. At Pemrang we found four lip types of unlaminated, tempered ware which are not found at other sites (fig. 3, f-i): f, slightly incurved rim, with rounded lip; j slightly incurved rim with flat lip; h, incurved rim with thickened, beveled lip; i, slightly incurved rim with thickened, rounded lip. The percentage of total weight found in Pemrang is as follows: lip type a, 7 per cent; type b, 48 per cent; type c, 12 per cent; type d, 1 per cent; type e, 1 per cent; type f, 38 per cent; type g, trace only; type h, trace only; type i, 2 per cent. (See table 15.) At Boldanig, only lip types b and c of the unlami- nated ware were found: type b, 71 per cent; type c, 29 per cent of total weight at this site. It seems ob- vious that Pemrang is the center of development of the unlaminated, tempered pottery of Yap; indeed, it yielded 429 oz. of rim lips as against only 20 oz. found in Boldanig. Drilled potsherds.-Twenty-one laminated potsherds with drilled holes were collected. The drilling is coni- cal and from the exterior, presumably done after the shaped vessel had dried, but before it was fired. Al- though a couple of specimens show some slight evi- dence of reaming of the hole from the inside, the true drilling is on the exterior of each sherd. One specimen has a second hole, apparently drilled because the first was too close to the lip of the vessel and broke out. Our crew said the purpose of these per- forations was the suspension of the bowls by cords. No unlaminated sherds with drilled holes were found. No drilled sherds were found at Boldanig. In Wal- gom, 1 comes from the 6-in. block. In Ruuway 2, in 183 ANTHROPOLOGICAL RECORDS the upper six inches of the deposit, 3 were found. In Penin, 4 specimens were recovered from the 0-6 in. level and 1 from the 6-12 in. level. The remaining 12 are from Pemrang, as follows: surface, 1; 0-6 in., 4; 6-12 in., 3; 12-18 in., 2; 36-42 in., 1; 42-48 in., 1. Depths seem to indicate that drilled potsherds are limited to late times in Yap and are definitely part of the laminated pottery complex. Gambreled shoulders.-Eight potsherds with gam- breled shoulders were recovered, 3 from Pemrang, 5 from Ruuway 2. The Ruuway specimens are all of laminated ware, four of them body sherds, the fifth (32092) a lip sherd. Depths at which the individual specimens were found are listed as follows: 32008, surface; 32068, 32088, 32092, all 0-6 in.; 32099, 6-12 in. The last specimen (32099) was said by our crew to be from a "pot to hold fire," similar in function to the bowl illustrated in plate 33, a. The Pemrang specimens are all lip sherds. Two (35804, 35933) are laminated ware from depths of 0-6 in. and 12-18 in., respectively; 35804 is illustrated in plate 32, f. The third (36766, from 36-42 in.) is un- laminated and tempered, with a unique decoration of incised crosshatching on its exterior, and may be from an imported vessel. It is illustrated in plate 32, i. Incised ware.-Four sherds with incised decoration were obtained. Three (36765, 36766, 36767) are from Pemrang (36-42 in.); 1 (36911) is from Walgom (6-12 in.). All four are unlaminated and tempered pottery; none of the laminated sherds were incised. The Wal- gom sherd and two of the Pemrang sherds are shown in plate 32, i-k. The fifth sherd, from Pemrang, is discussed in the preceding paragraph. Painted ware.-Four small body sherds and some additional tiny fragments constitute specimen 36767 from a depth of 36-42 in. at Pemrang; on the 4 sherds brick red pigment is applied in parallel lines on a terracotta-colored base. Three of these sherds show only the concave surface, which is painted; the convex surface is missing. Both surfaces of the fourth and largest are visible and both are painted. These sherds are apparently fragments of an open bowl, painted in- side and out. Fragments of china of Chinese and Japanese origin were occasionally found, usually on the surface or in the upper levels of sites. However, one fragment (32641) was found at a depth of 18-24 in. in the gar- bage dump at Penin. The depth has little significance, since the site slopes 53 in. in 21 ft. At Kanif, a large Spanish jar was in use as a lime container. Fragments of a similar jar, with the letter "C" impressed on its shoulder, were found on the surface at the Pemrang site. Petrography.-Miss Reba W. Benedict, a graduate student working under the supervision of Professor G. H. Curtis of the Department of Geology of the Uni- versity of California, has analyzed the mineral com- position of a number of laminated and unlaminated sherds collected by our expedition. The percentage of stone, which she interprets as temper, ranges from 50 to 20 in the laminated sherds; in the unlaminated, from 10 to 50. The percentages of tempering material are shown in table 16. Laminated 32015 (Ruuway 2, 6 in.): In this specimen it is difficult to distinguish between temper and matrix. If only the coarser material is included in the tem- per, the proportions are 10 per cent temper: 90 per cent matrix. There is, however, a great deal of material of intermediate grain size, less than 0. 3 mm. in diameter but much larger than the particles in the clay matrix. This material is composed al- most entirely of hornblende and plagioclase feldspar fragments. The hornblende and much of the plagio- clase are unaltered, so much of this material may have been added with the coarser temper. If this finer material is included with temper, proportions are 35 per cent temper: 65 per cent matrix. 32069 (Ruuway 2, 6 in.): Most of the feldspar and quartz grains are full of inclusions, probably hornblende. None of the alteration appears to be related to the present grain boundaries. 32583a (Penin, 6 in.): The feldspar contains much fine clay material, but it seems to have no systema- tic relation to the present crystal outlines. 32583b (Penin, 6 in.): The smaller grains of feld- spar are highly charged with clay; many of the larger ones are less altered, and some are fresh. Most of the quartz grains are full of small mineral inclusions. 32637 (Penin, 24 in.): Feldspar grains are highly charged with clay minerals; other mineral grains of the temper are quite fresh. 33868 (Boldanig, 6 in.): A number of the feldspar crystals appear to be penetrated by the clay material of the matrix, but most of the feldspar fragments are sharply bounded against the matrix. 35464 (Pemrang, 12 in.): One large quartz grain, otherwise clear, has blades of dark material of low birefringence projecting into it along one margin. Another large quartz grain contains many inclusions of red clay. Quartz grains in the temper of this specimen are all large ( 2- 1 mm.). Many feldspar grains contain numerous inclusions, probably horn- blende, and sometimes also contain some reddish clay. In some grains, this clay seems to be en- croaching on the grain from the matrix, or is most abundant around the edges of the grains. 36934 (Walgom, 6 in.): All the mineral fragments, except the clay and magnetite grains, are very small. Feldspar is fresh or has only light clay alteration. Unlaminated 33236 (Boldanig, 72 in.): Both quartz and feldspar often contain numerous mineral inclusions, but the feldspar is quite fresh. The rock fragments listed under "Other" in table 17 are aggregates of feldspar and epidote, feldspar and muscovite, and feldspar and pyroxene. These combinations of minerals in a single rock fragment are found only in this specimen.' Organic material consists of tests of animals [fora- minifera], probably composed of carbonate material. 35513 (Pemrang, 12 in.): Many of the quartz and feldspar grains contain numerous prismatic inclu- sions, probably hornblende. Feldspar crystals con- tain moderate amounts of clay, but it has no sys- tematic relation to grain boundaries. 35722 (Pemrang, 66 in.): The organic remains are the tests of various forms of animal life [fora- minifera], probably composed of fine-grained car- bonates. About half of the quartz grains are clear, and about half contain numerous inclusions. Some hornblende grains are fresh and some rather heavily altered, but there is no indication that the alteration is related to the grain boundaries. 36378 (Pemrang, 12 in.): The matrix contains, in 184 GIFFORD: EXCAVATIONS IN YAP addition to clay, many small (less than 0. 1 mm. in diam.) fragments of quartz, hornblende, and feld- spar. Several of the quartz-feldspar rock fragments (listed as "other" in the table) are distinguished from any rock fragments in any of the other speci- mens by their graphic texture, an angular interlock- ing of the two minerals. One of these rock frag- ments contains hornblende as well as quartz and feldspar; the hornblende also shows graphic texture. Much of the feldspar shows clay alteration, which is apparently unrelated to the crystal boundaries. The quartz fragments in this sherd are exceptionally large. 36930 (Walgom, 30 in.): Some feldspar grains are perfectly fresh; others are altered. SHELL ARTIFACTS The preponderance of shell tools in an archipelago where hard stones are present would seem to suggest that the Yapese ancestors arrived with a tradition of shell implements. Adzes.-The following report by Sylvia Broadbent is the result of a special study of the shell adzes col- lected by our expedition. (See figs. 4, 5 and pl. 37.) Fifty-five shell adzes were found, 41 at Pemrang, 12 at Boldanig, and 2 at Penin. None was found at Ruu- way or Walgom. (See table 18.) TABLE 16 Matrix and Temper of Pottery Temper Matrix Specimen Per cent Per cent Nature Laminated 32015 ......... 10 90 Red clay 32069 ......... 20 80 Red clay 32582a ........ 20 80 Red clay 32583b ........ 10 90 Red clay 32637 ......... 5 95 Mixed red and yellow clay 33868 ......... 5-10 90-95 Red clay 35464 ......... 5-10 90-95 Red clay 36934 ......... 5 95 Red clay Unlaminated 33236 ......... 50 50 Red clay 35513 ......... 25 75 Red clay 35722 ......... 40 60 Red clay 36378 ......... 10-15 85-90 Yellowish clay 36930 ......... 10 90 Yellowish clay TABLE 17 Relative Abundance of Constituents of Pottery Temper (4, abundant; 3, common; 2, scarce; 1, trace) Mineral fragments Rock fragments Specimen ei 0 ~~~0 CU 0) C co N 04 4 0 1 31 ... 2 .. . . ... .. . co . o x .,0 1 o,l 32583~~~~~~~~~0C a ............4 .. . . . . . . . . . 3544 0 04 2 1 to 3693Z4 .. 3 24 0 . . 0. 3 . C. .. 3 Laminated 32015 ............... 1 4 1 ... ... 3 1 ... 4 1 ... ... 32069......... 4 4 1 3 1 1 1 1 . 2.. .. 32583a ................. ..4 3 3 3 ... ... 2 ... ... ... ... 1 ... 32583b .............. 2 4 ... 2 ... ... 3 ... ... ... ... 32637......... 2 3 1 2 1 1 .. .. 1 .. .. .. .. 33868 . .............. 4 ... 1 ... 3 1 35464 ............... 2 4 2 1 3 . . . . ... 1 ... ... ... ... 36934......... 3 2 2... ..... 3 .. .. 3 .. .. .. .. Unlaminated 33236......... 3 3 2 ... 2 3 4 2 2 .. .. 2 35513......... 4 4 1 3 .. .. 4 .. .. 1 1 ... 35722......... 4 ... 2 .. ... . . 1 ... 4 .. .. .. .. .. 36378.........4 2 1...... ...... 1 3 36930......... 4 3 3 3 . .2 .....1 185 ANTHROPOLOGICAL RECORDS TABLE 18 Distribution of Types of Shell Adzes Type Site and level 1 2a 2b 3 4 5 Unclassified Total Penin 6-in 1 1 2 Boldanig 6-in ........... 1 1 12-in 1 1 18-in ........... 1 1 24-in ........... 1 1 2 30-in 0 36-in ........... 1 1 2 42-in ........... 1 1 2 48-in ........... 1 1 54-in ........... 1 1 60-in 0 66-in 0 72-in 1 1 Pemrang 6-in ........... 12 2 14 12-in ........... 3 1 4 18-in ........... 4 3 7 24-in ........... 1 1 2 30-in ........... 4 4 36-in ........... 2 1 1 4 42-in 0 48-in ........... 1 1 2 54-in ........... 2 1 3 60-in ........... 1 1 66-in ...l.l.1 | 1 The adzes which have enough original surface left for the molluscan species to be determined with rea- sonable certainty are all made of Tridacna. There are, however, a number of pieces whose natural shell sur- face has been completely ground away, making decisive identification of the species impossible. Some of these adzes may have been made of Hippopus shell; indeed, the characteristics of some suggest Hippopus rather than Tridacna. These characteristics include a general massiveness, more likely to be provided by the hinge area of Hippopus than by any but the largest Tridacna; translucency, particularly if the specimen appears mottled, more like Hippopus than Tridacna; and, on one or two of the pieces, what appear to be grooves of the inner surface of Hippopus. The majority of these adzes fall into two classes, depending on the position of the bevel, whether it is on the front face (exterior of shell), edge down (type 1), or on the back, edge up (type 2). The other adzes are classified, according to various criteria, in much smaller groups. In this discussion the terminology of Buck, Emory, Skinner, and Stokes (1930) is used. Type 1. Edge down: 3 7 (30 from Pemrang; 7 from Boldanig). Pl. 37, a-e. The edge of these adzes is formed between the bevel and the back, the chin being on the front. The shape is in general trapezoidal or, perhaps better, parabolic, the edge forming the base and the poll the apex of the parabola. This form ranges to a subrectangle with nearly parallel sides or even with a slight contraction of the width toward the edge. The adzes are not tanged, and there is no particular point at which one can distinguish butt from blade. The sides are normally more or less square, that is, with an angle separating them from back and front surfaces. They are sometimes "under. cut," so they form acute angles with the front sur- face. All but two of the specimens which can be certainly identified as Tridacna fall into this group; a few others are included, in which the shell speciees cannot be certainly identified, since the growth lines have been ground off. All the type 1 adzes in which it is possible to tell what portion of the shell was used are made from the margin farthest from the hinge (see fig.4). Fig. 4. Ideal section of shell adze, type 1. The part removed by grinding is stippled. 186 GIFFORD: EXCAVATIONS IN YAP One side is cut parallel to the edge of the shell. The edge is formed on the side toward the hinge. Right and left valves are used about equally. The adze was first shaped by chipping and then finished by grinding. Unfinished specimens (36734, 36823, 36509) indicate that almost all the actual shaping was accomplished by chipping. Something of the sequence of grinding is suggested by specimens 36674 and 36564, both broken in manufacture. These have completely formed edges and some grinding on the front and back; the right-hand side of each is ground, but the left-hand one is not. The two pieces are made from opposite valves, a fact which suggests that the left-hand side might have been the last part ground in the finishing process. Finished specimens show grinding on all surfaces except the poll, which is usually chipped. Most of the grinding to form the edge produces the bevel, which is on the front (exterior), but the back also shows some marks of grinding at this point. At the poll, almost all the grinding is usually on the back. Only one of these adzes (36430) shows any of the traits listed above as suggesting Hippopus; this specimen shows a mottled translucency. Type 2. Edge up: 10 (7 from Pemrang; 2 from Boldanig; 1 from Penin). P1. 37, f_i. This class is less easily characterized than the preceding; the main distinctive trait on which it is set up is the fact that the edge is formed between the bevel and the front, with the chin on the back of the adze. It is further distinguished by a lateral curvature of the front, in addition to a poll-to-chin curve, giving more or less of a turtleback effect, while the back is flat or even convex, not concave like the back of type 1. In some specimens the curvature of the front may be due to the natural fracture of the shell; in any event, no attempt has been made to eliminate the curve by grinding. Type 1 specimens tend to be concavo-convex in section; type 2 are plano-convex or double-convex. This type may be divided into two subgroups, depending on whether the lateral curvature is very marked or relatively slight. The species of the shell of which the type 2 specimens are made cannot be certainly identified. The following specimens show traits which suggest Hippopus: 33461, 33598, 36729, 35520 are trans- lucent; 33461 and 35944 are massive. 2a. Marked lateral curvature: 5 (1 complete, from Boldanig; 1 unfinished, from Pemrang; 3 fragments, 1 each from Pemrang, Boldanig, and Penin, respectively). Identification of the unfinished specimen (35944) and the three fragments (32383, 33598, 36792) with the complete adze (or gouge) from Boldanig (pl. 37, f) is somewhat dubious; the unfinished piece from Pemrang is a long way from completion and the three fragments are quite small. Indeed, some or all of these latter may be fragments of a pestle rather than an adze or gouge. The complete speci- men (33961), however, is very distinctive and merits a subclass of its own, whether or not the other specimens listed belong with it. It is roughly oval in outline and plano-cor!vex in section, the thickness (2.02 cm.) being about 70 per cent of the maximum width (2.87 cm.). The sides merge into the front with no dividing line, forming a continuous parabola. The bevel likewise merges into the back, with no definite chin. The surface is completely ground, except for the poll, which shows some chipping. 2b. Slight lateral curvature: 5 (more or less complete, one badly worn; all from Pemrang). P1. 37, g-. These vary somewhat in shape but are, generally speaking, roughly trapezoidal or parabolic. One (pl. 37, j) differs noticeably from the others (pl. 37, _, h, i ): its outline is more rectangular, it has a definite chin, and the back, sides, and poll show no signs of grinding to smooth out the chip marks. The edge shows signs of chipping, perhaps from use, which would indicate that this is probably a finished tool. The other adzes lack a sharp chin and are ground on all surfaces. The sides can be differentiated from the front and are generally some- what rounded. One specimen of type 2b (pl. 37, h) has a curious rippled grooving on the back, while on the poll a folding-over of the layers suggests the flange area of some large bivalve. The details of the grooves, however, and a small, natural-seeming depression on one side near the edge do not correspond topo- graphically to the flange areas of either Tridacna or Hippopus. Type 3. Oblong: 3 (2 from Pemrang; 1 from Boldanig). P1. 37, m-o. Two of these adzes are oblong in general shape and somewhat chisellike. One specimen (pl. 37, n) from Pemrang has a clear edge-down bevel. The back shows what may be the interior grooving of Hippopus, although the specimen is not particularly translucent. One of the adzes from Pemrang (pl. 37, o) is larger and broader than the other two. No. 34356 (pl. 37, m), from Boldanig, lacks any clear bevel, the edge being formed between back and front; it is quite translucent. Type 4. Oval: 3 (2 from Pemrang; 1 from Boldanig). P1. 37, k, 1. These adzes are made from Tridacna and have an edge-down bevel like Type 1, but they appear to be made from a different part of the shell (see fig. 5). They are oval in outline. The edge, when present, is almost halfway between the back and front sur- faces-an unusual feature, which may indicate that these are axes rather than adzes. Specimen 33237, from Boldanig, is large enough to be an axe, but the others are a little small. Unfortunately, no complete finished specimen of this type was found; the adze from Pemrang shown in plate 37, 1 is un- finished, being chipped to shape but not ground, while the other two, 35180 (pl. 37, k), and 33237, lack the butt end. Type 5. Terebra adze (or gouge): 1 (from Boldanig). P1. 37, r; pl. 40, a. This distinctive specimen (33074) is made from a whole shell of Terebra maculata. The lip and one side have been ground off, exposing part of 187 ANTHROPOLOGICAL RECORDS the columella; the back and bevel are formed by this grinding. The natural surface of the shell forms the front and sides of the adze; the spire of the shell, the poll. The edge is formed by the meeting of the bevel and the front; it is therefore, by definition, an edge-up adze. The bevel is rounded, and the transition between bevel and back is gradual, with no sharp chin. Fig. 5. Tridacna shell. The outlined areas show the form of adze blades, types 1 and 4, in their relation to the whole shell. Two unfinished specimens, of unassigned type, are shown in plate 37, p and q. Two other unclassifiable specimens were also found. Specimen 32483, from Penin, is so very badly weathered that it is impossible to tell whether its shape is natural or the result of human working. If it is an artifact, it is an edge-up adze, type 2b, of markedly aberrant shape; it is roughly triangular, somewhat hollowed out on the back. Specimen 36620 is unfinished; it is roughly oval in outline, and no grinding has been done. None of the origi- nal surface of the shell remains, but the piece is quite translucent. It is impossible to identify it with any of the types above. Table 19 gives the average measurements of these five types of adzes, while table 20 presents data on their pro- portions. These figures indicate several differences be- tween the types. Type 1, for instance, varies most in size but relatively less in proportions, considering the number of specimens. Type 2a is a little longer and a little nar- rower than types 1 and 2b; it is thicker than any other type. Type 2b averages a little shorter and narrower than types 1 or 2a but is thicker than type 1. It has the widest range in proportion of any group. Type 3 is noticeably narrow and thin. Only one measurable specimen each of types 4 and 5 was recovered. Tables 19 and 20 help to indicate the re- lation of these two types to more adequately represented types. Type 4 falls within the width-length ranges of types 1 and 2b, but is thinner than either. Type 5, of Terebra maculata, is similar in proportions to type 2a, though a little smaller. The comparison of actual measurements confirms the impression of similarities and differences. Straight knives.-Short, straight-bladed, shell knives were made from large shells. The edges of some were formed by grinding, of others by longitudinal fracture or by splitting the main whorl of the shell. Twenty-two knives made from Conus shell were found, 18 at Pemrang, 3 at Boldanig, and 1 at Ruuway 2. Most of these can be identified as Conus literatus (9 knives) or BLE 19 Types of Shell Adzes: Measurements (cm.) Type Measurement 2a 2b 5 (14 specimens) (2 specimens) (5 specimens) (2 specimens) (1 specimen) (1 specimen) Length Average ........ 6.91 8.56 6.13 6.28 Maximum ...... 8.67 9.10 7.05 6.60 5.57 7.54 Minimum ....... 4.92 8.02 5.23 5.93 Range .......... 3.75 1.08 1.82 0.67 Width Average ........ 4.31 4.35 4.05 2.17 Maximum ...... 6.51 5.82 5.11 2.28 3.79 2.76 Minimum ....... 2.60 2.87 3.09 2.06 Range .......... 3.91 2.95 2.02 0.22 Thickness Average ........ 1.25 2.63 1.58 0.82 Maximum ...... 2.20 3.23 1.96 0.87 0.72 1.57 Minimum ....... 0.76 2.02 1.30 0.78 Range .......... 1.44 1.21 0.66 0.09 1 ~ . .. 188 GIFFORD: EXCAVATIONS IN YAP TABLE 20 Types of Shell Adzes: Proportions (per cent) Type Proportion 1 2a 2b 3 4 5 (14 specimens) (2 specimens) (5 specimens) (2 specimens) (1 specimen) (1 specimen) Width: Length Average ........ 62 50 67 35 Maximum ...... 82 64 90 35 66 37 Minimum ....... 47 35 49 35 Range .......... 35 29 44 0 Thickness: Width Average ........ 29 63 41 38 Maximum ...... 49 70 56 38 19 57 Minimum ....... 20 55 25 38 Range .......... 29 15 31 0 . C. marmoreus (8 knives). One specimen is C. striatus Linne and one may be C. vexillum Gmelin. The shell species of 3 pieces cannot be determined. For the species distribution by site and level, see table 21. These knives were made from the body whorl of the shell, the lip forming the cutting edge. Their form is ap- proximately that of an isosceles triangle, the cutting edge being on one of the long sides. Almost the full width of the whorl along the lip was used, a portion being removed from the narrow end. About a third to approximately half of the circumference of the whorl was used, including the lip, which was ground down to function as the cutting edge. To judge from well-preserved specimens, this hard shell could be ground to an almost razor-sharp edge. On most specimens, the grinding was extensive enough to remove all traces of the natural edge. In fact, on many the edge has been ground so far back that it has a concave outline, TABLE 21 Distribution of Conus Knives Site Species Various Total and level C. literatus . marmoreus Ruuway 2 Surface 1a 1 Boldanig 6 in... 12 in... 18 in ... 1 1 24 in... 2 2 Pemrang 6in ... 3 3 12 in ... 3 2 lb 6 18 in ... 2 4 1 7 24 in ... 1 1 36 in 1 1 aVexillum. bStriatus. suggesting that these are worn-out specimens beyond fur- ther sharpening. Apart from the cutting edge, the other edges show little trace of careful working. In most speci- mens, the piece of shell appears to have been simply bro- ken to shape. The unusual specimen (35625) from Pem- rang (30-36 in.) comprises about two-thirds of the body whorl of an unidentified Conus, with one edge filed to a sharp cutting edge. It is figured in plate 40, m. Some of the knives (e.g., 35086, 36892) show evidence of battering at the suture edge of the body whorl. At this point, the cone-shaped body whorl meets the disk-spiral of the remaining whorls, forming, on the unmodified shell, a rather well-defined acute angle. This angle would be sus- ceptible to battering and wear under natural conditions, so the battering visible on the knives may not all be truly the work of man. On one specimen (36681) the narrow edge of the whorl appears to have been cut or ground to a more or less straight line rather than simply broken. The following list gives site and depth provenience of each of the catalogued Conus shell knives in our collection. Conus literatus: Boldanig, 12-18 in., 33945, pl. 38, b; Pemrang, 0-6 in., 34835; Pemrang, 0-6 in., 34907; Pemrang, 12-18 in., 35086; Pemrang, 12-18 in., 35971; Pemrang, 0-6 in., 36681; Pemrang, 6-12 in., 36697, pl. 38, a; Pemrang, 6-12 in., 36892; Pemrang, 6-12in., 36893. Conus marmoreus: Boldanig, 18-24 in., 33653; Boldanig, 18-24 in., 33971, pl. 38, c; Pemrang, 12-18 in., 35086; Pemrang, 12-18 in., 35088; Pemrang, 12-18 in., 35870 (2 specs.); Pemrang, 6-12 in., 36821; Pem- rang, 6-12 in., 36891. Conus striatus: Pemrang, 6-12 in., 36896. Conus vexillum: Ruuway 2, surface, 31989. Conus sp.: Pemrang, 30-36 in., 35625, pl. 40, m; Pemrang, 12-18 in., 35973; Pemrang, 18-24 in., 36485. The depth distribution of these knives suggests that their use is not very ancient on Yap. Only one specimen (35625) was found below 24 in.; of the 21 others, all but 3 are from depths of 18 in. or less. In addition to the 22 obviously worked specimens de- scribed above, three fragments of the lip area of Conus shells were found at Boldanig. Although these show no 189 ANTHROPOLOGICAL RECORDS signs of a ground cutting edge, they may have functioned as knives, for the natural lip edge is sharp enough to pro- vide a cutting edge. One of these (33653) is C. marmoreus, a striplike fragment 6.5 cm. long and 0.65 to 1.1 cm. wide found at a depth of 18-24 in. The other two are C. literatus. Specimen 33782, from a depth of 54-60 in., is a roughly triangular fragment, 4 cm. along the natural lip edge and 2 cm. wide at the widest point; 33744 is a strip, 4.8 cm. long and 0.9 to 1.3 cm. wide, from a depth of 42-48 in. These last two are from a greater depth than any others, but they may be only natural fragments rather than artifacts. Taro peelers.-Nine objects identified by Yapese informants as taro peelers were excavated at Pem- rang. Crew members suggested that these objects had been imported from Ngulu, an atoll with which Giror village has close ties. There are three types of peel- ers, two made from shells of Cassis cornuta Linne, the third from Charonia tritonis Linne. They appear to be contemporaneous and recent. None were found below 12 in. Type 1. Cassis cornuta. "Coat-hanger"-shaped. 6 specimens. This type has a general outline similar to that of a wooden coat-hanger. These peelers are made from the lip area of Cassis cornuta. In Cassis, the enamel of the inner lip grows out, forming a thin enamel layer over the rough exterior canal surface. However, near the canal this thin layer does not seem to adhere closely to the old surface, but may be broken away to reveal the old exterior. It is from this area that the type 1 taro peeler is made. The "neck" of the coat-hanger points toward and may include part of the convolutions formed by the natural orifice near the canal. The cutting edge is ground across the "bar" of the coat-hanger, that is, more or less transversely across the shell. The edge is ground from the outside of the shell only. Where the overlapping inner lip does not ad- here closely, it is broken or ground away, reveal- ing the old outer surface. In some specimens, prob- ably worn-out pieces, the cutting edge has been ground back until the blade is only a centimeter or two wide. In these specimens there is a decided slant to the cutting edge; if the shell exterior faces the observer, this slant is up to the right. Of the six specimens of this type, three are per- forated (pl. 38, f-h), the hole being approximately in the middle. In two (pl. 38, g, h) the hole, biconi- cally drilled, is circular, 0. 63 cm. and 0. 37 cm. in diameter respectively. The other perforated specimen (34955) has an oval hole, 2.89 cm. long and 1. 14 cm. wide, parallel to the cutting edge, ground from the outside. These three pieces all have cut or ground edges. The other peelers are not perforated but are merely broken to shape, ex- cept for the cutting edge. Plate 40, b shows the largest of the three unperforated pieces. Type 2. Cassis cornuta. Triangular. 2 specimens. P1. 40, c, d. These two taro peelers are also made of Cassis cornuta, but, unlike the preceding type, they were cut from a portion of the main whorl and do not include any part of the inner lip. However, they were probably cut from an area closer to the lip than to the apex of the spire, judging from the configuration of the outer surface as compared to a whole specimen of the shell. Like type 1, the edge was ground transversely across the shell, the corner opposite the cutting edge being toward the lip. The cutting edge was ground from the exterior of the shell. One of these (pl. 40, c) is perforated by a biconically drilled hole, 0. 63 cm. in diameter, in the middle. This specimen has one ground edge (the right-hand one, if the cutting edge forms the base of the triangle), while the remaining edge is broken to shape. The other piece (pl. 40, d) has no hole; the left-hand edge is slightly ground. The cut- ting edge has a slightly concave outline, perhaps from re- sharpening. Type 3. Charonia tritonis. 1 specimen. Pl. 40, e. This specimen consists of the spire portion, about 11 cm. long, of a Charonia tritonis shell. A little less than half the circumference has been neatly broken away to 4. 5 cm. from the tip. A cut- ting edge has been ground on the remainder of the circumference, more or less parallel to the natural suture, but concave in outline. The edge is ground from the exterior of the shell only. Type 1. Cassis cornuta: Perforated: Pemrang, 0-6 in. 34879, pl. 38, h; Pemrang, 0-6 in., 34955, pl. 38, f; Pemrang, 6-12 in., 35477, pl. 38, g. Unperforated: Pemrang, 0-6 in., 35434; Pemrang, 6-12 in., 35474, pl. 40, b; Pemrang, 0-6 in., 34903. Type 2. Cassis cornuta: Perforated: Pemrang, 6-12 in., 36822, pl. 40, c. Unperforated: Pemrang, 0-6 in., 36793, pl. 40, d. Type 3. Charonia tritonis. Pemrang, 6-12 in., 35478, pl. 40, e. Paring knives.-These are whole univalve shells with a circular or oval opening made in the body whorl by grinding or otherwise, thus producing sharp edges (cf. Gifford, 1951, fig. 1, a; Gifford and Shutler, 1956, pl. 7, m, pl. 8, h, ac, ad). These shells are used for paring fruit or vegetables, the peeling passing into the shell through its side opening and out its natural mouth. The Fijian specimen cited above (Gifford, 1951) is an ethno- logical piece, which was used as a breadfruit peeler. It is made of Fasciolaria filamentosa, as are the five from Pemrang. Seven specimens of this type were found, including two pieces tentatively identified as paring knives Five are from Pemrang, one is from Wolom and one from Boldanig. Each consists of a whole univalve shell with a hole broken or ground into the main whorl. Those from Pemrang (pl. 40, f ) all show considerably more wear on the side with the hole than on other surfaces. One of these pieces is badly broken and is included only tentatively. The paring knife from Wolom (pl. 40, g), which also shows wear, is Turbo argyrostomus. The one from Bol- danig, Strombus luhuanus (pl. 40, h), is listed dubiously, since the hole is small. This piece does not show much wear on the hole side. The stratigraphic distribution of this type is indicated in the list below. The specimens from Boldanig and Pemrang are not from any great depth, but the one from Wolom was found below 60 in. Fasciolaria filamentosa: Pemrang, 0-6 in., 34877; 0-6 in., 34878, pl. 40, f; 0-6 in., 34904; 0-6 in., 35445; 18-24 in., 35573. 190 GIFFORD: EXCAVATIONS IN YAP Turbo argyrostomus: Wolom, 60-66 in., 34781, pl. 40, g. Strombus luhuanus: Boldanig, 24-30 in., 34315, pl. 40, h. Scrapers.-A valve of Arca scapha (pl. 40, i), with abraded edge opposite the umbo, appears to have been a scraper, although it has a perforation at the umbo. It is from a depth of 6-12 in. at Boldanig. Three valves from Pemrang (2, numbered 37826, from the surface; the third, 36477, from a depth of 18-24 in.) seem to have been used as scrapers also. Bracelets and rings.-Seven fragments of shell brace- lets and rings were found, all at Pemrang. The shell species of very few of these can be identified, since in most of them the natural surface of the shell has been ground off. Most are probably Conus sp.; plate 38, 1 re- tains traces of what appear to be the markings of C. li- teratus, while plate 38, j seems to be a different species, perhaps C. striatus. One whole C. literatus shell (pl. 40, i) from the surface at Pemrang, shows a 1-mm. groove encircling the shell 2 cm. below the shoulder of the main whorl, presumably a step toward making a brace- let. This piece recalls a New Caledonian specimen (Gifford and Shutler, 1956, pl. 13, k). Each of these fragments consists of an arc of a circle, apparently made by cutting a horizontal strip about 2 cm. wide from the widest part of the main whorl of the shell, parallel to the suture. This process would yield a spiral rather than a true circle, with the ends overlapping. If, however, a Conus shell was used, the ends of the spiral might have remained con- nected on one side, where the main whorl joins the lesser whorls to form the flat spine of the shell. Most of the specimens are very well finished, having been ground to a smooth surface and a regular shape. The cross sections show various shapes and dimensions. The seven bracelets and rings from Pemrang are described individually below. Inner diameters are esti- mated from the degree of curvature. 35682. Depth 48-54 in. Inner diameter, ca. 0.5 cm.; thickness, 0. 75 cm.; width, 0. 70 cm. Cross section: pentagonal, the apex forming the outer equator of the bracelet. 36165. Depth, 42-48 in. Inner diameter, ca. 7.5 cm.; thickness, 0. 66 cm.; width, 0. 88 cm. Cross section: trapezoidal, with slightly curved sides, the widest side towards the interior. 36201. Depth, 48-54 in. Inner diameter, ca. 7.5 cm.; thickness, 0. 67 cm.; width, 1. 21 cm. Cross section: triangular, with slightly curved sides, the apex forming the external equator of the bracelet. P1. 38, k. 36542. Depth, 36-42 in. Inner diameter, ca. 4 cm.; thickness, 0.49 cm.; width, 1. 11 cm. Cross section: trapezoidal, with slightly curved sides, the longest side towards the wrist. Pl. 38, 1. 36676. Depth, 0-6 in. Inner diameter, ca. 5 cm.; thickness, 4. 2 cm.; width, 0. 74 cm. Cross section: lenticular. 36712. Depth, 12-18 in. Unfinished. Inner diameter ca. 6.5 cm.; thickness, 0.45 cm.; width, 1. 38-1. 73 cm. The widest diameter edge has been ground to about a 450 bevel, while the other edge is simply broken. The inner and outer surfaces show no signs of working. Pl. 38, j. 36834. Depth, 12-18 in. Inner diameter, ca. 6 cm.; thickness, 0. 26-0. 32 cm.; width, 1. 28 cm. Cross section: subrectangular; the side towards the wrist is flat with sharp corners; the outside has a wide, shallow central groove between rounded ridges at the sides. P1. 38, i. Not all of these circlets could have been worn on the wrist of an adult. It is doubtful whether a grown person could put on a bracelet with an inner diameter of less than 6 cm. Three of these seven specimens (35682, 36542, 36676) were smaller than this, while 36834 is on the borderline. The grooved C. literatus (pl. 40, j), presumably intended for a bracelet, has an external diameter of only 5.2 cm. Some rings may have been worn as ear pendants (cf. Muller, 1917, fig. 17, p. 24). Although bracelets were excavated only at Pemrang, they were there distributed within a considerable range of depth, down to 54 in. Their use at this site, there- fore, is presumably of some antiquity. At Merur village, Tomil, Magistrate Roboman gave us a man's arm ornament or bracelet, a modern piece consisting of a nearly whole Conus literatus shell (pl. 38, w). The disklike spire of the shell has been cut out to provide a circular opening. A portion of the body whorl opposite the lip has also been cut to form a lateral opening 3. 6 cm. wide, extending from a point 2. 7 cm. from the apical end to the canal end of the natural orifice. This end has been truncated, the brace- let thus having the form of a truncated cone, 13 cm. long, 9 cm. maximum diameter, and 3. 5 cm. minimum diameter. The mouth of the shell has been enlarged to 3. 6 cm. wide and 9.9 cm. long. The natural surface of the shell has been ground off, except inside the lip. All surfaces are smoothly ground and neatly finished. The internal diameter of the circular opening is 6.5 cm. Miller illustrates a similar specimen (1917, fig. 35, p. 30). The only method of putting this bracelet on an adult hand appears to be as follows: the hand is inserted through the circular aperture and out through the side slit, keeping the thumb on the side away from the slit; that is, the fingers go through the slit before the thumb. It goes on most easily on the right hand, but can be put on the left. In position on the arm, it lies some- what obliquely, the lip of the shell being toward the palm on the left arm and toward the elbow on the right arm. Disks.-A disk (pl. 40, k) 35 by 37 mm. in diameter, apparently of Tridacna or Hippopus shell, was excavated at a depth of 24-30 in. at Boldanig. It has a maximum thickness of about 5 mm. in the middle and is a bit thinner on the edges. We learned nothing about its probable use. A second disk (pl. 40, 1), 61 to 66 in. in diameter and about 10 mm. thick in the center, tapers to a thin periphery. It was excavated at a depth of 24-30 in. at Pemrang. It is also probably Tridacna. Disk beads.-Nineteen shell beads were found at two of the five sites, Ruuway and Boldanig. Their absence at the rich Pemrang site is difficult to understand. The beads are mostly Spondylus or Lambis (cf. Muller, 1917, pl. 12). Drilling is usually biconical but in some beads perforation from one side penetrated all the way through. Identification of the shell species used in beads is often difficult. Of the 19 beads, 12 appear to be Spondylus 191 ANTHROPOLOGICAL RECORDS (4 are illustrated in pl. 38, m, n, r, u). Five, which appear to be Lambis lambis, are more or less wavy disks because of the conformation of the portion of the shell from which they are cut (pl. 38, p, q, s, t, v). These Lambis beads have a larger diameter than those of Spondylus. Two beads (33141 and 33733) are yellowish and appear to be clamshell, possibly Hip- popus hippopus; one is shown in plate 38, o. Spondylus: Ruuway 2, 6-12 in., 32112, pl. 38, m; Ruuway 2, 12-18 in., 32135, pl. 38, n; Ruuway 2, 0-6 in., 32295; Ruuway 2, 12-18 in., 32351; Boldanig, 18-24 in., 33016; Boldanig, 30-36 in., 33072a, pl. 38, r; Boldanig, 30-36 in., 33072b, pl. 38, u; Bolda- nig, 42-48 in., 33140; Boldanig, 24-30 in., 33667b; Boldanig, 48-54 in., 33755; Boldanig, 36-42 in., 34038; Boldanig, 18-24 in., 33960b. Lambis lambis: Boldanig, 18-24 in., 33646, pl. 38, s; Boldanig, 24-30 in., 33667a, pl. 38, p; Bolda- nig, 18-24 in., 33959, pl. 38, v; Boldanig, 18-24 in., 33960b, pl. 38, q; Boldanig, 30-36 in., 34010, pl. 38,t. Hippopus hippopus (?): Boldanig, 42-48 in., 33141, pl. 38, o; Boldanig, 42-48 in., 33733. Perforated univalves.-We excavated a few whole univalve shells with either drilled or cut holes in the body whorl. The 4 drilled shells were either Turbo ticaonicus or Nerita undata: 34745, Turbo ticaonicus, Wolom, 48-54 in., pl. 40, n; 34762, Turbo ticaonicus, Wolom, 54-60 in., pl. 40, o; 35356, Nerita undata, Pemrang, 66-72 in., pl. 40, p; 35417, Nerita undata, Pemrang, 84-90 in., pl. 40, q. There are three Cerithium nodulosum shells with holes cut or chopped in them: 34091, Boldanig, 48-54 in., pl. 41, a; 34223, Boldanig, 0-6 in.; 36115, Pem- rang, 30-36 in., pl. 41, b. The fact that the drilled shells are from greater depths than those with cut holes is interesting. The Wolom and Boldanig excavations were not less than 50 yds. apart, so these specimens may be regarded as from one locality. Perforated bivalves.-More than a dozen bivalve shells, with either drilled or cut holes, were excavated. Six species were found: Arca scapha, 7 specimens; Cardium mindanense, 3; Gafrarium tumidum, 2; Pinctada mar- garitifera, 1; Ostrea sp., 1; Spondylus radians, 1. The two Gafrarium tumidum shells have drilled holes near the edge of the valve; one is shown in plate 41, g. The Pinctada margaritifera specimen (pl. 41, i) with a drilled hole is a thin fragment of a larger piece, probably money. The Ostrea piece (pl. 41, h) is a small valve with a drilled hole, and the Spondylus radians (pl. 41, j) has two holes drilled near the edge opposite the umbo. All of the Arca scapha and Cardium mindanense shells have cut holes near the unmbo. These holes vary in size, as the illustrations (pl. 41, c-f) show. The Arca scapha pieces are reminiscent of the net sinkers of New Caledonia (Gifford and Shutler, 1956, pl. 3, a). The Yapese crew said that these shells were not used as net sinkers in Yap. The occurrence of these perforated Arca scapha and Cardium mindanense shells in Yapese archaeological aites is given below. Illustrated specimens are indicated. Arca scapha: Ruuway 2, 18-24 in., 32171; Penin, 0-6 in., 32650; Boldanig, 6-12 in., 33911; Pemrang, 0-6 in., 34888, pl. 41, d; Pemrang, 48-54 in., 35262, pl. 41, c; Pemrang, 54-60 in., 35298; Pemrang, 24- 30 in., 36066. Cardium mindanense: Penin, 0-6 in., 32675, pl. 41, e; Boldanig, 78-84 in., 33837; Pemrang, 48-54 in., 35690, pl. 41, f. Drilled shells of four species of bivalves were ob- tained at Boldanig-Wolom. Gafrarium tumidum: 84-90 in., 33851, pl. 41, j; 42-48 in., 34723. Ostrea sp.: 0-6 in., 33303, pl. 41, h. Pinctada margaritifera: 18-24 in., 33393; 0-6 in., 33875, pl. 41, i. Spondylus radians, 12-18 in., 34561, pl. 41, i. Conus cap.-The spire top of a Conus literatus shell (pl. 41, k), found in the 12-18 in. block of Pemrang, resembles a New Caledonian disk cut from the top of a Conus shell (Gifford and Shutler, 1956, pl. 8, i). The Pemrang specimen is not ground as smooth on the top and bottom as the one from New Caledonia. Pestles.-Four large pestles of shell and coral lime- stone from our collection are shown in plate 34, a-d. Three of these (a-c) were obtained from Yapese owners. The fourth (d) was collected on the surface at the shrine of the iguana god, Magrogoy, in Tomil municipality. The pieces shown in plate 34, a and b are taro pounders of coral limestone, the first presented by Magistrate Roboman at Merur village, Tomil munici- pality, the second purchased at Kanif village, Dalipebi- nau municipality. The slender pestles shown in plate 34, c and d are of shell, either Tridacna or Hippopus. The smaller specimen (c) is a pestle for areca nut. It was given us, along with a bamboo mortar, by Loochaz, a crew member. These pestles are used by elderly people unable to chew the whole areca nut along with the betel leaf and lime. Shell pestles for areca nuts, as well as wood and bamboo mortars, are pictured by Muller (1917, figs. 92-94, p. 63). The large, incomplete pestle shown in plate 34, d is in three pieces, with one or more other sections lacking. It seems unduly heavy for an areca pestle. We found it lying on the ground, along with many pieces of pearl-shell money, at the shrine of the god Magrogoy in Tomil. Thus it may well be a piece of money, like that pictured by Muller (1917, figs. 174 and 175, p. 127). A fifth pestle, a short shell object which the crew members thought was for pounding medicine, is illus- trated in plate 41, m. It was found on the surface of Rugog's grave at Teb, Tomil municipality. This pounder is short and heavy, with a drilled hole in one side. The hole does not pass through the piece but penetrates about halfway at the point where the narrow diameter of the pestle is approximately 50 mm. Pottery smoothing tools.-On the basis of the pottery smoother of Cypraea tigris shell (pl. 38, e) used by Finek, we have tentatively classified as potters' tools 11 specimens recovered in excavation. One is illus- trated in plate 38, d. All are Cypraea tigris shells except one, which is a large Cypraea arabica (33995). These tools consist of whole shells from which a large, discoidal portion of the body whorl has been 192 GIFFORD: EXCAVATIONS IN YAP removed by simple fracture. The large round hole thus obtained, opposite the natural aperture of the shell, renders the columella accessible to serve as a grip. The curved or somewhat flat surfaces on either side of the natural aperture may then be used to smooth the surface of pottery. For comments on ethnographic use see section on "Pottery Making"; also Muller (1917, p. 124). Presumably our cowry-shell pottery smoothers-if this is what they are-are used in the making of lami- nated pottery. Seven of the specimens are from depths down to 18 in.; 4 are from depths of 24-42 in. The archaeological specimens were found at the following sites: Penin, 1; Boldanig, 4; Pemrang, 6. Cypraea tigris: Penin, 0-6 in., 32602; Boldanig, 12-18 in., 33631; Boldanig, 6-12 in., 33921;Boldanig, 12-18 in., 33946; Pemrang, 0-6 in., 34910; Pemrang, 0-6 in., 34967 (2 specimens); Pemrang, 30-36 in., 35196 (2 specimens); Pemrang, 36-42 in., 36150, pl. 38, d. Cypraea arabica: Boldanig, 24-30 in., 33995. Money.-The Yapese interest in various types of currency has been well set forth in ethnological ac- counts, for instance, by Furness (1910, pp. 126-133) and Muller (1917, pp. 126-127). Plate 35 shows both archaeological and ethnological pieces of pearl-shell (Pinctada margaritifera) money. The ethnological specimens are easily recognized by the attached sinnet cords (pI. 35, a, d); the piece shown in plate 35, c is also ethnological, though lacking the cord. Plate 35, b shows one of the 82 specimens col- lected on the surface at Magrogoy's shrine. Two drilled fragments were collected at Boldanig, one (33875) at a depth of 0-6 in., the other (33260) at 12-18 in. Two cut fragments (32030, 0-6 in. at Ruuway 2; 36832, 12-18 in. at Pemrang) are probably pieces of shell money. A perforated fragment from Boldanig is illustrated in plate 41, i. Mflller pictures shell money in his plate 36. Trumpets.-Seven trumpets of Charonia tritonis shell were collected, but only two from underground: one (pI. 36, b) from a depth of 6-12 in. at Boldanig, the other (34901) from 0-6 in. at Pemrang. Two more (34511, 34512) were obtained on the surface at Bol- danig; 34512 (pl. 36, c) has a drilled hole in the lip. Three others are from the surface of the shrine of the god Magrogoy at Teb in Tomil municipality. One of these had been collected earlier by the magistrate, Roboman, who had inserted a cord in the drilled lip. Plate 36 pictures four of the seven trumpets we collected. Only two had the lip perforated for a cord by which the trumpet could be suspended. In each, of course, one of the middle whorls had been perforated for the mouthpiece. BONE ARTIFACT The only bone artifact (35291) in our collection is a fragmentary perforated disk of turtle bone (pl. 38, x) found between 54 and 60 in. at Pemrang. The Yapese crew suggested that it might be an imitation of stone money, but we think it may rather be the forerunner of stone money. No evidence of stone money was found below the surface anywheire. Consequently, if this bone disk has any relation to stone money, disk has any relation to stone money, which we doubt, it is probably a forerunner. STONE ARTIFACTS Adze.-Except for the conspicuous stone money disks quarried overseas in Palau and latterly in Guam, stone artifacts are rare in Yap. No stone tools that may have been used in quarrying were found, unless it is the small stone adze (32379), the only complete one in our collection. This was picked up on the sur- face at Penin, close to the edge of trench line Z1-Z2. This piece is rectangular in section and trapezoidal in outline, with a rounded poll. There is very little difference between front and back, but the front is slightly convex. The edge is formed between the front and the bevel. The chin is not very clearly marked. Professor Howel Williams has identified this stone as probably amphibolite (or hornblende schist), composed of green hornblende and feldspar. The dimensions are: maximum length, 7. 51 cm., maximum width 5.26 cm., maximum thickness 1.28 cm. The maximum proportions are as follows: width, 70 per cent of length; thickness, 24 per cent of width. See plate 37, s. Disk.-A flat disk of white stone (34086), from a depth of 48 to 54 in. at Boldanig, is illustrated in plate 41, 1. The stone, according to Professor Francis J. Turner, Department of Geology, University of Cali- fornia, is perhaps phosphate. The disk measures 34 mm. in diameter; its thickness varies from 3 mm. on one edge to 5 mm. on the opposite edge. One surface has a chalky white coating. The edge looks as though it had been first chipped and then ground, at least a little. There is no evidence of any attempt to perforate the center. Stone money.-It was hoped that our excavations would throw light on the time of origin of the unique stone disk money for which Yap is famous, but they provide~ no such evidence. Indeed, we found no buried money whatsoever. Three small pieces of stone money were presented to the expedition. (See pl. 39.) Our excavations did produce a few tiny fragments of calcite that we thought might be money stone, but some of this may, actually, be shell. These pieces were examined by Professor Howel Williams, but the results were inconclusive, since calcite and aragonite (a form of calcite with different crystalline arrange- ment) occur in both inorganic and organic forms. The following quotation from a letter, dated March 7, 1957, from Dr. Charles G. Johnson of the U. S. Geological Survey, bears on the problem. As you undoubtedly know, the Yap "stone money" was carved from dripstone in the limestone caves of Palau and Guam. There are no limestone caves on Yap and, therefore, no dripstone. If your frag- ments are large enough to allow positive identifica- tion as dripstone, you can be certain they came from Palau or Guam and probably are chips from stone money. The fragments of calcite you found in the middens might be from several sources. I do not know of any veins of calcite or aragonite on Yap that are large enough to furnish stone for the Yap stone money. There are a few veins of calcite up to 2 inch in width that could furnish small 193 ANTHROPOLOGICAL RECORDS fragments of crystalline material. Fragments ob- served up to 6 inches in diameter of dense, white, finely crystalline limestone are present, though very rare, in the breccia that underlies much of eastern Gagil-tomil, Map, and Rumung. The lime- stone of Garim Island, located out on the reef off the east side of southern Yap, and the two small limestone mushroom stacks off the southern tip of Yap could furnish large limestone fragments; these can easily be recognized under a microscope by their being composed of detritus of coral, shells, calcareous algae, and foraminifera. I can think of but one other possible source of calcite or arago- nite fragments and that is the heavy shelled mol- lusks. The shell of the giant Tridacna is com- pletely crystallized in the massive umbo and hinge regions. This is also true of the larger Hippopus and Spondylus shells. When these shells are broken up, shattering frequently occurs along cleavage planes so many of the pieces have several smooth plane surfaces much like crystal faces. This last appears to me to be a likely possible source of small calcite or aragonite fragments that are pres- ent in the kitchen middens. We saw the stone money at all the villages we visited; there are many hundreds of pieces in Yap. The so-called "Money Bank" on both sides of the road on the shore of Tomil harbor, near Balebatt village, is a good example of the manner of placing the money. Sogon volunteered information about two pieces here. A medium-sized piece on the shore side of the road directly opposite the foot of the stone- paved path leading upslope to Ruuway is called Tumou, after the man who voyaged by canoe to Palau and brought it back. A large piece at the foot of the path belonged to Sogon's father's father and was brought over on O'Keefe's sailing vessel. There is a huge piece, more than 10 ft. in dia- meter, at Uanajan, near Gaychapar, Gagil munici- pality. A piece was pointed out that had been quar- ried in Guam and had also been brought over on O'Keefe's ship. The stone on Guam was quarried only after access to the source in Palau was closed to the Yapese. The Guam stone proved unsatisfactory, so the quarrying came to an end. Muller (1917, pp. 126-133) and Furness (1910, chap. 7) discuss various kinds of money used in Yap. "Toy stone money."-A fist-size piece of soft stone with a hole drilled through it was found at Penin at a depth of 0-6 in. in cookhouse 1. This was said to be a child's toy stone money (pl. 38, z). A flat per- forated piece of hard stone, which may also be toy money, was found in the 0-6 in. block at Pemrang (pl. 38, y). The Yapese crew members had no idea what it might be. Christian (1899, p. 319) relates the story of the inven- tion of stone money, which is attributed to a Tomil man named Anagumang. Mlller (1917, pl. 60, fig. 2) pictures children's toy money made of sea-urchin shells. Whetstone, pestle, and hammerstones.-Plate 24, c, shows a large grindstone at Cheen, within the limits of Teb village in Tomil municipality, described in the account of Walgom. It may have been used to grind adze blades. A small whetstone fragment (pl. 41, n) was excavated at a depth of 0-6 in. at Penin. It is convex on top, flat on the bottom. The distal end of a small stone pestle (pl. 41, o) was excavated in the 30-36 in. block at Boldanig. It bears some resemblance to the short pestles illustrated by Spoehr (1957, fig. 70). Seven possible hammerstones were excavated. Boldanig: 24-30 in., 33417 (2 specimens), wts., 1.5 oz. and 3 oz.; 42-48 in., 33734, wt. 4.5 oz.; 42-48 in., 33735, wt. 6 oz.; 24-30 in., 33988, wt. 10 oz.; 42-48 in., 34386, wt. 28 oz. Pemrang: 48-54 in., 35260, wt. 25 oz. All these are natural cobbles, small enough to be held in the hand, and all show battering from use. None is illustrated. Quartz strike-a-light.-The present-day Yapese know how to produce sparks by striking quartz fragments against an iron file. It is likely that quartz specimens from Ruuway and Penin were so used. The quartz pieces found at some depth at Penin are from a hillside garbage dump and may have been buried under slides. Therefore they may not be as ancient as the depths at which they were found would seem to indicate. The specimens from Boldanig and Pemrang come from so great a depth that we cannot assume that they were necessarily used for striking fire. Iron tools were not found at these depths. There are, however, deposits of iron ore in Yap, and it is possible that iron was used for tools. The following quartz specimens were collected. Ruuway 1, 0-6 in., 31973. Penin: 0-6 in., 32495, 32509, 32530, 32556, 32599; 6-12 in., 32578, 32620; 12-18 in., 32732. Boldanig, 54-60 in., 33196. Pemrang, 48-54 in., 35688. Mineral pigments.-Lumps of mineral pigment, red mostly but some yellow, were found in excavations. Yapese crew members said this substance was used as paint, the red in particular being used on canoes. Muller (1917, p. 117) gives an account of the use of pigment. Lumps of soft stone pigment, identified by Professor Howel Williams as goethite with traces of hematite, were found at several sites. The depths at which they occurred at Boldanig and Pemrang indicate that they were used in very ancient times. The specimens were distributed as follows: Boldanig: 12-18 in., 33932; 42-48 in., 33494; 48-54 in., 33754; 60-66 in., 34444; 66-72 in., 34460; 72-78 in., 34477. Pemrang: 6-12 in., 36692, 36819; 12-18 in., 35942, 36839, 36432, 36711; 18-24 in., 35999, 36001, 36730; 24-30 in., 36058, 36504, 36747; 30-36 in., 36104; 36-42 in., 35641, 36135; 42-48 in., 35658, 36168; 48-54 in., 36202; 78-84 in., 36656. Walgom: 0-6 in., 36903; 12-18 in., 36953; 18-24 in., 36925, 36955; 24-30 in., 36932. Rugog's grave: 32 in., 36971. 194 RADIOCARBON DATES Seven samples of charcoal, the only ones of suffi- cient size in our collection, were measured by Pro- fessor H. R. Crane, Randall Laboratory of Physics, University of Michigan. Only two samples from the deeper levels of the excavations were large enough for carbon- 14 measurement. The dates obtained are listed in table 22. Four levels (48-54 in., 54-60 in., 60-66 in., 66-72 in.) were combined for sample M- 634; the other samples are from 6-in. levels, except the charcoal from Rugog's grave, obtained at 20 in. An eighth sample (our sample 11, Crane's sample M-769) is omitted from the following discussion be- cause of its mixed character. It is composed of four lots of charcoal from four separate pits and from depths ranging from 42 to 60 in. at Boldanig-Wolom. It yielded a date of A.D. 1857, which is entirely out of line with the other dates from this site. TABLE 22 Radiocarbon Dates U.C. Crane ie Depth Age Middle No. No. Sie (in.) (yrs. before date present) (A.D.) 1 M-626 Rugog's grave 20 200 + 200 1756 4 M- 629 Penin 24-30 200 + 200 1756 6 M- 631 Boldanig 36-42 320 + 200 1636 10 M- 791 Boldanig- Wolom 60-66 1110 + 200 847 7 M-632 Pemrang 18-24 250 + 400, 1706 - 250 8 M- 633 Permrang 24-30 100 + 200, 1856 - 100 9 M-634 Pemrang 48-72 1780 + 250 176 No samples of sufficient size were obtained from the two northernmost sites, Walgom in Tomil munici- pality, and Ruuway in Rull municipality. The sample from Rugog's grave, however, was obtained perhaps a mile from Walgom. Rugog, the Noah of Yap mythology, lived only 200 years ago, according to the charcoal from his grave. The only dates in our table that are reversed and not in harmony are those of the two upper Pemrang samples, 7 and 8. This fact, however, is perhaps not too significant, since sample 7 (Crane's M-632) was dug in rectangle A10-11 Z10-11, whereas sample 8 (M-633) was derived from four other rectangles, ex- cept for a negligible amount of 0.2 gm. from A10-11 10-11. The dates within the standard deviations of amples 7 and 8 overlap considerably. With the ex- eption just cited, both Boldanig and Pemrang dates are consistent with the depths from which the char- oal samples were derived. Boldanig dates range from A.D. 847 to A.D. 1636, [ 195 corresponding with depths of 36-42 in. and 54-66 in. respectively-practically 800 years for a possible 30 inches of accumulation. At Pemrang the dates of A.D. 176 and A.D. 1856, correspond with depths of 48-72 in. and 24-30 in., respectively-practically 1,600 years for a possible 48 inches of accumulation. Obviously, the rate of accumulation was slower at Pemrang than at Boldanig: 4 feet in 1,600 years at Pemrang, 2.5 feet in 800 years at Boldanig; one foot in 400 years at Pemrang, one foot in 320 years at Boldanig. These figures suggest intermittent rather than continuous use, especially in view of the accumu- lation at Penin, 2.5 feet in 200 years, more than a foot in less than a century. All five sites were apparently occupied into historic or at least protohistoric times. Rugog's grave also is dated in this recent period. The only dates which lie completely outside this range are A.D. 176 for the lower levels at Pemrang, and A.D. 847 for those at Boldanig. In summary, the radiocarbon dates obtained from Yap suggest a considerable antiquity for the oc- cupation of the island and also for pottery- making and the working of shell. Since the dates of A.D. 176 at Pemrang and A.D. 847 at Boldanig are based on a combination of charcoal from depths of more than two 6-in. blocks (54-66 in. at Boldanig, 48-72 in. at Pemrang), there is every reason to believe that the beginning of the deposit was much earlier than this. An increasing number of Neolithic radiocarbon dates for the Oceanian area is being published and the dates seem to be, in general, harmoniously related, as indi- cated by the earliest dates recorded for several places: Saipan, Marianas Islands (Spoehr, 1957, p. 177) ........................ 1527 B.C. Hoifung, Hong Kong area (Broecker, Kulp, and Tucek, 1956) .............. 1175 B.C. New Caledonia (Gifford and Shutler, 1956, p. 89) ......................... 847 B.C. Masbate Island, Philippines (Broecker et al., 1956) ....... ................. 754 B.C. Viti Levu Island, Fiji (Gifford, 1955) .... 46 B.C. Yap Island ......... ................... A.D. 176 Oahu, Hawaiian Islands (Spoehr, 1957, p. 177) ........ ................ A.D. 1004 Spoehr (1957, p. 177) has presented his ideas of the significance of the dates from the Marianas, Hawaii, Fiji, and New Caledonia in relation to Malayo-Poly- nesian movements eastward into the Pacific. Presum- ably the dates from Yap and Masbate also pertain to speakers of Malayo-Polynesian. Naturally, however, there is an inherent risk in attributing unwritten lan- guages to the prehistoric inhabitants of areas where certain tongues are spoken today. Probably the risk is not too great, but we would hesitate to name the language of the Hoifung people in 1175 B.C. Possibly, too, the New Caledonians of 847 B.C. did not speak Malayo-Polynesian, considering the languages used to- day in parts of Melanesia which are not Malayo- Polynesian. I LOCAL DISTRIBUTIONS IN SPACE AND TIME Yap is so tiny a speck of land (39 sq. mi.) that we are uncertain about the significance of the occurrence or absence of artifacts in our five sites. With such a small area it is difficult to conceive of local develop- ments, except as features of the caste system. Thus, in modern times the making of laminated pottery was a specialty of the low-caste village of Gitam, but lami- nated pottery was used in every village. At both Boldanig-Wolom and Pemrang the amount of unlaminated, tempered pottery (Marianas Plain) in- creases with depth, whereas the amount of laminated, untempered pottery decreases. At Boldanig-Wolom un- laminated pottery constitutes 1 to 35 per cent of the total weight in the 0-42 in. level, 14-61 per cent in the 42-66 in. level, and 61-100 per cent in the 66-90 in. level. Potsherds of this type were found down to 90 in. At Pemrang, unlaminated pottery constitutes 16-45 per cent in the 0-24 in. level, 45-99 per cent in the 24-48 in. level, 93-99 per cent in the 48-72 in. level, 78-96 per cent in the 72-90 in. level. It is difficult to explain the limited distribution of various scarce types of artifacts. Are the occurrences purely fortuitous or are these artifacts merely local- ized manufactures? Table 23 gives examples of both limited and wide distributions. Pemrang site is richest in artifact types, Boldanig a lagging second. These are the deepest and southernmost sites, in sandy terrain. Table 23 lists fifty-three artifact types, distributed in our five sites as follows: Pemrang 45, Boldanig 30, Ruuway 14, Penin 13, Walgom 7. Obviously, Pemrang is the richest site, with 45 types present out of the total of 53. Boldanig is second with 30 types. Both sites are 90 in. deep. The other three sites, each 30 in. deep, have far fewer types. TABLE 2 3 Local Distribution of Artifact Types (x, present; ellipses, absent) Sites Artifacts I X X X Walgom Ruuway Penin Boldanig Pemrang Pottery: Laminated Untempered ...................... x x x x x Lip type a ....................... x x x x Lip type b ....................... x x x x x Lip type c ....................... x x x x x Lip type d ....................... ... x x x x Lip type e ....................... ... x ... ... x Gambreled ... .................... x x Drilled hole ....... .............. x x x ... x Pottery: Unlaminated Tempered (Marianas Plain) ....... x x ... x x Lip type a ....................... ... ... ... x Lip type b....................... ... ... ... x x Lip type c....................... ... ... ... x x Lip type d....................... ... ... ... ... x Lip type e....................... ... ... ... ... x Lip type f....................... ... ... ... ... x Lip type g....................... ... ... ... ... x Lip type h....................... ... ... ... ... x Lip type i....................... ... ... ... ... x Incised .......................... x ... ... ... x Pottery: Painted .................... ... ... ... ... x Shell Clamshell adze Type 1 ....... ................ ... ... x x Type 2a ....... ............... ... x x x Type 2b ................... ... ... ... ... ... x Type 3 ................ ... ... x x 196 197 GIFFORD: EXCAVATIONS IN YAP TABLE 23 (c ontinued) Sites Artifacts I I I Artifacts_______________________ Walgom Ruuway Penin Boldanig Pemrang Shell (continued) Clamshell adze (continued) Type 4 ........ ................ .. ... ... x x Type 5 ........ ................ .. ... ... x Terebra adze, type 5 .... ... ... ...x... Straight Conus knives .............. ... x ... x x Taro peeler Type 1 ........ ................ ... ... ... x Type 2 ...... ... .. ... ... x Type 3 ......... ............... ... ... ... x Paring knives of whole univalves ... ... ... x x Arca scraper ....... .............. ... ... x x Conus bracelets and rings .... ..... ... ... ... ... x Disk ............................. ... ... ... x x Disk beads ........................ x ... x ... Perforated univalve, drilled hole ... ... ... x x Perforated univalve, cut hole ...... ... ... x x Perforated bivalve, drilled hole .... ... ... ... x Perforated bivalve, cut or abraded hole ........................... x x x x Conus cap ........................ ... ... ... ... x Cowry pottery smoother .............. ... x x x Pearl- shell money ..... ........... ... x ... x x Trumpet ... ....................... x x Bone: Perforated disk ..... .......... ... ... ... ... x Stone A dze ............................. x Disk ............................. ... ... ... x Toy money ....................... ... x ... ... Perforated hard, flat pebble ....... ... ... ... ... x Quartz strike-a-light .............. ... x x x x Mineral pigment ... ............... x ... x x W hetstone ......................... x Pestle ............................ ... ... ... x Hammerstone ............ ... ... ... x x Our excavations have yielded the following number of types limited to a single site: Penin 3, Boldanig 4, Pemrang 16. No types are found only in Walgom or Ruuway. The distribution of artifacts in the relatively shal- low deposits of the three northern sites, Walgom, Ruuway, and Penin, seems to have less significance than in the two deep deposits of Boldanig-Wolom and Pemrang. From Walgom and Ruuway we obtained no adequate charcoal samples, but at both Boldanig- Wolom and Pemrang we collected scanty amounts of unlaminated, tempered pottery (see table 12 and ac- companying text). No unlaminated pottery was found at Penin, although the deposit there was the same depth (30 in.) as at Walgom and Ruuway. The char- coal sample from a depth of 24-30 in. at Penin yielded a radiocarbon date of A.D. 1756. At this site we found three unique artifacts: a piece of toy stone money and a whetstone fragment, both excavated in the 0-6 in. block, and a stone adze found on the surface. These objects are presumably not very old. The two deep deposits of Boldanig-Wolom and Pemrang are another story. In both these sites we find certain artifact types found in only one level. For both sites, also, we have radiocarbon dates from charcoal from low levels. It seems possible therefore to make some rough correlation of artifact types and dates. At Boldanig-Wolom, we have two impressive radiocarbon dates: A.D. 1636, from charcoal taken at a depth of 36-42 in., and A. D. 847, from char- coal taken at a depth of 60-66 in. At Boldanig, unlaminated tempered potsherds of Marianas Plain ware were found down to 90 in. Hence artifacts from below 66 in. would appear to be earlier than this date. These dates seem to warrant a three- fold chronological classification of artifacts from Boldanig- Wolom, corresponding to depth: artifacts found at levels from the surface down to 42 in.; those from 42 to 66 in.; those from 66 to 90 in. (See tables 24 and 25.) ANTHROPOLOGICAL RECORDS TABLE 24 Distribution of Artifact Types, Boldanig-Wolom (x, present; ellipses, absent) Artifacts Depth (in.) 0-42 42-66 66-90 Pottery Laminated, lip type a.... x Laminated, lip type b .... x x x Laminated, lip type c.... x x ... Laminated, lip type d .... x x Unlaminated, lip type b ... x x x Unlarninated, lip typec... x x x Shell Clamshell adze, type 1 x x ... Clamshell adze, type 2a x Clamshell adze, type 3 x... ... Clamshell adze, type 4 ... ... ... x Terebra-shell adze, type 5 x ... ... Straight Conus-shell knife x Paring knives of whole univalves .x x Arca-shell scrapers.. x Shell disk ...x... Shell disk beads .x x Perforated univalves, drilled hole .......... x ... Perforated univalves, cut hole .x x Perforated bivalves, cut or abraded hole .x . ... x Perforated bivalves, drilled hole .x x x Cowry pottery smoothers . x ... Pearl-shell money .x x ... ... Trumpets . x x Stone Stone disk ................ ... x Quartz strike-a-light. ... x Mineral pigment ......... x x x Stone pestle ............. x Hammerstone ........... x x TABLE 25 Distribution of Artifact Types in Pemrang (x, present; ellipses, absent) Depth (in.) Axtifacts -- 4 0-24 [ 24-48 48-72 72-90 Pottery: Laminated Lip type a ....... x x ... Lip type b ....... x x Lip type c ....... x x Lip type d ....... x x x ... Lip type e ....... x ... Gambreled ....... x ... Drilled hole ...... x x ... TABLE 2 5 (continued) Depth (in.) Artifacts l 0-24 24-48 48-72 72-90 Pottery: Unlaminated Lip type a ....... Lip type b ....... Lip type c ....... Lip type d ....... Lip type e ....... Lip type f ....... Lip type g ....... Lip type h ....... Lip type i ....... Incised .......... Pottery: Painted .... Shell Clamshell adze Type 1........ Type 2a ...... Type 2b ...... Type 3........ Type 4........ Straight Conus knives ........ Taro peeler Type 1........ Type 2........ Type 3........ Paring knives of whole univalves Arca scrapers Conus bracelets and rings ... Disks ........... Perforated uni- valves, drilled hole .......... Perforated uni- valves, cut hole .......... Perforated bi- valves, cut or abraded hole.. Conus cap Cowry pottery smoothers Pearl-shell money Trumpets........ Bone: Perforated disk Stone Perforated hard flat pebble, drilled........ Quartz strike-a- light.......... Mineral pigment.. Hammerstone .... x x x x *x x x x * * x x x x x x x x x . . x x x x *. . * * x x x x x x x . x x x x * * x x x ... x ... ... x ... x ... x ... ... x x x ... x ... x x ... x ... x x x x ... ... x * . . * . . x x x ... ... . x x ... ... x * . . x * . . x x x x 198 GIFFORD: EXCAVATIONS IN YAP It is evident from table 24 that certain artifacts oc- cur only in certain levels in Boldanig-Wolom and thus may be correlated with the chronological grouping sug- gested above. The following tabulation lists fifteen types and the levels in which they occur; no one of these types is found in any other level in Boldanig- Wolom, though some occur in other sites. From the 0-42 in. level: 1. Pottery, laminated, lip type a. (Found also in Ruuway, Penin, Pemrang) 2. Clamshell adze, type 2a. (Found also in Penin and Pemrang) 3. Clamshell adze, type 3. (Found also in Pem- rang) 4. Terebra-shell adze, type 5. (Found only in Boldanig) 5. Straight Conus-shell knife. (Found also in Ruuway and Pemrang) 6. Arca-shell scraper. (Found also in Pemrang) 7. Shell disk. (Found also in Pemrang) 8. Cowry-shell pottery-smoother. (Found also in Penin and Pemrang) 9. Pearl- shell money. (Found also in Ruuway and Pemrang) 10. Shell trumpet. (Found also in Pemrang) 11. Stone pestle. (Found only in Boldanig) From the 42-66 in. level: 12. Perforated univalve shell, with drilled hole. (Found also in Pemrang) 13. Quartz strike-a-light. (Found also in Ruuway, (Penin, and Pemrang) 14. Stone disk. (Found only in Boldanig) From the 66-90 in. level: 15. Clamshell adze, type 4. (Found also in Pem- rang) Two of the types from the 0-42 in. level (4 and 11 above) are found only in Boldanig; nine are found in other sites as well. Of the three types found in the 42-66 in. level only one, the stone disk (14 above), is found only in Bolda- nig. The other two occur in other sites as well: a perforated univalve shell, with drilled hole (12), found also in Pemrang, and a quartz strike-a-light (13), found in Ruuway, Penin, and Pemrang. One artifact type (15), a clamshell adze, type 4, which occurs in the 66-90 in. level in Boldanig, is found also in the Pemrang site. The distribution of types found in only one of the four chronological divisions or strata of Pemrang is as follows: 0-24 in. 24-48 in. 48-72 in. 13 types 3 types 5 types Types found only in the 0-24 in. level: 1. Pottery, laminated, lip type e. (Found also in Ruuway) 2. Pottery, unlaminated, lip type d. (Unique to Pemrang) 3. Pottery, laminated, gambreled. (Found also in Ruuway) 4. Clamshell adze, type 2a. (Found also in Penin and Boldanig) 5. Shell taro peeler, type 1. (Unique to Pem- rang) 6. Shell taro peeler, type 2. (Unique to Pem- rang) 7. Shell taro peeler, type 3. (Unique to Pem- rang) 8. Paring knives of whole univalves. (Found also in Boldanig) 9. Arca-shell scraper. (Found also in Boldanig) 10. Conus-shell cap. (Unique to Pemrang) 11. Pearl-shell money. (Found also in Ruuway and Boldanig) 12. Shell trumpet. (Found also in Boldanig) 13. Perforated hard flat pebble. (Unique to Pem- rang) Six of these types (2, 5, 6, 7, 10, 13) are found only in Pemrang; seven are found in other sites also. This distribution contrasts with that at Boldanig, where of the ten types found only in the upper stratum (0-42 in.) only one is unique to Boldanig, whereas nine are found in other sites too. Types found only in the 24-48 in. level: 14. Pottery, unlaminated, lip type h. (Found only in Pemrang) 15. Pottery, unlainated, lip type i. (Found only in Pemrang) 16. Pottery, painted. (Found only in Pemrang) None of the three types found in the 24-48 in. stratum at Pemrang is found in any other site exca- vated. Types found only in the 48-72 in. level: 17. Pottery, unlaminated, lip type e. (Found only in Pemrang) 18. Clamshell adze, type 3. (Found also at Bol- danig) 19. Perforated bone disk. (Found only in Pem- rang) 20. Quartz strike-a-light. (Found also in Ruuway, Penin, and Boldanig) 21. Hammerstone. (Found also in Boldanig) Two of the five types found in the 48-72 in. level in Pemrang (types 17 and 19) are not found in any other site. The four types of artifacts found in the 72-90 in. level in Pemrang occur also at higher levels in the site. 199 FOREIGN CULTURAL RELATIONSHIPS Dr. Alexander Spoehr, in his admirable Marianas Prehistory, makes the following comment (p. 174): Until archaeology is further advanced in the western Pacific, it is too early to determine the prehistoric cultural affiliations of the Marianas. Logically, on the prehistoric time level, the Mari- anas should exhibit relations with the Palaus and with Yap. Our excavations in Yap have demonstrated a close connection between the archaeology of that island and the Marianas, evidenced by at least two important types of ancient artifacts, potsherds and shell adzes. Fourteen of our unlaminated, tempered sherds from Boldanig and Pemrang were sent to Dr. Spoehr for comparison with his Marianas sherds. Our pieces, he says, in a letter of September 4, 1957, "are identical with the Marianas sherds I have classed as 'Marianas Plain.' " Comparison of the Tridacna-shell adzes shown in our plate 14 with those illustrated in Spoehr's figure 81 also shows close resemblance between the Yapese and Marianas specimens. In addition, Spoehr figures a fragmentary Terebra gouge, perhaps made of the same shell, Terebra maculata, as our specimen, plates 37, r, and 40, a. Shell adzes from elsewhere in the Caroilnes seem to differ in shape from most of the Yap ones, as Matsumura's illustration (Matsu- mura, 1918, fig. 30) shows. This negative evidence complements the positive resemblances cited above. Matsumura (p. 79) comments that in Ponape and Truk, where basalt is found, the natives prefer shell, "since it is easier to make their implements of shells than of stone." This preference, as has been pointed out, is also evident in Yap. There are other parallels between the archaeologi- cal artifacts found in Yap and in the Marianas. Some of our drilled sherds from Yap resemble the drilled sherds of Marianas Plain ware shown in Spoehr' s figure 46, all ours, however, being of laminated ware. None of Spoehr's stone adzes (his figs. 63-65) matches exactly our one amphibolite adze from Penin (pl. 37, s). Possibly ours is imported. However, his short, stubby stone pestles (fig. 70) suggest our coral limestone pestle from Rugog's grave (pl. 41, m), which, like Spoehr's, was said to be for grinding medicines. See also our stone pestle fragment (pl. 41, o). Ham- merstones from the Marianas (Spoehr, fig. 73) resem- ble some of our specimens from Yap, which are not illustrated. Marianas scrapers of Turbo shell, shown in Spoehr's figure 83, are similar to the artifacts we have called paring knives (pl. 40, jg). The univalve shell with drilled hole near the edge of the lip, illustrated in Spoehr's figure 84 and said to be used as a lime container for betel-nut chewing, suggests our similarly drilled uni- valve specimens (pl. 40, n-q). Marianas shell beads (Spoehr, 1957, fig. 86) resemble the Yapese beads in plate 38, m-v, and a shell bracelet, shown in Spoehr's figure, is comparable to our Yapese examples in plate 38, i-1. There are, however, some archaeological artifacts from Yap which resemble, not the Marianas specimens discussed in Spoehr's monograph, but other Oceanian types. Two taro pounders of coral limestone (pl. 34, a, b) suggest somewhat two more elaborate ones pic- tured by Matsumura (fig. 18, a, b) from Truk and Kusaie, and also resemble the plainer Polynesian types. Trumpets of triton shells (Charonia tritonis) (pl. 36) are similar to trumpets of this kind in other parts of the tropical Pacific. The Terebra gouge mentioned above (pl. 37, r and 17, a) has an interesting Melanesian parallel. It re- sembles a gouge (perhaps of Mitra shell) from Male- kula in the New Hebrides, shown in figure 10, a of the British Museum Handbook to the Ethnographical Collections. It seems obvious that Yap has, archaeologically, its closest affiliations with the Marianas. Excavations in the Palaus might demonstrate equally close or closer relations with that group, but speculation on this point must await further work in the Palaus. [ 200 LITERATURE CITED Abbreviations American Antiquity University of California Publications Anthropological Records American Archaeology and Ethnology British Museum 1910. Handbook to the Ethnographical Collections. London. Broecker, W. S., J. L. Kulp, and C. S. Tucek 1956. Lamont Natural Radiocarbon Measurements, III. Science, 124 (no. 3213): 154-163. Buck, Peter H., Kenneth P. Emory, H. D. Skinner, and John F. G. Stokes 1930. Terminology for Ground Stone Cutting Imple- ments in Polynesia. Jour. Polynesian Soc., 39:174-180. Christian, F. W. 1899. The Caroline Islands: Travel in the Sea of the Little Lands. Methuen & Co., London. Furness, William Henry 1910. The Island of Stone Money. Uap of the Caro- lines. J. B. Lippincott Co., Philadelphia. Gifford, E. W. 1916. Composition of California Shellmounds. UC- PAAE 12:1-29. 1949. Diet and Age of Californian Shellmounds. A Ant 14:223-224. 1951. Archaeological Excavations in Fiji. UC-AR 13:189-288. 1955. Six Fijian Radiocarbon Dates. Jour. Poly- nesian Soc., 64:240. Gifford, E. W., and Dick Shutler, Jr. 1956. Archaeological Excavations in New Caledonia. UC-AR 18:1-148. Kubary, J. S. 1895. Ethnographische Beitrige zur Kenntnis des Karolinen Archipels. Leiden. Matsumura, Akira 1918. Contributions to the Ethnography of Micro- nesia. Jour. College of Science, 40:1-174. Tokyo Imperial University. Miller, W. 1917. Yap. Ergebnisse der Suidsee-Expedition, 1908-1910. Pt. 1, pp. 1-380. 1918. Ibid., pt. 2, pp. 381-811. Text. Spoehr, Alexander 1957. Marianas Prehistory: Archaeological Survey and Excavations on Saipan, Tinian, and Rota. Fieldiana: Anthropology, Vol. 48 (June 24, 1957), Chicago Natural History Museum, Chicago. Walleser, S. 1913. Religi6se Anschauungen und Gebriuche der Be- wohner von Jap. Anthropos, 8: 607-630, 1044-1069. [ 201 ] A Ant UC -AR -PAAE PLATES EXPLANATION OF PLATES PLATE 24 Archaeological features in Teb village area. a. Shrine of the god, Magrogoy. Pitug squats on stone money which has fallen over. b. Stone pillar, 9 ft. 7 in. high, at Cheen; Loochaz (left) and Pitug. c. Grind- stone at Cheen. d. Grave of Rugog. Loochaz stands at northwest corner of grave. PLATE 25 Archaeological site Ruuway 2. a. Looking north; Muguy collecting potsherds on surface. b. Wall B4- C4, viewed from south; dug to depth of 30 in.; clayey soil shows near base. c. Wall B2-C2, viewed from north; dug to depth of 24 in., exposing coral blocks; the block in the foreground overhangs the lowest six inches (24-30 in. level) of wall B3-C3. PLATE 26 Archaeological sites Penin and Boldanig. a. Sacred stone at Penin; pencil as gauge of height. Photograph by Porter E. Ward, U. S. Geological Survey. b. Stone cairn at Penin; a whole Hippopus hippopus shell, con- taining pearl-shell money, was found here. Photograph by Porter E. Ward, U. S. Geological Survey. c. Bolda- nig house mound, viewed from north; one piece of stone money; stones from retaining wall scattered by search- ers for land crabs. The north stone of fire box is vis- ible behind young coconut at left. d. Stone fire box on Boldanig house mound; close-up view from south. The stone at the left edge of the picture forms the west side of the fire box; the northern and western stones are in shadow; south stone shows in middle foreground; wall Bl-2 of excavations was dug beside the west stone (see diag. 4, p. 157). PLATE 27 Archaeological sites Boldanig and Wolom. a. Wall E1-F1, Boldanig, dug to a depth of 78 in.; ground water at bottom; sand from excavation in foreground. The workmen descended and ascended by means of the lean- ing branch at left. b. Crew sitting on edge of pit at Boldanig. Left to right: Francis Nuuan, Matam, Joseph Sogon, Rosmor, Vincent Untaman, Loochaz. c. Wolom shrine, after the vegetation had been cut away. Two leaning tall stones supported the ridgepole; the three low stones mark the two sides of the shrine. PLATE 28 Pottery-making at Gitam, March 17, 1956. a. Finek enlarges the clay cylinder to form the walls of the ves- sel. The work is done by the hands alone; no tool is used. b. Finek expands the sides and smoothes the in- terior with a tool of cowry (Cypraea tigris) shell (31967). She supports the vessel with her left hand. c. Finek smoothes the interior of the bowl and shapes the in- curved rim with the cowry shell. d. Finek shapes and expands the rim, moistening her fingers with water from the coconut cup at her right. The cowry tool lies in the bowl. PLATE 29 Pottery-making at Gitam, March 17, 1956, a and b. Finek smoothes the interior of the bowl with the cowry shell. In b note the solid base of the original cylindri- cal mass of clay resting on board. c. Finek touches up the rim with wet fingers. The coconut cup contains water; the cowry lies in the vessel. PLATE 30 Pottery-making at Gitam, March 24, 1956. a. fRu- waureng, using a split bamboo stick as a drawing knife, scrapes off the excess clay after the bowl has been cut from the clay base. b. Finek (left) moves the inverted vessel to get at another part to shave off the excess clay. Ruwaureng uses a bamboo spokeshave; note the pared-off clay on the areca spathe. c. Finek trims the side of the bowl with a bamboo spokeshave. PLATE 31 Pottery-firing at Gitam, July 25, 1957. Firing the vessels formed on March 17 and 24, 1956. a. Before firing Finek smoothes the surfaces of a dried pot with a piece of coconut husk dipped in water contained in a coconut shell. b. Pot being fired is placed on coals by Finek, who has added coconut spathe as fuel. c. Finek removes the fired pot from the coals, using a stick because the pot is still hot. Photographs by Shi- geru Kaneshiro and Defngin. PLATE 32 Potsherds. a. Paste texture, laminated; Ruuway 2, surface; 32007. b. Paste texture, pebble-tempered, un- laminated; Pemrang, 6-12 in.; 35465. c. Paste texture, calcite-tempered, unlaminated; Pemrang, 24-30 in.; 36500. d. Surface texture, laminated; Pemrang, 0-6 in.; 35423. e. Surface texture, pebble-tempered, unlami- nated; Ruuway 2, 24-30 in.; 32064. f. Gambreled shoul- der, rim sherd, laminated; Pemrang, 0-6 in.; 35804. g. Beveled lip, tempered, unlaminated; Pemrang, 24-30 in.; 36854. h. Beveled lip, tempered; Peleliu I., Palau Group, surface; 11926; for comparison with g. i. In- cised rim, tempered, unlaminated; Walgom, 6-12 in.; 36911. ij. Incised, laminated, tempered; Pemrang, 36- 42 in.; 36766. k. Incised rim, laminated, tempered; Pemrang, 3 -42 in.; 36765. 1. Drilled, laminated; Pem- [ 204 1 GIFFORD: EXCAVATIONS IN YAP rang, 6-12 in.; 35006; cross-section showing conical perforation. m. Drilled, laminated, exterior; Penin, 0-6 in.; 32508. n. Drilled, laminated, exterior; Pem- rang, 0-6 in.; 35431. o. Drilled, laminated, exterior; Pemrang, 0-6 in., 36780. p. Drilled, laminated, in- terior; Pemrang, 6-12 in.; 35007. q. Drilled, lami- nated, interior; Pemrang, 12-18 in.; 35934. Length of o is 112 mm.; all others to scale. PLATE 33 Pottery bowls. a. Purchased at Kanif village from owner, who bought it in Gitam village; has served as a hearth; 32746. b. Boldanig, surface; 34510. Length of b is 343 mm.; a is same scale. PLATE 34 Pestles of coral limestone and shell. a. Coral, ob- tained at Merur village, Tomil; 32742; gift of Magis- trate Roboman. b. Coral, purchased at Kanif village, Dalipebinau; 32433. c. Shell, obtained at Lamer vil- lage, Rull; 32748. d. Shell, obtained at Magrogoy's shrine, Teb; 36961. Length of c is 277 mm.; others to scale. PLATE 35 Shell (Pinctada margaritifera) money. a. Yap; 14642; gift of Professor David M. Schneider. b. Magrogoy's shrine; 36957. c. Obtained at Rumu village, Fanif municipality; 32747. d. Obtained at Gitam village, Rull; 31962. Length of b is 131 mm.; others to scale. PLATE 36 Shell (Charonia tritonis Linne) trumpets. a. Magro- goy's shrine, Teb; 36964; gift of Magistrate Roboman. b. Boldanig, 6-12 in.; 34533. c. Boldanig, surface; 34512. d. Magrogoy's shrine, Teb; 36963. Length of d is 380 mm.; others to scale. PLATE 37 Shell adzes; stone adze. a. Shell adze, type 1, Bol- danig; 30-36 in.; 33448. b. Shell adze, type 1; Boldanig, 48-54 in.; 33165. c. Shell adze, type 1; Pemrang, 30- 36 in.; 36095. d. Shell adze, type 1; Pemrang, surface; 34823. e. Shell adze, type 1; Boldanig, 42-48 in.; 33493. f. Shell adze (gouge), type 2a; Boldanig, 18-24 in.; 33961. .. Shell adze, type 2b; Pemrang, 0-6 in.; 34880. h. Shell adze, type 2b; Pemrang, 12-18 in.; 36431. i. Shell adze, type 2b; Pemrang, 0-6 in.; 35808. i. Shell adze, type 2b; Pemrang, 12-18 in.; 35520. k. Shell adze, type 4; Pemrang, 30-36 in.; 35180. 1. Shell adze, type 4; Pem- rang, 6-12 in.; 35023. m. Shell adze, type 3; Boldanig, 36-42 in.; 34356. n. Shell adze, type 3; Pemrang, 48- 54 in.; 35683. o. Shell adze, type 3; Pemrang, 42-48 in.; 36564. p. Unfinished specimen; Pemrang, 18-24 in.; 36509. q. Unfinished specimen; Pemrang, 12-18 in.; 35944. r. Terebra-shell adze (gouge); Boldanig, 30-36 in.; 33074; see pl. 40, a, showing manner of grinding the shell. s. Stone adze of amphibolite, or hornblende schist; Penin, surface; 32379. Length of f is 81 mm.; others to scale. PLATE 38 Shell, bone, and stone artifacts. a. Conus literatus knife; Pemrang, 6-12 in.; 36697. b. Conus literatus knife; Boldanig, 12-18 in.; 33945. c. Conus marmoreus knife; Boldanig, 18-24 in.; 33971. d. Cowry (Cypraea tigris) pottery smoother; Pemrang, 36-42 in.; 36150. e. Ethnological pottery smoothing tool of Cypraea tigris; Gitam village, Rull; 31967. f. Shell taro peeler, type 1; Pemrang, 0-6 in.; 34955. &. Shell taro peeler, type 1; Pemrang, 6-12 in.; 35477. h. Shell taro peeler, type 1; Pemrang, 0-6 in.; 34879. i. Shell bracelet fragment; Pemrang, 12-18 in.; 36834. j. Shell bracelet fragment; Pemrang, 12-18 in.; 36712. k. Shell bracelet fragment; Pemrang, 48-54 in.; 36201. 1. Shell bracelet fragment; Pemrang, 36-42 in.; 36542. m. Shell bead; Ruuway 2, 6-12 in.; 32112. n. Shell bead; Ruuway 2, 12-18 in.; 32135. o. Shell bead; Boldanig, 42-48 in.; 33141. p. Shell bead; Boldanig, 24-30 in.; 33667a. q. Shell bead; Boldanig, 18-24 in.; 33960b. r. Shell bead; Boldanig, 30-36 in.; 33072a. s. Shell bead; Boldanig, 18-24 in.; 33646. t. Shell bead; Boldanig, 30-36 in.; 34010. u. Shell bead; Boldanig, 30-36 in.; 33072b. v. Shell bead; Boldanig, 18-24 in.; 33959. w. Man's dance bracelet, Conus literatus, ethnological specimen; Merur village, Tomil; 32744; gift of Magistrate Roboman. x. Drilled turtle-bone disk; Pemrang, 54-60 in.; 35291. y. Drilled stone (schist) (toy money?); Pemrang, 0-6 in.; 36675. z. Drilled stone (toy money?); Penin, 0-6 in.; 32380. Length of b is 51 mm.; others to scale. PLATE 39 Stone money, ethnological specimens. a. Yap; 32750; gift of Mr. Kent R. Groote. b. Lamer village, Rull; 32430; gift of Miss Ada Ariadna Tretnoff. c. Merur village, Tomil; 32743; gift of Magistrate Roboman. Greatest diameter of b is 258 mm.; others to scale. PLATE 40 Shell artifacts. a. Adze (gouge) of Terebra maculata showing manner of grinding; 33074; see pl. 37, r. b. Unperforated taro peeler of type 1; Pemrang, 6-12 in.; 35474. c. Perforated taro peeler, type 2; Pemrang, 6-12 in.; 36822. d. Unperforated taro peeler, type 2, Pemrang, 0-6 in.; 36793. e. Taro peeler, type 3; Pem- rang, 6-12 in.; 35478. f. Paring knife of Fasciolaria filamentosa; Pemrang, 0-6 in.; 34878. g. Paring knife of Turbo argyrostomus; Wolom, 60-66 in.; 34781. h. Paring knife of Strombus luhuanus; Boldanig, 24-30 in.; 34315. i. Scraper of Arca scapha; Boldanig, 6-12 in.; 33911. i. Conus literatus shell with encircling groove, apparently for cutting off a piece for a ring; Pemrang, surface; 34834. k. Disk of Tridacna or Eippopus shell; Boldanig, 24-30 in.; 33985. 1. Disk, probably Tridacna; Pemrang, 24-30 in.; 36748. m. Conus shell knife, about 2/3 of shell, showing ground edge; Pemrang, 30-36 in.; 35625. n. Drilled Turbo Ticaonicus shell, Wolom, 48- 54 in.; 34745. o. Drilled Turbo ticaonicus shell; Wolom, 54-60 in.; 34762. . Drilled Nerita undata shell; Pem- rang, 66-72 in.; 35356. q. Drilled Nerita undata shell; 205 ANTHROPOLOGICAL RECORDS Pemrang, 84-90 in.; 35417. Length of a is 77 mm.; others to scale. PLATE 41 Shell and stone artifacts. a. Perforated univalve, Cerithium nodulosum, with cut hole; Boldanig, 48 - 54 in.; 34091. b. Perforated univalve, Cerithium nodulo- sum, with cut hole; Pemrang, 30-36 in.; 36115. c. Perforated bivalve, Arca scapha, with cut hole near umbo; Pemrang, 48-54 in.; 35262. d. Perforated bi- valve, Arca scapha, with cut hole near umbo; Pemrang, 0-6 in.; 34888. e. Perforated bivalve, Cardium min- danense, with hole near umnbo; Penin, 0-6 in.; 32675. f. Perforated bivalve, Cardium mindanense, with hole near umbo; Pemrang, 48-54 in., 35690. g. Perforated bivalve, Gafrarium tumidum, with drilled hole; Bolda- nig, 84-90 in.; 33851. h. Perforated bivalve, Ostrea sp., with drilled hole; Boldanig, 0-6 in.; 33303. i. Per- forated bivalve, Pinctada margaritifera, with drilled hole; probably shell money; Boldanig, 0-6 in.; 33875. L Perforated bivalve, Spondylus radians, with 2 drilled holes; Boldanig, 12-18 in.; 34561. k. Filed-off cap of Conus literatus; Pemrang, 12-18 in.; 35087. 1. Stone (phosphate ?) disk; Boldanig, 48-54 in.; 34086. m. Shell pounder for medicine; Rugog's grave, surface; 36972. n. Whetstone fragment; Penin, 0-6 in.; 32707. o. Stone pestle fragment; Boldanig, 30-36 in.; 33449. Length of a is 75 mm.; others to scale. 206 a b d Plate 24. Archaeological features in Teb village [ 207 1 c a c b Plate 25. Archaeological site Ruuway 2 [ 208 1 # '4~~~~~~~~~0 c d Plate 26. Archaeological sites Penin and Boldanig [ 209 1 a b b a c Plate 27. Archaeological sites Boldanig and Wolom [ 210 ] a Plate 28. Pottery-making at Gitam, March 17 [ 211 ] c d b b c Plate 29. Pottery-making at Gitam, March 17 [ 212 ] a a b c Plate 30. Pottery-making at Gitam, March 24 [ 213 1 41~~~~~~~ PlZate31. Pottery-firing at Gitam, July 25, 1957 Plate 31. Pottery-firing at Gitam, July 25, 1957 [ 214 1 b 0 ~ ~~~~~~~~ 8 4 Plat 3. Poshrd _~~~~~~~~~~~~~ 215 7: i ;................. 00 a b Plate 33. Pottery bowls [ 216 ] -? 't IN~~~' -b I y ' . 'a .,, , i Plate 34. Pestles of coral limestone and shell [ 217 ] I Ix :. 141 - i-4 - -* ?:' , , I - "N I ? ? 0 ?? ??? I" t . I S: I X1 Plate 35. Shell money [ 218 ] i 7A,a A'4) A:0 > M Plate 36. Shell trumpets [ 219 ] ppp,- 0 -q ..:?? A 1, _s . 4, , .j I I I .. : ... } - r 8&' 1-1::? Plate 37. Shell adzes; stone adzes [ 220 ] t N.'i Pr # *#., I I I , i I* :; s I : . . v t vqE 4?./jI? 1 N . .t . . '.I 1. " . " ?41 -% .V??? ? .1 . 1, t -o' q I i?? 1?1 . . Ir 4 -m_ Plate 38. Shell, bone, and stone artifacts [ 221 ] k-.,. ? 11 k ? -I !?- ?,? , ??, .1 i?l * "A " ? , , ? - I w, : ?: " "A, Plate 39. Stone money [ 222 ] 7k *., jj- t Iv Plate 40. Shell artifacts [ 223 ] i iAt, F- < 4I 2 i k I.. '. ft S 4 , -, IO .% *: 2E :i - : I* 3j' Plate 41. Shell and stone artifacts [ 224 ] V4 AVA =. I ,- 4 r* 1