572.05 FA N.S. no. 22-27 1994-96 FIEI pnthropology i.s 10.24 NEW SERIES, NO. 24 Paugvik: A Nineteenth-Century Native Village on Bristol Bay, Alaska Don E. Dumond James W. VanStone Published August 31, 1995 Publication 1467 PUBLISHED BY 1 iJbJLD MUSEUM ut inai MPAIGM ij7c ci page x-essing, which reduces the publication time. Contributions from staff, research ' t regai-dless of ability to pay page charges, however, the t. I ascripts. Three complete copies of the text (including ti. ;act; and of the illustrations shoulo be submitted (one original copy plus two review copies which may be s). No manuscripts will be considered for publication or submitted to reviewers before all Tiaterials arc co;npii it' and in the hands of the Scientific Editor. \i I'M ,.■'•ir)t'^ slioiiUi 1)(^ submiited U) Sricniifif Fditor. Fu'IdLina. Field Museum of Natur.il Ilisi'. Tej-t: .Marmscnpts must be typCHTittcn double-spaced on standarcl-weigtvt, Ky?- t5y ii-mcn paper wan wiac margins on all four sides. If typed on an IBM-compatible computer using MS-DOS, also submit text on SVi-inch diskette (WordPerfect 4.1, 4.2, or 5.0, Muiii' ' ^ Wang PC, Samna, Microsoft Word, Volksv - WordStar programs or ASCII). I'or papers over 100 manuscript pages, authors are requested to submit a Table of Contents," a "List of Illustrations," dJuJ a "List of Tables" immediately following title page. In most cases, the text should be preceded by an "Abstract" and should conclude with "Acknowledgments" (if any) and "Literature Cited." All measurements should be in the metric system (periods are not used after abbreviated i;c.>su;cnicntsV T!ic forina! and style of headings should follow that of recent issues ol Fieldiana For more detailed style information, see The r^>'"'"' az-^k,,,^/ ( licago Press, and also recent issues of FietSiana. References: In "Literature Cited," book and journal titles should be given m full. Where abbreviations are desirable fc ; , in citation of synonymies), authors consistently should follow Botanico-Periodicum-Hunnanum and TL-2 Taxonomic • ■ by F. A. Stafleu & R. S. Cowan (1976 ei seq.) (botanical papers) or Serial Sources for the Biosis Data Base (1983) , . . . cd by the BioSciences Information Service. Names of botanical authors should follow the "Draft Index of Author Abbreviations. Rojal Botanic Gardens, Kew," 1984 edition, or 7 ' References should be typed in the following form: Croat, T. B. 1978. Flora of Barro Colorado Island. Stanford University Press, Stanford, Calif., 943 pp. Grubb, P. J., J. R. Lloyd, and T. D. Pennington. 1%3, A comparison of montane and lowland rain forest in 1 1 I y„- ("orest structure, physiognomy, and floristics. Journal of Ecology, 51: 567-601. I^ngdor ; 979. Yage among the Siona: Cultural patterns in visions, pp. 63-80. In Browman, D. L.,and R. A. .->ciiwi»r,^, eds., Spirits, Shamans, and Stars. Mouton Publishers, The Hague, Netheriands. Murra, J. 1946. The historic tribes of Ecuador, pp. 785-821. In Steward, J. H., ed.. Handbook of South Ameri Indians. Vol. 2, The Andean Civilizations. Bulletin 143, Bureau of American Ethnology, Smithsonian Institution, Washington, D.C. Siob:e, R. G. 1981. Ferns and fern allies of Guatemala. Part II. Polypodiaceae. Fieldiana: Botany, n.s., 6: 1-522. Illustrations: Illustrations are referred to as "figures" in the text (not as "plates"). Figures must be accompanied bv V ime indication of scale, normally a reference bar. Statements in figure captions alone, such as "x 0.8," are not accepta! Qiptions should be typed double-spaced and consecutively. See recent issues oi Fieldiana for details o' ■•"''■ All illustrations should be marked on the reverse with author's name, figure number(s), and "to )mitted should, whenever practicable, be 8Vi x 11 inches (22 x 28 cm), and may not cxceca u •/2 x io:-^ Illustrations should be mounted on boards in the arrangement to be obtained in the printed work. This o suitable for transmission to the printer as follows: Pen and ink drawings may be originals (preferred) (1 drrA-inps nn;st be originals, but within the size limitation; and photostats must be high-quality, glossy, ' - inding author upon pi)'- ihors who wish to publish figu; quire costly special paper or color reproductior lii^rit'; vxiih th'" Scipntinr Fjiitor p correction systci ndingauthc on which dclclion&, additions, and changes can be made and queries answered. Uniy one set or All desired corrections of t]/pe must be made on the single set of page proofs Changes in page lo corrections) are very ex, page proofs can < ■ the ivance to pay for them. This paper meet* the requirements of ANSI/NISO Z39.48*1992 (Permaner)pe of Paper). FIELDIANA Anthropology NEW SERIES, NO. 24 Paugvik: A Nineteenth-Century Native Village on Bristol Bay, Alaska Don E. Dumond Professor Emeritus Department of Anthropology University of Oregon Eugene, Oregon 97403-1218 James W. VanStone Curator Emeritus Department of Anthropology Field Museum of Natural History Chicago, Illinois 60605-2496 Accepted October 28, 1994 Published August 31, 1995 Publication 1467 PUBLISHED BY FIELD MUSEUM OF NATURAL HISTORY U" © 1995 Field Museum of Natural History Library of Congress Catalog Card Number: 95-78799 ISSN 0071-4739 PRINTED IN THE UNITED STATES OF AMERICA M.S. . .>. Table of Contents 1 . History of the Region 1 The Region and Its People 1 Russian Explorations 4 The Aglurmiut 4 Russian Explorers at Paugvik 6 Paugvik and the Fur Trade 7 The Russian Orthodox Church 8 Paugvik in the American Period 9 2. History of Archaeological Research ..15 Work Before 1985 15 Summary of the 1985 Field Season 18 Mapping 18 Excavations 19 3. Excavation Detail 21 Trench 1 21 House 1 22 House lA 23 House 2 25 House 2A 25 House 3 27 House 4 27 House 5 29 House 6, Area 6A 29 Trench 4 30 Stratification of Cultural Features 31 4. Collections 33 Procurement Network 33 Hunting 33 Fishing 43 Trapping 44 Transportation 44 Maintenance Network 45 Tools 45 Household Equipment 48 Personal Adornment 53 Smoking Complex 57 Toys 57 Ceremonial Objects 57 Miscellaneous 58 Protective Network 58 Clothing 58 Imported Building Materials 59 Unidentified Objects 59 Wood 59 Antler, Ivory, Bone 59 Miscellaneous Debris 60 5. Paugvik in Historical Context 91 Subsistence 91 Mammals 91 Birds 92 Fishes 93 Shellfish 93 Vegetal Foods 94 Conclusion 94 The Fur Trade 95 Seasonality 95 Dating 96 Ceramics 96 Beads 96 Metal 97 Glass 97 Summary and Conclusions 97 Paugvik in the Russian Period 98 Relations with Neighbors 99 Acknowledgments 101 Literature Cited 102 Appendix 106 List of Illustrations 1. Map of the Alaska Peninsula 2 2. Map of southwest Alaska showing ethnic groups 5 3. Takhuty natives of the Naknek River re- gion 6 4. Native houses near Naknek, 1900 10 5. Native houses and fish-drying racks near South Naknek, 1900 11 6. Native house and cache, 1900 12 7. Aerial view of the Paugvik site and en- virons 16 8. Contour map of the Paugvik site 17 9. Profile of the major occupation area at Paugvik 18 1 0. Profile of a portion of Trench 1 22 1 1 . Plan and sections of Houses 1 and 1 A . . 24 12. Plan and sections of Houses 2 and 2 A . . 26 13. Plan and sections of House 3 28 14. Plan and profiles of House 6 30 1 5. Toggle harpoon head 40 16. Harpoon parts, float parts, arrowheads, bow, wound plug 62 17. Boat hook, end blades, arrow parts, gun part, fishing items 63 1 8. Kayak deck beam, net weights, sled shoes, showshoe part 64 19. Pelt stretchers 65 20. Sled upright, umiak part, wedges 66 21. Sled runner, wedges, sled stanchion .... 67 22. Maul, axe heads, wedge, whetstone .... 68 23. Knife parts, skin scraper blade blank, whetstones, engraving tool 69 lU 24. Whetstone 70 25. Ulus, scrapers or knives, awls, pick or mattock blade 71 26. Pick or mattock blade, compound vessel, shovel blade 72 27. Snow beaters, rake prong, ice pick or chis- el 73 28. Compound vessel parts 74 29. Spoons, lamp, float and bag nozzles, com- pound vessel part, lug 75 30. Dippers, spoon, ladles, kettle rim frag- ment 76 3 1 . Mat or bag fragment 77 32. Mat or bag fragment 77 33. Mat or bag fragment 77 34. Braided grass cordage 78 35. Braided grass cordage 78 36. Lamps 79 37. Lamps 80 38. Chinaware fragments 81 39. Native pottery profiles 82 40. Mask, figurines, respirator 83 4 1 . Skin and shoe fragments, buttons, knotted sealskin 84 42. Unidentified objects 85 43. Unidentified objects 86 44. Harpoon parts, knife blade, bow frag- ment, arrowhead 87 45. Ulu handle, kettle lug, knife blade, mask part, engraving tool, brass box, spoons, kettle lid, wedge, kayak part 88 46. Projectile point, bifaces, end blades, stone saw, slate blanks, adze blade, ulu blade, skin scraper blade 89 47. Chapel at Naknek, 1900 98 List of Tables 1. Potentially time-sensitive artifacts from Trench 1 23 2. Distribution of artifacts and detritus ... 34 3. Chinaware sherds from 1985 excavations 52 4. Glass beads from 1985 excavations .... 54 5. Types and varieties of beads in the 1985 sample 56 6. Comaline d' Aleppo beads from south- western Alaskan sites 56 7. Rotation analysis of six hearths 61 -^ 8. Faunal remains by numbers of elements and individuals 92 9. Birds and fishes identified from previous work 94 10. Proportions of mammal foods probably available 95 1 1 . Counts of potsherds from sites in south- western Alaska 96 IV Paugvik village is well represented in Russian records from southwestern Alaska, suggesting that it was an important settlement in the 19th century. Excavations in 1985 cleared all or parts of nine houses, where faunal and other evidence indicates participation in the commercial fur trade. Although glass trade beads were present throughout, there were rela- tively few other industrial trade items and a profusion of objects of traditional native manufacture. Collections, site layout, and historical documents suggest the village was oc- cupied from after a.d. 1800 to about 1870. History of the Region Herein we report the results of archaeological excavations at the 1 9th-century native set- tlement of Paugvik on the Alaska Peninsula in southwestern Alaska. The major fieldwork was in 1985, when a crew of six devoted two months to excavations at the settlement. In describing the Paugvik collections, however, we have added to materials of 1985 those recovered in abbreviated tests at the site in 1961 and 1973. Although our aim is primarily to describe these archaeological results, we also attempt in a preliminary way to place the people of Paugvik within their social and economic surroundings. The Region and Its People The Alaska Peninsula juts southwestward from the Alaska mainland and, with its partly submerged extension in the long chain of Aleutian Islands, forms the boundary between the Bering Sea on the north and the Pacific Ocean on the south (Fig. 1 ). Toward its wider, northeastern end the peninsula is 160 km or more in width. Throughout its length its backbone is the Aleutian Range of volcanic mountains, peaks of which rise to elevations above 1 800 m and form a divide that in the northeast is 15-25 km from the abrupt, fjorded coast of Shel- ikof Strait on the Pacific but as much as 145 km from the coast on Bristol Bay of the Bering Sea. Toward that coast the ground slopes as a soggy plain built by outwash of the Pleistocene glaciers that carved the basins of the lakes that now stretch in series along the northwestern foot of the moun- tains, which is the source of meandering streams and the spectacular runs of red or sockeye salmon for which Bristol Bay is famous. The village of Paugvik was located on the right bank of the Naknek River 1 km above its mouth on Bristol Bay and 2 km below the modem village of Naknek. Bristol Bay forms the southeast comer of the Bering Sea, and the flat peninsula coastal plain and shallow seas partake of the arctic climate of the north. The plain is treeless and tundra cov- ered, save for a few protected spots in stream val- leys where pioneer stands of stunted spruce ap- pear. On the bay there is a substantial ice cover for much of normal winters. Summers are punc- tuated by periodic storms that rage inland from the unpredictable Bering Sea. Faunal food resources are plentiful in the region. Although the shallow seas of the upper bay dis- courage the approach of larger whales and other sea mammals— walrus, for instance, are found no closer than 200 km to the west, where they haul out in summer on islands fronting Togiak Bay- harbor seals are abundant, and beluga (white whales) inhabit Bristol Bay the year around, cours- ing up the Naknek River in spring and summer in pursuit of mns of smelt and salmon. Some clam species are available in the upper bay, with mussel colonies on intertidal rocks such as those visible at the mouth of the Naknek at low tide. Seabirds and migratory waterfowl are also plentiful in sea- son. The major Alaska Peninsula caribou herd calves in spring in the lowlands of the Bering Sea plain near Port Heiden. In early fall the herd drifts northeastward to winter, usually between the Uga- shik and Naknek rivers, although in the mid- 19th- century caribou were so numerous that they would move seasonally across the Naknek and even the Kvichak River (Hemming, 1971, pp. 39^4). But the most dependable and major resource is pro- vided by the five species of Pacific salmon, which History of the Region 1 Fig. 1 . Map of the Alaska Peninsula. are present in great numbers in streams during their migrations and also available offshore in Bristol Bay. Runs begin in early June with king or Chinook salmon, continue through July with sock- eye (red) and chum (dog) salmon, and last through August with pink and silver (coho) salmon (UA, 1974, pp. 422-440). Although all of these species occur in the Naknek River, red salmon are es- pecially plentiful, with annual upstream escape- ments even under modem fishery pressure running well over 1 million fish (ADFG, 1991); in aborig- inal times the runs into the river must have been substantially larger. Smelt also run into the Nak- nek in spring or early summer, and freshwater fish are abundant, including salmonids, such as rain- bow trout and char, grayling, pike, and whitefish (UA, 1974, p. 444). The upper peninsula lies within the region of aboriginal Western Eskimo or Yupik speech. Nineteenth-century Paugvik itself was occupied by people referred to in the most modem literature as Aglurmiut, known to the Russians as Agleg- miut. Their nearest ethnic and dialectic neighbors were the Aglurmiut of settlements located at the mouth of the Egegik River on the peninsula coast to the southwest, and near the mouth of the Nush- agak River across Bristol Bay. These Aglurmiut were the southemmost speakers of the language designated Central Yupik (Krauss, 1982). They were reported by the early Russians to have been driven from the lower Kuskokwim River vicinity in a series of bloody battles of the late 1 8th century, known more recently in Kuskokwim native tra- dition as the "bow and arrow wars" (see, for in- stance, Ackerman & Ackerman, 1973; Fienup- Riordan, 1990). Although some early U.S. sources credited the Aglurmiut or Aglegmiut with control of all of the Bering Sea slope of the northem Alaska Peninsula (e.g., Petroff, 1881, 1884; Porter, 1893), the Russians knew them to have been restricted to the Bristol Bay littoral, from which they had displaced people called by the Russians "Sever- novtsy (Northemers) and Ugashentzy" (Wrangell, 1980, p. 64). Although the second of these were people of the Ugashik River located well to the T southwest of the region of immediate interest here (Fig. 2), the former were people of the upper por- tion of the Naknek River drainage and hence of relevance to the condition of Paugvik and its neighborhood. About 1 00 km above Paugvik within the Nak- nek River drainage system, immediately above Iliuk Arm of Naknek Lake and in the geographic center of the peninsula, was the multi village com- munity called by the Americans Savonoski, known to the earlier Russians as the Sevemovsk (i.e., northemer) settlements, with their inhabitants the Severnovskie Aleuty, or Sevemovsk Aleuts. A sense of contrast in the identities of people of these set- tlements is made plain by records of births entered by the Alaska Russian Church (ARC, 1 8 1 6-1936, Nushagak mission) between the 1840s and 1895. At Paugvik, births were recorded as 74% "Agleg- miut," 21% Kusquqvagmiut (i.e., people of the Kuskokwim River region), 2% Kiatagmiut (of the upper drainage of the Nushagak River system or the vicinity of Iliamna Lake), and 3% "Aleut." In the Sevemovsk settlements, 92% were recorded as "Aleut," 5% as "Aglegmiut," and 2% as Kiatag- miut (see also Dumond, 1986, p. 5). There may have been some tradition of hostility between villages at the two extremes of the Naknek Lake and River system, as indicated by the Rus- sian accounts of Aglurmiut history. In 1953 a Sev- emovsk native alleged that in very old days the two peoples had fought each other with bow and arrow. In those same olden days, he said, the peo- ple of the lower Naknek River never went up- stream, and the Sevemovsk people never went Part One downriver but repaired to the Pacific coast rather than to Bristol Bay to hunt sea mammals (Davis, 1954). A similar course for trading was reported for the Sevemovsk people in the 1 880s by the first U.S. census official in the region, Ivan Petroff, who remarked that the people of two villages ... in the vicinity of lake Walker [his designation for Naknek Lake] came down to Katmai [on Shelikof Strait] to do their shopping and to dispose of their furs, undertaking a very fatiguing tramp over mountains and glaciers and across deep and dangerous streams in preference to the canoe journey to the Bristol Bay stations. (Petroff, 1884, p. 25) And he reported a local tradition in which hostil- ities probably involving the two Naknek River peoples are alluded to, when at a feeder stream of Naknek Lake there was a night attack made by the "bloodthirsty" Aleuts long years ago, when every soul in the place was dispatched without mercy, with the exception of one man, who hid himself under a waterfall close by, and thus survived to tell the tale. (Petroff, 1884, p. 24) In 1912 the violent volcanic eruption in the vi- cinity of Mt. Katmai, which deposited 30 cm or more of pumice and ash on upper Naknek Lake, caused the permanent abandonment of the two Sevemovsk settlements then occupied. Despite any residual hostile feeling for the 20th-century de- scendants of the Aglurmiut, most of the survivors relocated on the left bank of the Naknek River 10 km above Naknek village. Not only had the Sev- emovsk people fled there as the emption began, but canneries on the lower Naknek promised oc- casional employment, and the Pacific coastal set- tlements they had been inclined to visit in earlier times were totally destroyed by the emption. Unlike the people of Paugvik and their descen- dants, the social and linguistic affinity of the Sev- emovsk people is not clear in generally available sources. However, the matter is important to some considerations stemming from the work reported here and will be pursued briefly. When the Russian fur hunters followed the path of Vitus Bering to the New World after his unlucky voyage of 1741-1742, they applied their appel- lation "Aleut" to native peoples of what we now know as the Aleutian Islands— peoples who spoke one or more languages that are now called (after the Russian innovation) Aleut (Fig. 2). But as the Russians moved eastward around the northem edge of the Pacific, they applied the same term, Aleut, to people they met on Kodiak Island. These were a people who spoke a language entirely unintelli- gible to natives of the Aleutian Islands. It is now recognized as the southemmost of the Yupik lan- guages and designated Alutiiq or Sugpiaq, and the people are called Koniag. The Russian fur hunters also applied the designator "Aleut" to the Eskimo- speaking peoples they began to meet on the Alaska Peninsula (shown as Peninsula Eskimo in Fig. 2). This practice continued until they had crossed the peninsula to the Bering Sea, where they gave sep- arate ethnic designations to the larger ethnolin- guistic groups, such as Kusquqvagmiut of the Kus- kokwim, Kiatagmiut of the upper reaches of the Nushagak, Wood, and Kvichak rivers, and "Aglegmiut" of the Bristol Bay coast. Is there, then, any affinity implied among those Eskimo-speaking peoples they had designated as Aleuts— a designation applied from Kodiak in the south to people of the Ugashik River and of the Sevemovsk settlements of the Naknek drainage in the north? Certainly the native people of the Pacific coast of the Alaska Peninsula were related to those of Kodiak. As one traveler in the first decade of the 19th century reported of people of the peninsula's Kukak Bay (as near to the Sevemovsk settlements as was Katmai), "the customs, the manners, and in a great degree the clothing and language . . . are the same as those of the people of Kodiak" (Langs- dorff, 1814, II, p. 236). And in census and vital statistics documents of the Russian Orthodox church (ARC, 1733-1938, 18 16-1 936), the people of that coast were as often as not referred to as "Kodiak Aleuts." With regard to people farther north on the peninsula, at least one 1 9th-century traveler reported a dialectal difference between Sevemovsk people and those of Katmai (Spurr, 1900, pp. 92-93), although in recent decades na- tive informants in Naknek village have reported that natives of the Sevemovsk villages spoke a language essentially identical to that of both Ko- diak and Ugashik but differing in significant re- spects from speech current around Naknek in the earlier years of this century (Dumond, fieldnotes of 1974, 1985). In 1961, an account of the 1912 Katmai emption was recorded in the native speech of one of the few surviving members of the original Sevemovsk migrants to the lower Naknek River, a woman who was bom in a Sevemovsk settlement in 1 879 according to church records (ARC, 1816- 1936, Nushagak mission). This account has been recognized to be in Alutiiq, although with some Central Yupik elements (Michael Krauss, personal communication to Dumond, 1979). Thus it seems History of the Region reasonable to conclude that the people of both the Sevemovsk villages and Ugashik (i.e., the Penin- sula Eskimo of Fig. 2), like those of Kodiak (the Koniag), were native speakers of some form of Alutiiq. It also seems reasonable to conclude that the Russian ethnic designator "Aleut," when ap- plied to Eskimo-speaking peoples, was reserved for speakers of that same language. Seen in this way, the designation of the upper Naknek community by the Russian fur hunters and priests as "northerner settlements" makes considerable sense. The Sevemovsk people were the northernmost of the "Aleuts" or Alutiiq speak- ers, their villages located directly north of and ac- cessible by trail from the Russian-controlled hunt- ing station of Katmai on the Pacific coast. As Alu- tiiq-speaking southerners, however, they contrast- ed with the Central Yupik Aglurmiut of Paugvik, who may now be seen to have occupied a beach- head in enemy territory until peace was imposed by the Russians. Russian Explorations As early as the mid- 1 8th century, Russian fur trad- ers began to expand into areas north of the Gulf of Alaska. The tip and the southern shore of the Alaska Peninsula were to some extent within the Russian sphere of influence by 1 76 1 , possibly even earlier. In 1 799 the Bristol Bay-Iliamna Lake area was controlled by the Lebedev-Lastochkin Com- pany, and some areas of Bristol Bay probably were explored during the last two decades of the 1 8th century (Black, 1984, p. 27). Early in the 19th century, as the number of fur- bearing animals declined in traditionally exploited regions, the Russian-American Company focused its attention on the vast area of southwestern Alas- ka north of the Alaska Peninsula. There, they be- lieved, new profits could be achieved through trade with the Eskimo and Indian inhabitants for beaver pelts and other furs. The company dispatched an expedition in April 1818 under the command of Petr Korsakovskiy to explore part of the Alaska Peninsula and the coast from uppermost Bristol Bay to the mouth of the Kuskokwim River. The party crossed the peninsula at what is now known as Becharof Lake and moved down its outlet stream to Bristol Bay. In August, leaving some of his party at the mouth of the Nushagak River, Korsakov- skiy led a detachment eastward to lakes Iliamna and Clark. On Iliamna Lake he met Eremy Ro- dionov, a local trader, who offered to lead a small party north into the interior, a difficult journey during which they may have reached the Kusko- kwim River. In the fall Korsakovskiy and his men returned to Kodiak Island by way of Iliamna Lake (VanStone, ed., 1988). In the summer of 1819 Korsakovskiy led an- other exploring party to Bristol Bay. The party planned to explore the Kuskokwim River, but for a variety of reasons was not successful. The 1819 expedition did, however, establish a trading post, Aleksandrovskiy Redoubt, at the mouth of the Nushagak River at what would become the site of the mission and settlement of Nushagak (Fig. 1). Fedor Kolmakov, a company employee who had accompanied Korsakovskiy on both his expedi- tions, was placed in charge (Fedorova, 1973a, p. 8; 1973b, pp. 68-69). The two expeditions of Kor- sakovskiy and the coastal explorations undertaken by V. S. Khromchenko and A. K. Etolin between 1819 and 1822 (VanStone, ed., 1973) provided the Russian-American Company with its first re- liable information concerning relations among na- tive groups in the Bristol Bay region and the extent to which they would be inffuenced by the estab- lishment of Aleksandrovskiy Redoubt (Berkh, 1823a, pt. 2, pp. 1-20, 1823b; RAC/CS, vol. 3, no. 164, 4 May 1823). The Aglurmiut The native people whom Korsakovskiy and other explorers encountered on the upper Alaska Pen- insula and in Bristol Bay were the Central Yupik- speaking Aglurmiut or "Aglegmiut." Korsakov- skiy brieffy described the Aglurmiut in his 1818 journal (VanStone, ed., 1988, pp. 29-31), and the first published account of subcultural groups in the Bristol Bay region was derived from the explorer's 1819 journal as reported by Berkh ( 1 823b). In this account the coastal inhabitants are referred to as "Glakmiut" and are said to have been constantly at war with the Kusquqvagmiut of the Kuskokwim River. V. S. Khromchenko, during his coastal ex- plorations of southwestern Alaska, in 1822, noted that the Aglurmiut were the most warlike people along the coast between Bristol Bay and Norton Sound. His account included a brief description of their culture and a rather extensive vocabulary (VanStone, ed., 1973, pp. 52-53). Wrangell (1970, p. 1 7), Khlebnikov (Lyapunova & Fedorova, eds., 1 979, p. 77), and early reports of the general man- Part One . ^ ^ y< ;^ ^- cv^Tl ;0 100 HHH t^ ^ i!<»W. 90 ortion of the river bluff seen in middle background above the pond to the left. Photo by James Thompson, 1985. in the main part of the Paugvik site, in an area that examination of the eroding river bluff indi- cated contained thick deposits of midden but in which there were no recognizable house depres- sions. Two detached 1 x 2-m pits extended the excavation line an additional 8 m to the east (Du- mond, 1981, Fig. 4.2; present Fig. 9). In the trench and its extensions, characteristic postcontact Pa- vik phase artifacts were found in heavily streaked brown midden soil in which layers of peaty sod marked former ground surfaces. Beneath the heaviest of these old sod layers in the continuous trench, in places undisturbed by intrusions from overlying Pavik levels, there was a layer of white volcanic ash about 1 cm thick. This was thought to be an ash that had been fairly securely dated to about A.D. 1450 in the upper Naknek River drain- age, where it immediately overlies cultural depos- its of the Brooks River Camp phase. In the lower, southwestern end of the trench the excavation of 1973 was carried less than 1 m, at which point sterile glacial outwash was encoun- tered. There the upper surface of the glacial clays was capped by a stratum of peat some 50 cm thick that tapered upslope to the northeastward and vanished entirely about 12 m along the trench (Dumond, 1981, Fig. 4.3). This peat was inter- preted as the remains of vegetation at the edge of the kettle lake that once covered what is now the dissected basin west of the hill, a counterpart of the ponds that still remain east of it (Dumond, 1981, Fig. 4.2; present Fig. 9). In the higher, north- eastern end and in the detached pits, sterile out- wash was not reached by the end of the field sea- son, with excavations carried to about 1.5 m below the surface. The nature of the artifact-rich deposits in the northeastern section of the continuous trench strongly suggested the presence of a habitation noit visible from the surface. A deeper test in a limited area of the section revealed a trace of the white volcanic ash only a few centimeters lower than the trench floor. Again, scattered potsherds of Brooks River Camp phase type— now 1 5 in number (in a total of 356)— were recovered, half of them from otherwise Pavik phase deposits and half from the underlying peaty layer at the western end of the trench. No datable charcoal was recovered with the sherds, most of 16 Part Two History of Archaeological Research 1 7 indistinct depression outline of excavation or clearly defined depression projected outline House 4 Houses 20 meters Fig. 9. Plot of the major occupation area at Paugvik. which could represent items lost or discarded near the edge of the former pond at a time well before the establishment of the known historic Paugvik settlement. Surveillance of the eroding river bluff at the Paugvik site was maintained during the en- suing 1974 field season, and a thin band of white volcanic ash was discovered above scattered ar- tifacts in the bluff immediately opposite the 1973 trench. A small excavation carried below the vol- canic ash at that point yielded 14 Brooks River Camp phase potsherds next to charcoal that pro- duced a radiocarbon date of about a.d 1255 (695 ± 65, SI-2070), finally providing some confir- mation of prehistoric occupation at the site and contributing to the identification of the white vol- canic ash (Dumond, 1981, pp. 65-67). Neverthe- less, it seemed clear that very few remains of any such earlier occupation remained, presumably having been erased by the heavy tidal erosion that as late as the 1 970s was taking its toll on the bluff in the area of heaviest 19th-century occupation. In none of this work, however, had any attempt been made systematically to sample the habitation area of the site to determine its extent, nor had any attempt been made to examine the surround- ing area carefully to determine the probable lo- cation of Hrdlicka's excavations. The need for some such attempt became evident in 1 983, when a road was constructed immediately north of the site to connect the modem village of Naknek with a new municipal sewage facility. Thus, in 1985, with fi- nancial support from the Alaska Historic Preser- vation Office, it was possible to expand the ex- cavation program sufficiently to permit such ad- ditional mapping and testing to be accomplished. Summary of the 1985 Field Season Mapping Although major effort in 1985 was directed to the excavation of houses and middens, a new contour map of the Paugvik site also was plotted by alidade and planetable, supplemented by the use of three datum points at the same elevation— assigned an arbitrary value of 100 m— from which triangula- tions were made and elevations were measured (Fig. 8). For outlying areas some elevations were estimated by handlevel. In the actual area of the 10 well-defined house depressions, horizontal po- sitions were pinpointed by extension of a single metric grid system over the entire site, laid out 18 Part Two with reference to the hne that had been established for the trench of 1973 (Fig. 9). To determine the area of actual major occupa- tion, a soil sampling device with a 1-inch barrel was used where soil conditions permitted. Because the underlying sterile layer in the entire area is a greenish, clay-laden till that is unmistakable even in modest amounts, this small sampler was en- tirely adequate where frost did not impede pene- tration. Shovel tests were used where a larger ex- posure was necessary. In the habitation area in particular, the peaty soil was consistently frozen beneath the thick layer of overlying sod (of Cal- amagrostis canadensis [Canadian bluejoint]). Here it was necessary to first cut away the sod to give the underlying soil a chance to thaw enough that either the shovel or the soil sampler could be used. The southern edge of the site is on the bluff of the Naknek River, which was actively eroding un- til only recently, and on the face of that bluff the signs of occupation could be traced in profile until they disappeared both to the east and to the west. North of the bluff and outside of the area marked by surface depressions (in which the presence of occupation was obvious), tests were made at ap- proximately 10-m intervals as far north as the westernmost of the two small ponds shown in Fig- ure 8. Tests were made at about double this in- terval in the area through the swale just west of the obvious habitation area and also on the rela- tively high ground east of the easternmost pond. Both of these last areas were essentially devoid of any trace of occupation debris, with glacial ma- terial appearing within about 40 cm of the surface. In the higher ground surrounding the habitation area, a search was made for traces of earlier ex- cavations. A total of 1 fairly clearly defined holes, obviously dug a number of years ago, each about 30 X 1 20 cm in extent and 30 cm in depth, were located in the area marked with the bold dashed rectangle in Figure 8. In one of these a fragment of tooth enamel gave evidence of the almost cer- tain presence at one time of a buried human, sug- gesting strongly that this was the place of Hrd- licka's burial excavations. In the same area, seven other depressions, smaller and less clearly delin- eated, were also counted, but higher up the rising ground to the west no additional depressions were noted, and the soil testing revealed no burials that could be identified. Within the habitation area, the limit of signifi- cant occupation debris is bounded by the 95-m contour that appears as a thick line in Figure 8. Within that area the location of debris was some- what irregular, with occupation fill anywhere from 20 cm to 2 m in depth above the irregular surface of the underlying glacial till. As a general rule, the midden deposits appeared to be deepest between the 97- and 99-m contours, although there were exceptions: for instance, the knoll that lies between the concentration of visible houses and the eastern pond was almost devoid of midden on its summit. The area of heaviest occupation trash was not sole- ly confined to the immediate vicinity of visible house depressions. Excavations Although no formal grid had been established in 1973, when the trench of that year was backfilled the key stakes were driven completely into the ground to permit relocation of the trench if de- sired. In 1985 these were used to orient the ex- cavation grid, which was labeled according to car- dinal directions, although in fact the nominal grid north was oriented 4 1 degrees west of magnetic north, or about 20 degrees west of true. Unless specifically indicated otherwise, the nominal ori- entation will be used in the site descriptions here- inafter. Three houses were excavated completely, two more were tested, and likely midden areas were trenched. Traces were found of at least five and possibly six additional houses that were not rec- ognized on the surface. Because frost hampered excavations throughout the season, the major ex- cavation units were attacked for short periods at a time, rotating from one to another as thawing permitted. Trench 1 of 1985 was 20 m in length, its (nom- inal) south edge lying along the coordinate desig- nated NIO, which was just 2 m north of the south edge of the 1973 trench, and its west end (coor- dinate E20) coinciding with the east end of the continuous portion of that 1973 trench (Fig. 9). Frozen ground was encountered a short way under the sod. House 1 was located at the highest point of the knob on which the remains of the village were discerned. Definition of the relatively shallow house was clear except for the south comer, where an earlier disturbance was encountered, the cause of which was only later understood. Concurrent with this excavation, sod was removed from House 2 and House 3, which were selected because they were apparently completely undisturbed by loot- History of Archaeological Research 19 ers' pits and because initial shovel tests showed each of them to have well-defined floors in the vicinity of substantial rock-lined central fireplaces. The amount of frost remaining in all but the cen- tral areas of their floors, however, dictated some delay before excavations could begin in earnest. As frost permitted, work in Trench 1 was con- tinued until Pleistocene-age glacial till was en- countered throughout, at depths varying from 1 .0 to more than 9 m. At its western end, the trench penetrated the floor of a house, including a rock- lined fireplace, that apparently had been perma- nently frozen in its position beneath 1 m or more of peaty overburden. In the course of the summer a small area to the south of Trench 1 was opened up to expose the hearth, and a larger expansion of 2 X 6 m was opened to the north, which was field- designated Trench 2. At its opposite or eastern end. Trench 1 penetrated a heavy deposit of wood ash that suggested the presence of a house floor underlying a portion of House 1 and might have been responsible for the difficulty encountered in defining one comer of that house. A 2 x 2-m area was opened up in this vicinity to test this suppo- sition; this small unit was then designated Trench 3. Meanwhile, thawing had proceeded rapidly enough in Houses 2 and 3 to permit a shift of the crew to those units, while the newly opened trench- es were allowed to thaw. The initial promise of rapid clearance of the new houses, with their par- ticularly well-marked floors, was not realized, however. House 2, in particular, had been excavated over an earlier habitation (designated House 2A), which had its subterranean entrance lying squarely be- neath the stone-lined fireplace and its floor be- neath the rear bench of House 2. House 2 itself was both deeper and larger than expected, and the complexity introduced by the underlying structure caused considerable delay. The excavation of House 3 also began decep- tively, with the structure promising to be rather small and shallow, although with some weU-pre- served structural members and a fairly clear floor deposit. Complexity arose particularly with the discovery that the edges of Houses 2 and 3 had either coincided or overlapped slightly. Further- more, in the vicinity of their conjunction there was also the buried entranceway to an earlier house, the floor of which was never discovered. In ad- dition, the northernmost comer of House 3 had been cut away by a still later house or other stmc- ture, which was so faintly indicated on the surface as to have been imperceptible during our earlier examinations. Because of these complications, progress was sufficiently delayed that full-scale excavations of additional houses could not be undertaken. Trench 2 was cleared to underlying glacial till, exposing about one third of the frozen floor of the buried house, which is now designated House 6. The yield of artifacts of organic material— wood, bone, fur, baleen, etc.— was especially good from this house floor and the adjacent area. Trench 3 was shallow, but the excavation cleared the stone-lined fireplace that had clearly pertained to the earlier house (now House lA) that underlay portions of House 1. Trench 4 was de-sodded to provide some sample of the midden near the entries of Houses 2 and 3. By this time the season was nearly over, and additional sampling of the northeastem portion of the site, in which we had intended to completely excavate at least one relatively undisturbed house, was perforce confined to the testing of two house depressions (Houses 4 and 5). Both had only weak- ly defined floors, although both also revealed the large rock-lined fireplaces that were now recog- nized as characteristic of the site. In Trench 4 an area of 1 x 6 m, less than that originally opened, was carried to a depth of about 1 m, at which time sterile till had not been encountered, and the sea- son was brought to an official close. AU excava- tions were completely backfilled. 20 Part Two Excavation Detail Trench 1 In 1973 the main trench, which was 1 m wide, had been laid out parallel to the bluff with its nominal western end beginning at the base of the western side of the hill on which the main house remains of Paugvik could be discerned. The west- ernmost 1 8-m section was opened as a continuous trench, the 2-m segments of which were simply designated sections 1 through 9, beginning at the low western end. East of this, the trench was con- tinued as two detached 2-m sections, which were designated sections 1 1 and 1 3 (Fig. 9), but neither was carried to the sterile layer because of ham- pering frost. When the trench was backfilled at the close of the season, some key stakes were pounded into the ground to make relocation of the trench possible if desired. In 1985 Trench 1 was laid out with its nominal southern edge (on the coordinate designated NIO in the arbitrary grid of that year) exactly 2 m north of the southern edge of the 1 973 trench, its western end (at coordinate E20) coinciding with the eastern end of section 9 of the 1 973 trench (Fig. 9). Twenty meters to the east (at E40) the trench ended almost exactly 1 m short of the edge of the depression at the top of the hill that was designated House 1 . Through intermittent excavation as permitted by thawing, excavation of the trench was carried to sterile glacial silt and till throughout its length; a total of 27 m^ of material was removed. As with all of the units excavated in 1985, the trench was everywhere overlain by the pinkish streak of volcanic ash marking the 1912 eruption from the vicinity of Mount Katmai. The eastern end of the trench was shallow, with glacial till, clays, and loess encountered within 50 cm of the surface in most of the area east of E36, although in the southern wall of the easternmost meter of the trench there was clear evidence of charcoal and wood ash less than 20 cm below sod that marked the northern edge of a hearth that pertained to the house designated House lA. At the opposite or nominal western end of the trench, where midden overlay the glacial material 1 m or more in depth, a stone-lined hearth was encountered that was considered probably a fea- ture of the house that had been suspected to lie beneath the eastern end of the 1973 trench, and evidence of a vertical aboriginal cut 4 m to the east of that hearth in Trench 1 was considered the edge of the same house. This area of Trench 1 was solidly frozen but slowly yielded plentiful scraps of wood and twigs at the presumed house-floor level. This excavation led to the opening of a 2 x 6-m section north of the east end of Trench 1 , which was designated Trench 2 in the field. This new area was cleared to reveal additional portions of what is here designated House 6. In the 4.5-m section east of the eastern edge of House 6, two more aboriginal cuts were found (Fig. 10). The easternmost of these almost cer- tainly represented a cut for the tunnel of a house entrance, for at that point frozen remains of struc- tural members were found slumped into a trench of aboriginal date that crossed Trench 1 at about right angles; to the west and stratigraphically later was evidence of a second deep cut suspected of having been a part of yet another house that had in turn been partly obliterated when House 6 was constructed (Fig. 10; some of the area designated House 6A almost certainly pertained to this house, although the jumbled logs and sticks of House 6 A did not reveal any clearly decipherable pattern). Thus there appeared to be three generations of Excavation Detail 21 - 97 00 ,t?^d Silt and till Fig. 10. Profile of a portion of the south face of Trench 1 (grid in meters). Structures, all of Pavik phase age, represented in the western portion of Trench 1 . We had hoped that the lengthy expanse of the trench would lead to discovery of an undisturbed deposit of the gray volcanic ash that in the 1970s had been determined to separate materials of the historic Pavik phase from those of the earlier Brooks River Camp phase and thus to mark the location of some significant quantity of the earlier material. However, in most of the trench the ash was nonexistent, evidently (as with all the houses excavated) eradicated entirely by Pavik phase oc- cupants of the site. But in the 4 m of Trench 1 east of E30, and coinciding with the easternmost portion of the deepest part of the trench, discon- tinuous traces of the ash were noted at the trench's southern edge. The overall deposit in that area appeared simply as midden, rather than as a house cut or other aboriginal excavation, and because it was relatively undisturbed beneath the remnant ash, it was thought to promise the recovery of some Camp phase materials (Fig. 10). Unfortu- nately, the yield from that section was the lightest from anywhere in the trench. Potentially diagnostic artifacts from the trench section E30 to E34 are listed in Table 1. Unfor- tunately, the edge of the cut for the apparent house entrance (Fig. 10) was not clearly identified in the generally mixed fill until excavation of that 2-m section of the trench (E30-E32) was nearly com- plete. With few exceptions artifacts were recorded only by level and 2-m section; it is therefore not surprising if the sample from the section is mixed. In the 2-m section to the east (E32-E34), however, there was no such disturbance. The distribution of artifacts there might at first be construed to suggest the presence of at least a trace of a lower component in that area. That is, the lower portions of the section yielded no glass beads and a single thick sherd of a type characteristic of the earlier Brooks River Camp phase (Table 1). However, the presence of four sawed slate pieces (generally rare in materials of the Camp phase, a time when slate was almost universally chipped to shape) seems to indicate with equal strength that the en- tire deposit was predominantly Pavik in age. In any event, given the paltry scale of this uncertain evidence and the small overall proportion of thick- paste sherds from the entire excavation (46 of 930)— where many of the thick fragments may well be from Pavik phase lamps rather than Brooks River Camp phase pots— there seems no reason to suppose that the Pavik collection overall is se- riously contaminated by earlier materials. House 1 The surface depression marking the location of this house was visible enough that the collapsed structure was confidently desodded in its entirety, only to encounter frost a short distance below sod. As thawing permitted work to resume, the abo- riginal floor, with a fireplace in a shallow pit un- lined by stones, was revealed 1 cm below modem ground surface at the center of the depression. Altogether, an area of about 35 m^ was opened, from which some 1 2 m^ of fill was removed by season's end. The house had evidently been con- structed with horizontal logs outlining the base of the walls, and on the northeastern side and along the eastern side of the front and the eastern portion of the rear the limits were defined with some ease 22 Part Three Table 1 . Potentially time-sensitive artifacts from two sections of Trench 1 , Paugvik. Artifact^ No. Label E32-E34 E30-E32 Upper Lower Upper Lower Total Trade objects 21. bullet mold half 1 62. metal ulu 1 107. glass bead 12 133. nail 1 Stone artifacts 15. slate end blade, type 1 1 55. whetstone, type 1 56. whetstone, type 2 ~ 60. stone saw 63. stone ulu blade 1 64. untyijed ulu fragment 1 77. misc. sawed slate pieces 2 78. misc. jKjlished stone 2 140. slate chips and chunks Naknek ware potsherds 100. thin plain, variety unknown 101. thin plain, Pavik variety 105. thick plain, variety unknown 1 2 21 1 2 1 1 1 1 1 6 5 6 7 1 1 " Number and label refer to descriptions in Table 2. by the telltale channel and adjacent cut in the greenish glacial till. In much of this area where the edges were so clearly defined it was also possible to estimate the height of aboriginal ground surface (Fig. 11). In sections of the western side and to- ward the northwestern rear comer of the house, however, the till surface dipped below the floor of the house and the definition of floor limits was less certain. A similar cut in the till was located in parts of the right (southwestern) front comer, but in other parts of the same comer the till had been dug out by an earlier disturbance. As the tunnel entrance of the house was cleared, it became clear that the disturbed area extended to the tunnel side. At first suspected to mark a side chamber, excavation showed that the disturbance had pre- dated the constmction of House 1 , for the buried rock pile that was a major part of the disturbed area, which penetrated about as deeply as the floor of the entrance, had been cut cleanly by the House 1 tunnel edge. The disturbance was explored only to the limits indicated in Figure 1 1 . House 1 was the smallest and shallowest of the four houses in which substantial excavations were made, and it departed from all of the others in the discemible outline of its base logs and in the ab- sence of a barricade of stones around the centrally located fire. The tunnel, too, was unusually stubby and extended an uncharacteristically short dis- tance into the house floor. Although deteriorated posts and horizontal wood fragments were plen- tiful, and although some of these and some shallow depressions without traces of wood (Fig. 1 1) could be imagined to be at strategic locations on the floor, it was not possible to identify the locations of major roof support posts with confidence. The total area of the actual house floor was about 25 m\ House lA Excavations in the eastern end of Trench 1 had revealed a heavy deposit of ash on the south wall, and toward the end of the season the trench was expanded south by a 2 x 2-m cut designated in the field as Trench 3. In a rather hasty excavation, the expansion uncovered a large stone-lined hearth that was almost certainly— to judge from the other houses excavated— the central feature of a dwell- ing, designated House 1 A (Fig. 1 1). The excava- tion was carried somewhat below the level of the supposed house floor, and its relation to the floor of House 1 is shown in Figure 1 1 , section C-C. When the House 1 A fireplace was first exposed to confirm the existence of the house itself, it was thought likely that the House 1 A entry tunnel was the cause of the disturbed fill at the front of House 1. Although this possibility cannot be ruled out Excavation Detail 23 modem sod aboriginal surface rock wood post cliarcoai, wood ash projected edge pit or depression 2 meters SECTION A - A SECTION B - B 1/ SECTION C - C Fig. 1 1. Plan and sections of House 1 and excavated portion of House lA. 24 Part Three entirely, a careful examination of the modem sur- face south of the House 1 A excavation revealed a slight but abrupt depression, and the adjacent bluff showed what seemed to be a section of truncated entrance tunnel, suggesting rather that the en- trance had opened more directly toward what is now the river bluff (as projected in Fig. 9). This orientation is more likely, as indicated by a surface of uncut glacial till that rose some 1 cm above the floor of House 1 at its western edge (Fig. 1 1 , sect. B-B, C-C). Because the till remnant rose even higher above the level of the floor of House 1 A and yet was within 75 cm of the House 1 A hearth, its presence seems explicable only by the existence of a low earthen bench in House 1 A immediately behind (i.e., north or northeast of) its central fire, a bench high enough to lie completely above the till at that point. Whatever its precise orientation. House 1 A must have predated House 1 . First, the position of House lA was almost impossible to discern by any sur- face indication, whereas House 1 was clearly vis- ible. Second, nothing in the fill of House 1 indi- cated any overlying disturbance such as would be caused by a later structure. Finally, if the distur- bance at the front of House 1 was in any way related to House 1 A, the excavation evidence was clear that the House 1 entrance postdated it. In any event, there were clearly two generations of houses located at what we called House 1 . House 2 The central, rock-lined hearth of House 2 was en- countered in the initial test pit at a depth of about 30 cm. As thawing permitted the hearth area to be exposed, the surrounding floor was found to be fairly distinct but not cleanly overlying sterile till as hoped. A transverse trench 40 cm wide was then laid out along the north-south excavation grid and excavated slowly into and through the floor. What it revealed was that the hearth and the center of the house covered earlier excavations of considerable size. Comparable complexity was found over much of the floor, little of which over- lay the recognizable greenish glacial clay; indeed, only a portion of the south wall and the southwest comer were clearly defined where the position of the major comer post was circled by a shallow cut in a small bed of remnant till. The southeastern comer had been entirely eradicated by the deep house lying immediately to the south of House 2. That house had been passed over in choosing the excavation sample both because of its depth and because of all the 1 house depressions visible on the site it had suffered the most serious damage through pot hunting. The positions of the other walls of House 2 were defined only with some difficulty; there was no evidence remaining of any horizontal base logs, but the two northern comers were marked by fairly substantial post remnants. Efforts to positively lo- cate the front wall included a fairly expansive but shallow cut— the limits of which are shown in Fig- ure 12— that produced a few timbers and rocks that seemed at first to mark the existence of some kind of storm shed at the outer end of the sunken entrance, but when the cut was completed there was no clear evidence of any such stmcture. Like the other two houses with completely excavated entries, the entrance tunnel of this one seemed to open directly to the outside. Whereas the transverse trench, which crossed the house diagonally, provided evidence that the southeastern comer had been eradicated by a later structure, it also yielded evidence of some under- lying structural remnants at the north, above which the floor of House 2 was traced. The biggest sur- prise came upon clearing the deep disturbance be- neath the hearth, which the trench had also re- vealed. This disturbance was the entrance of an earlier house, the front part of the floor of which was hopelessly confused with what had been taken to be a slightly raised bench at the back of House 2. As the final result, it was not entirely clear whether the piles of firecracked rock and the area of charcoal and ash found on what was first thought to be the back bench (Fig. 1 2) were, in fact, features of House 2 or of the house underlying it. House 2A With clearance of the earlier entrance to a point slightly west of the back wall of House 2, charcoal and ash appeared that invited an extension west- ward by an arbitrary cut to reveal the rock-filled hearth of what is now designated House 2A, clearly the major structure underlying House 2. No at- tempt was made to carry the clearance beyond that area shown in Figure 1 2. On the south, the original cut for the House 2A floor had not eradicated the till now remaining at the comer of House 2. Al- though the actual northern and southern edges of the House 2A floor were not identified, its width Excavation Detail 25 secnoN D-D FiG. 12. Plan and sectioiis of House 2 and excavated portions of House 2A. appears to have been some^iiliat less than that of the later House 2. The separation of arti£u:ts per- taining to Houses 2 and 2A is provisional only. Altogether, some 55 m* was opened to uncover the total House 2 complex, in which the lai^ge House 2 itself was well over 30 m^ in floor area. In this complex, three generations of houses are indicat- ed: House 2A, overlain by House 2, which then had one firont comer obliterated by construction of the (unnumbered) house immediately south of it In addition, the structure thought to lie beneath the floor within a cut in the till on the north side (Fig. 12), which for lack of time was not cleared when it was concluded to be unassociated with the floor of House 2, could have been contemporary with House 2A. When it became clear that the base of the joint feature of Houses 2 and 2A was reached and glacial till appeared in patchwork &shion around the floor, work was ended without attempting to clear all underlying pockets of mid- 26 Part Three den and other disturbed material below the house and without carrying the excavation everywhere to sterile ground. House 3 Although more deeply buried than House 1 , House 3 was not so deeply covered as House 2 and upon testing promised a substantial floor under only some 20 cm of fill at the stone-lined fireplace. Unfortunately, as in House 2 there were no rec- ognizable base logs. About 40 m^ was finally opened, from which at least 1 5 m' of material was removed, although this amount was insufficient to expose the entire floor. Complexity arose with the discovery that the edge of House 2 had either coincided with or very slightly overlapped that of House 3 (of which the limit shown in Fig. 1 3 is the best approximation). From the relatively greater clarity of the side of House 2, which we actually excavated concur- rently with House 3, it was evident that House 2 was built later. Furthermore, near the conjunction of houses, the underlying log structure that had been found within the northern edge of House 2 (Fig. 1 2) ex- tended beneath the floor of House 3. There, at the northern end of the set of short parallel logs (Fig. 1 3), the pile of fire-cracked rocks continued down below those logs; both logs and rocks were located within a cut in the till. This structural arrangement was thought immediately to be remains of a sunk- en house entrance, but because the floors of both Houses 2 and 3 completely overlay it and because of the persistence of frost in the hole and the rapid passage of the excavation season, this structural arrangement was never explored to its base. What- ever it was, it probably was not a passageway con- necting Houses 2 and 3. In addition, the northernmost comer of House 3 had been eradicated by a still later house or other structure lying north of it but so faintly indicated on the modem surface as to have been missed completely upon earlier examination. This area is now labeled simply "disturbance" (Fig. 1 3). The northern edge of House 3 was perforce left unex- cavated as the season drew to a close, and the floor was found to extend farther in that direction than anticipated and into frozen ground under a very large pile of backdirt. Although, with the exceptions noted, the floor was clearly defined, there was nothing to con- vincingly mark the location of aboriginal ground surface, although it is presumed to have lain no more than 1 cm below the top of modem sod and only a few centimeters below the telltale streak of Katmai volcanic ash. Apparently the house was in most places excavated between 40 and 70 cm into the contemporary surface of the ground. As is often the case, the entranceway, thoroughly fro- zen before it began to be uncovered, yielded a substantial portion of the organic artifacts recov- ered from the house. Added to evidence from House 2, the excava- tion of House 3 suggested that not three but four generations of stmctures could be traced at Paug- vik: ( 1 ) the entry or other stmcture underlying both Houses 2 and 3, (2) House 3, (3) House 2, and (4) the unnumbered house south of House 2. House 2A was a feature of either the first or second of these stages, whereas the northern comer "distur- bance" of House 3 was a feature of either the third or fourth stage. House 4 As the end of the excavation season approached it was clear that time remaining was not sufficient to allow complete clearance of any of the houses that showed clearly on the surface of the remaining eastem end of the site. Two were selected for lim- ited tests. The surface depression designated House 4 was about 5 X 5 m, with an evident entrance channel pointing toward the bluff" above the river (Fig. 9). A 2-m-square cut was made in the center of the visible depression, but although the floor was ev- ident within 20 cm of surface, no hearth appeared; rather, as the floor stain was traced through ex- pansion of the pit southeastward, a substantial hearth appeared in what would have been the ex- treme southeastem comer of the house if the sur- face depression were taken as an accurate indi- cation of the actual house location (which it was concluded not to be). The apparently elliptical rock- ringed hearth, only one side of which was exposed, was 90 cm north-south, thus apparently virtually identical in size and shape to those of all the other houses except House 1. In all, 9.5 m^ was cleared to the relatively shallow floor of the house, below which appeared no indication of earlier occupa- tion fill. The conclusion, therefore, was that later disturbances had modified the visible surface de- pression of a house originally constructed over Excavation Detail 27 ...... ,. -\B ••■'' «»"^ ■** "'" ^"^ "<«»" around/ SECTION A - A SECTION D-D Hc 2 meters modern sod ^ ^^3 rock wood © post ..rt^j^jg^ charcoal -- projected (J"; pit O post hole Fig. 1 3. Plan and sections of House 3. 28 Part Three undisturbed ground, to give an erroneous picture of the house's true orientation: what seemed to be backdirt from the excavation of the next house to the east, designated House 5, overlay much of the House 4 hearth. House 5 Unhke House 4, here a 3 x 3-m excavation re- vealed the substantial rock-lined hearth, about 1 .0 X 0.7 m in plan size, with its long depression north-south, to be in the center of the visible sur- face depression. Like House 4, the relatively shal- low overburden, the modest overall size (about 5 X 5 m), and the lack of evidence of any structure beneath the hearth and center of the floor indicated that we would have done well to begin the season by excavating these outliers, which might have provided a simpler introduction to the archae- ology of the Paugvik houses than did the habita- tions we actually chose to begin with. The most unusual find from House 5 was the single gun part recovered from the Paugvik site. Together, Houses 4 and 5 appeared to represent two generations of structures in the eastern edge of the remnant Paugvik site. House 6, Area 6A The eastern end of the continuous portion of the 1973 trench was suspected of having penetrated a habitation of some kind, although there was no sign of a house on the existing surface. It was partly the hope of exploring this possibility further that dictated the placement of Trench 1 of 1985 next to the eastern end of the 1973 cut. The lower por- tions of Trench 1 were uniformly frozen, but when the base of occupation material was finally reached at its western end, the suspicion of 1973 was con- firmed by the presence of the substantial rock- lined hearth that, on the basis of evidence from other houses being uncovered, was supposed to mark the approximate center of a semisubterra- nean structure (Fig. 14). Accordingly, a 3 x 6-m cut was laid out north of Trench 1 to open more of the presumed house, although the degree of frost encountered at the base of Trench 1 made it clear that excavation of the new section would not be speedy. This northern cut was designated Trench 2 in the field, but for present purposes the house revealed by Trench 2 and the western 5 m of Trench 1 is designated House 6, which is described here as a unit. The eastern edge of the house was discernible in the wall profiles of both northern and southern edges of the cut (Fig. 14, profiles NIO, N13), and the limit of the floor within the trenches was thought to be located accurately, although a jum- ble of preserved logs and sticks, probably collapsed from a wall or roof, tended to obscure portions of the actual floor edge. For present purposes the section judged to be outside of the house within Trench 2 and the western 5 m of Trench 1 is des- ignated Area 6 A, probably but not certainly a hab- itation; the lower 20 cm in the appropriate areas is taken to be floor deposit of House 6 and Area 6 A. As suggested by the number of organic items, preservation of the House 6 and Area 6A floors and of the excavated portion of the House 6 en- trance tunnel was excellent, yielding grass cordage, much of the hair collected from the site, and wood- en artifacts, including mask parts and five clear examples of flat wooden pelt stretchers. In Area 6A, evidently predating House 6, there were three fairly well-defined pits, on an apparent (house?) floor at the approximate level of the floor of House 6, the easternmost of which (Fig. 14) yielded a number of flat sections of worked wood that at first were thought to be remnants of skin stretchers such as were found on the floor of House 6 itself, although examination in the laboratory cast doubt on this initial interpretation. The sec- ond pit yielded major fragments of twined netting, thought to be a fishbag, and the third produced an evident cache of leafstalks of the spreading wood fern {Dryopteris expansa (Presl) Fraser-Jenkins &. Jermy), a native foodstuff'of the region. It is likely, although not clearly demonstrable from the 1985 evidence, that this section of floor represented the still earlier house betrayed by the westernmost ab- original cut indicated in Figure 10, which was part- ly destroyed in the construction of House 6. Thus, despite the nicely frozen condition of House 6, its invisibility from the modem surface, and the ab- sence of occupation debris immediately beneath it, that house was almost certainly not the earliest habitation in its part of the site but was rather at least a second- and probably a third-generation structure in that vicinity, to judge by profiles of Trench I (Figs. 10, 14). Unfortunately, the slowness of thaw in the deep overburden, which totaled about 1 m, ruled out further extensive expansions of excavations in House 6 in the time available. But in the last days Excavation Detail 29 E20 E22 E24 E26 modem sod <^ (^3 rock c2iI^S3) wood ® post '«»- ^__,^^ «v^«Sjp»^ charcoal P ~— ^ ^®j___2;0rganic projected ( ~' pit N13 — N11 — N10 — — 96.00 L^^Lj layers PROFILE AT N13 fish bag 0(t\ ® V^^organic layers^ 96.00 PROFILE AT N10 Fig. 14. Plan and profiles of excavated portion of House 6 (grid in meters). of the season a 1 x 1.5-m southward extension at the west end of Trench 1 permitted the complete exposure of the House 6 hearth. For purposes of artifact provenience both the eastern 2 m of the continuous section of the 1973 trench (sect. 9) and the first detached segment of the trench to the east (sect. 1 1) were counted as portions of House 6. Trench 4 As excavations of Houses 2 and 3 progressed, it was speculated that the relatively flat area near their entrances might produce an informative sample of midden material associated with one or both houses. Accordingly, a 2 x 6-m trench was 30 Part Three laid out at that place, oriented north-south on the site grid, and was desodded to permit thawing. As the season wore on, however, it became clear that not all excavations projected could be completed, and so the area finally excavated was reduced to 1 X 6 m (Fig. 9). Work came to a hurried close as at least some glacial till showed throughout the length of the trench at depths of 1.0-1.4 m below the modem surface. Materials recovered did ap- pear to represent the midden expected; the matrix was jumbled and without definable strata. The sec- tion of the trench almost immediately in front of the entrance to House 3, in particular, yielded plentiful bone remains that appeared to represent largely animals of fur-bearing species that were mostly articulated at the time of deposition, pre- sumably skinned carcasses thrown out in front of the house. With cessation of excavation of Trench 4, the summer's work was brought to an end. All units were backfilled. Stratification of Cultural Features All major units of excavation revealed evidence of a sequence of construction at the Paugvik site and within the relatively brief temporal limits of the historic Pavik phase. There are indications of two generations of house structures both at House 1 and at combined Houses 4 and 5, of apparently three generations of construction in combined Trench 1 and House 6, and of at least three and probably four generations of structures represent- ed at combined Houses 3 and 4. Excavation Detail 3 1 4 Collections In the following discussion, artifacts from the Paugvik site excavated in 1961, 1973, and 1985 are described under three headings: procurement network, maintenance network, and protective network. Within these three broad categories, fur- ther subdivision was made on the basis of the activity for which the artifacts were intended. Al- though no exhaustive comparative treatment is attempted, some comparative data derived from Nelson (1983) and from published and unpub- lished reports dealing with sites closest to Paugvik, both spatially and temporally, are included with the descriptions when relevant. Numbers in pa- rentheses refer to numbered items in Table 2. Procurement Network Hunting Recovered objects associated with sea and land hunting reflect the diversity but not the complexity of early historic Eskimo weaponry. The toggle har- poon head ( 1 ) is represented by a single antler spec- imen, an antler spur fragment, and an ivory frag- ment. The complete head has a blade slit parallel to the round line hole, a closed socket (which is broken) and a single spur. Incised lines on the spur below the line hole depict a human face when the head is held upside down. A single incised line extends along one side to the tip of the blade slit (Figs. 1 5, 44h; Dumond, 198 1 , PI. XVII, Ab). Small harpoon heads like this one were probably used with a light sealing harpoon thrown with the aid of a throwing board. A similar head was recovered from the Old Togiak site on Togiak Bay (Kowta, 1963, pp. 68, 71, PI. 5c). The spur fragment in- cludes the lower half of a grooved line hole from which a straight incised line extends to near the end of the basal spur. The ivory fragment includes one side of the blade slit. There are eight harpoon dart heads of antler in the collection, seven of which are complete or nearly so. Six of the complete heads and the in- complete specimen, here designated type 1 (2), are identified as having been used with a light sealing harpoon (Nelson, 1983, Pis. XIV, LV 1-5). They are symmetrically or asymmetrically barbed bi- laterally, with a centrally located, triangular line hole. The complete heads have sharp or sloping shoulders and plain conical tangs (Figs. 1 6h, 44i- k; Dumond, 1981, PI. XVII, Ah-k). Similar har- poon dart heads have been recovered from a num- ber of late prehistoric and historic sites in south- western Alaska, including Hooper Bay village in the Yukon delta (Oswalt, 1952a, p. 49, PI. 1, 2- 5), Old Togiak (Kowta, 1963, pp. 78-79, PI. 7), and House 15 at Chagvan Bay (Staley, 1990, p. 239, Fig. 50e,f), and from earlier excavations at the Paugvik site (Larsen, 1950, Fig. 55 A, 2). The eighth dart head, designated type 2 (3), is heavier and may have been used for taking salmon; barbs are on one side only, and the round line hole is off"-center; the shoulders slope to a wedge-shaped tang (Fig. 1 6f). This style of harpoon dart head has been previously reported from Old Togiak (Kowta, 1963, pp. 132-136, PI. 19), Platinum South Spit on Goodnews Bay (Larsen, 1950, Fig. 55B, 3), the Tikchik site on the Nushagak River (VanStone, 1968, p. 58, PI. 8, 5-9, 12), and late prehistoric sites of the upper Naknek River drain- age (Dumond, 1981, PI. XVII, Ad-j). The collection contains two harpoon foreshafts (4), one of antler and the other of ivory. The antler specimen is broken at the proximal end but prob- ably had a wedge-shaped base. There is a centrally Collections 33 Table 2. Distribution of artifacts and detritus from Paugvik." HI H2 H2A Description A B HIA A B A B Procurement network Hunting 1 . toggle harpoon head 2. harpoon dart head, type 1 3. harpoon head, type 2 4. harpoon foreshaft 5. harpoon socket piece 6. float mouthpiece 7. bladder float plug 8. harpoon ice pick 9. wound plug 10. lance blade sheath 1 1 . bow fragment 12. arrowhead 13. blunt arrowhead 14. metal end blade 15. slate end blade, type 1 16. slate end blade, type 2 17. slate end blade, unclassified 18. chipped proj. point 19. arrow shaft 20. gun side plate 21. bullet mold half 22. boat or meat hook Fishing 23. lurehook 24. lurehook shank 25. barbless antler point 26. leister prong 27. fish spear point 28. net weight 29. net float 30. net mesh gauge 3 1 . fish scaler Trapping 32. pelt stretcher Transportation 33. kayak deck beam 34. kayak keel protector 35. umiak rib or riser 36. sled stanchion 37. sled upright 38. sled runner 39. sled shoe 40. snowshoe crosspiece Maintenance network Tools and manufacturing 4 1 . antler splitting wedge 42. steel wedge 43. wooden maul 44. metal axe head 45. stone adze blade 46. stone skin scraper blade 47. skin scraper blade blank 48. crooked knife handle 49. crooked knife blade 1 2 34 Part Four Table 2. Extended. H3 H4 H5 H6 H6A Tl T4 73T 61T Total 1 1 1 2 1 1 1 9 7 3 1 14 3 3 7 1 2 17 2 1 3 1 2 2 10 2 7 93 17 19 I 3 1 1 1 5 8 13 1 20 3 1 1 1 1 1 1 1 1 30 I 30 8 1 2 2 1 1 1 Collections 35 Table 2. Continued. HI H2 H2A Description HIA 50. composite knife handle 5 1 . end-bladed knife blade 52. rodent incisor knife 53. engraving tool 54. metal knife or engraver 55. whetstone, type 1 56. whetstone, type 2 57. whetstone, type 3 58. whetstone, type 4 59. whetstone, type 5 60. stone saw 6 1 . metal bladed ulu 62. metal ulu blade 63. stone ulu blade 64. untyped ulu fragment 65. ulu handle 66. metal scissors 67. awl 68. stone scraper or knife 69. bottle glass scraper 70. pick or mattock blade 7 1 . shovel blade 72. rake prong 73. ice pick or chisel 74. snow beater 75. unidentified metal object 76. sawed slate blanks 77. misc. sawed slate pieces 78. misc. polished stone 79. chipped bifaces 80. ochre anvil 81. hammerstone Household equipment 82. compound vessel 83. vessel side fragment 84. vessel bottom fragment (2 types) 85. spoon 86. ladle 87. dipper 88. water bag nozzle 89. nozzle or float part 90. large bag fragment 9 1 . mat or bag fragment 92. grass cordage 93. birch bark basket 94. metal kettle parts 95. brass box 96. pottery lamp 97. stone lamp 98. bottle glass 99. chinaware fragments Naknek ware potsherds 1 00. thin plain, variety unknown 101. thin plain, Pavik var. 102. thin plain. Camp var. 103. thin plain, Brooks R. var. 104. thin plain, exterior ridged 105. thick plain, variety unknown 106. thick plain. Camp var. 13 4 2 3 8 2 5 1 146 26 20 1 7 9 6 5 1 3 2 3 1 3 1 4 1 36 Part Four Table 2. Extended. Continued. H3 H4 H5 H6 H6A Tl T4 73T 61T A B A B A B Total 1 2 3 2 1 1 1 1 1 2 2 / 2 ._ 1 1 1 4 3 1 1 1 1 8 8 3 1 1 3 1 2 1 3 1 3 1 2 1 2 3 1 10 1 I 5 1 1 3 1 1 1 1 1 1 4 2 1 1 2 1 3 3 2 1 1 3 9 9 - 2 11 2 IS 2 68 1 1 8 9 12 33 2 1 1 1 1 1 2 10 2 1 1 4 18 1 8 3 f 4 2 1 7 3 3 2 1 1 3 1 17 1 1 2 2 3 1 11 1 1 1 1 1 5 2 1 1 1 3 1 2 2 1 2 2 7 1 3 2 1 1 1 8 1 1 1 1 1 4 1 1 1 3 2 2 11 , 4'^ 6 . 6 3 4 2 2 5 2 6 48 135 147 5 9 63 41 30 4 60 28 127 305 1163 29 28 1 49 6 69 2 5 14 40 8 250 14 il A 1 6 10 2 1 6 2 14 2 57 2 1 2 1 1 7 Collections 37 Table 2. Continued. HI H2 H2A Description B HIA Personal adornment 107. glass bead 35 108. 109. 110. 111. 112. native bead ring bracelet necklace segment hair comb Smoking equipment 113. snuffbox (?) 1 14. birch fungus Toy 115. bow Ceremonial objects 1 1 6. mask, unfinished 1 1 7. mask appendage 118. figurine Miscellaneous 1 1 9. sweatbath respirator Protective network Clothing 1 20. mukluk sole fragment 121. skin garment fragment 122. skin patch 123. gut raincoat (?) fragment 124. button 125. shoe fragment 1 26. sewn skin fragment 1 27. cut skin fragment 128. uncut skin fragment 1 29. knotted sealskin line 1 30. knotted baleen 131. wool cloth fragment Imported building material 1 32. window glass fragment 133. nail 134. screw 135. mica fragment 136. brick fragment Unidentified objects Wood 137. stake 138. unidentified Antler, ivory, bone 139. unidentified Debris 1 40. slate chips and chunks 141. chert, quartzite chips 142. pumice pieces 143. bone fragments 144. ivory fragments 145. antler fragments 146. iron fragments 55 12 1 1 1 11 5 1 1 1 7 1 1 3 3 7 2 2 6 2 21 8 5 3 1 1 1 -. 9 I 3 1 1 I 3 2 1 3 2 5 1 3 4 1 2 1 12 8 3 3 2 4 1 1 38 Part Four Table 2. Extended. Continued. H3 H4 H5 H6 H6A Tl T4 73T 61T Total 85 30 49 20 32 193 44 281 36 909 1 2/ 1 1 1 12 S 4 1 13 1 14 5 8 4 1 1 2 6 3 3 2 1 5 1 2 2 4 3 48 18 3 64 1 c 7 AS 2 1 1 3 1 13 1 11 1 2 1 2 1 1 1 _ 2 8 9 5 3 1 3 11 21 3 5 3 1 1 3 1 11 11 30 107 7 1 1 4 3 27 2 2 1 18 4 3 29 11 101 6 5 8 10 70 Collections 39 Table 2. Continued. Description HI H2 H2A HIA 147. brass, copper fragments 148. cut baleen 149. mammoth tusk, tooth Hair (no. samples) 1 50. Canis (dog, wolf, fox) 151. Cos/or (beaver) 1 52. Homo sapiens 153. <9rt^a/ra (muskrat) 1 54. Phoca (harbor seal) 155. Rangifer (caribou) 1 56. Urst4s (bear) " H = house; T = trench. 73 = trench dug in 1973; 61 = trench dug in 1961. Level A includes everything above the basal floor; level B is the lowest floor deposit. located, elongated line slot with incised lines ex- tending from each end; the specimen is round in cross section at the distal end (Fig. 1 6i). The ivory foreshaft, much larger and heavier, has an asym- metrical tang and an oval line hole from which extends a pronounced line groove on each side (Fig. 44m; Dumond, 1981, PI. XVII, Al). Seventeen objects are identified as harpoon socketpieces (5), only six of which are complete or nearly so. Four are similar in form, being drilled Fig. 15. Toggle harpoon head (point is downward). at the distal end to receive the dart head and hav- ing sharp shoulders and plain conical tangs (Fig. 1 6a, c, d); on one specimen the tang is asymmet- rical (Fig. 16c). Three are made of ivory, and one, which is not drilled at the distal end, is made of walrus penis bone. One of the ivory socketpieces has a projecting piece in the center of the drilled hole that presumably served to wedge the tang of the dart head in place (Fig. 16d). An ivory sock- etpiece is blunt and heavy with sharp shoulders and a rectangular tang (Fig. 44a; Dumond, 1981, PI. XVII, Bj); a badly weathered specimen of wal- rus penis bone has sloping shoulders and a plain, conical tang (Fig. 44d; Dumond, 1981, PI. XVII, Bi). These socketpieces would appear to have been used with a thrusting harpoon having a float of seal intestine similar to a "sea otter harpoon" col- lected in Bristol Bay in 1 88 1-1 883 by C. L. McKay (described and illustrated by Mason, 1902, p. 293, PI. 1 2). A similar socketpiece from House 1 5 at Chagvan Bay was described and illustrated by Staley (1990, pp. 245-246, Fig. 52b). A single unfinished harpoon socketpiece is made of ivory and has a long bifurcated tang. It is rough- ly worked and not drilled at the distal end (Fig. 16e). Another incomplete specimen of antler ap- parently had a wedge-shaped tang (Fig. 1 6b). Four ivory socketpieces are so badly weathered that their form when complete cannot be determined with certainty. One of these contains, in the drilled end, a wooden plug that permitted a more secure seat- ing of the dart head. Two socketpiece tang frag- ments, one of bone and the other of ivory, have sharp shoulders and are asymmetrical with pro- nounced knobs near the tip (Fig. 44e; Dumond, 1981,P1. XVII, Dc). 40 Part Four Table 2. Extended. Continued. H3 H6 H6A A B H4 H5 Tl T4 73T 61T Total 3 y 1 5 4 4 1 2 8 1 4 1 1 1 21 1 6 4 17 3 2 2 12^ 4 1 6 2 2 2 1 1 10 The remaining three socketpiece fragments, one of ivory and two of bone, were apparently used with much Ughter implements, possibly feathered harpoons thrown from a kayak with the aid of a throwing board (Nelson, 1983, PI. LIV). These are basal fragments with wedge-shaped tangs; one has a slight projection near the tip (Fig. 16k,l). Sock- etpieces similar to all the forms from Paugvik were recovered from the Old Togiak site and were de- scribed and illustrated by Kowta (1963, pp. 73- 78, PI. 6). The collection contains two ivory float mouth- pieces (6), one of which is incomplete. The com- plete specimen has an enlarged lip and a projecting spur with a drilled hole at the proximal end (Fig. 1 6g). The incomplete mouthpiece is for a smaller float of the type that is attached to the shaft of a harpoon (Fig. 16j). There is a projection at the proximal end that is pierced for the attachment of a line to bind the mouthpiece to the harpoon shaft (Nelson, 1983, p. 142, PI. XVI top, 17, 21, 25). A similar mouthpiece was recovered at Paugvik by Larsen (1950, Fig. 55 A, 9). A very small, round ivory object 0.8 cm in di- ameter and deeply grooved for lashing may be a bladder float plug (7). A probable harpoon ice pick (8) of antler tapers to a point at the distal end and has a drilled line hole approximately in the center (Fig. 1 6m). This implement may have been simply hafted to a shaft and used as a fishing ice pick. Two additional ob- jects are pointed at one end and may also be ice picks. One of ivory has a deep groove along one side (Fig. 44g; Dumond, 1981, PI. XVII, Cd), while the other of antler is thinned at the proximal end (Fig. 44f; Dumond, 1981, PI. XVII, An). A wooden object, rounded at one end and notched on either side, tapers toward the opposite end (Fig. 16s). It may be a wound plug (9) used to stop the flow of blood from a carcass. The collection contains two wooden lance blade sheaths (10), convex on the outer surface and hol- lowed out on the inside, for the protection of stone or metal blades. There are lashing notches at the point of maximum width (Fig. 1 6n). A similar sheath was described and illustrated by Nelson (1983, p. 146, PI. LVIIa, 27). Three comparable two-piece sheaths are included in an undated cache of hunting weapons (never published) that was accidentally encountered by a USGS geologist near the shore of Lake Grosvenor in the upper Naknek River drainage system in 1964. (.As of 1985, the sheaths, four associated shaft fragments, and eight polished lance heads were held by the National Park Service in Anchorage.) Two spruce bow fi-agments (11) are ovoid in cross section with nocks that are simple rounded projections with sloping shoulders. The smaller fragment may represent a toy (Figs. 1 6r, 44c; Du- mond, 1981,P1. XVII, Cc). Arrowheads (12) for large game or war are rep- resented in the collection by 10 specimens, only two of which are complete; all are made of antler. One complete arrowhead has a pair of barbs on one side, a sharp shoulder, and a conical tang; there is no blade slit. An incised line along one side may be an ownership mark (Fig. 1 6p). The other com- plete arrowhead is short with paired barbs at the slightly broken tip, sharp shoulders, and a plain conical tang (Fig. 441; Dumond, 1981, PI. XVII, Be). Two specimens, apparently unfinished, have sharp shoulders and plain conical tangs but no Collections 41 barbs (Fig. 16o,q). On one there is a series of par- allel incised lines, possibly ownership marks (Fig. 1 60). Of the remaining six fragments, three are tips (two with blade slots and one with a pair of barbs along one side), two are center sections showing a single barb, and one is a basal fragment, rectangular in cross section, with a conical tang (Fig. 16t). Two blunt arrowheads (13) of wood for use as bird arrows are roughly the shape of an elongated diamond. From a sharp tang, broken on both spec- imens, they swell to a point approximately one third of the distance from the tip and then taper to a point at the distal end (Fig. 1 6u, v). The collection contains seven metal end blades (14), two of brass and five of steel. The brass end blades are very thin and flat across the base (Fig. 17d; Dumond, 1981, PI. XVII, Db). Three steel blades are similar in shape but heavier (Fig. 1 7c; Dumond, 1981, PI. XVII, Dh), and the other two are long and narrow with short tangs (Fig. 17b; Dumond, 1981, PI. XVII, Dg). Because toggle har- poon heads are nearly absent from the Paugvik collection, it is assumed that these blades were used primarily with arrowheads. There are 1 29 whole and fragmentary slate end blades (15-17), presumably for the same purpose as the somewhat rarer metal end blades. These Paugvik insert blades are divided into two types, of which the first is by far the more common. Type 1(15) has a faceted butt, always more than 20 mm in length, that extends a variable distance over each face of the blade and was almost invariably formed by rubbing a narrow whetstone lengthwise to the blade, often grooving it deeply. Relatively whole examples among the 81 blades identified vary from 40 to nearly 90 mm in length and are 1 7-30 mm in maximum width (Fig. 46g-l). Type 2 (16) is similar in overall form and size, but the butt facet has been carved out with abrupt edges (Fig. 46m,n). Those remaining unclassified (17) are too fragmentary for recognition or are variant, most of the latter being very thin and lacking facets but of shapes variable enough that they form no coherent type; a few are thick and may have been in process of manufacture. Those that retain signs of their mode of manufacture were formed largely by abrasive sawing, snapping, and subsequent grinding, although some chipping before grinding is also in evidence (Fig. 46e,f). Butt-faceted slate insert tips similar to type 1 blades are found in many sites in northern Alaska (e.g., Ford, 1959) and are especially common in late prehistoric and historic sites around the south- em Bering Sea. They appear in the upper Naknek drainage sequence after a.d. 1000 and become the dominant projectile arming device after a.d. 1400 in the Brooks River Blufls phase (Dumond, 1981), at about which time they also appear on Kodiak Island (Jordan & Knecht, 1988; Dumond, 1991). Inserts of type 2 appeared in the Naknek region only with the beginning of the historic period (i.e., after about a.d. 1 800), with a geographical distri- bution much more limited than that of type 1, although they have been reported from some of the latest sites on Kodiak Island (e.g., Clark, 1974, PI. 16P). Larsen (1950) suggested that such im- plements with deeply carved facets that he recov- ered from Paugvik in 1 948 were derived from cast brass prototypes, but there is no evidence for this. The single chipped projectile point (18) is rem- iniscent of artifacts from the Naknek region of the early first millennium a.d. (i.e., of the Smelt Creek phase [Dumond, 1981]), although with a length of 56 mm and a width of 24 mm it is somewhat more elongated than is common in that phase. Although it might be compared with the form of some chipped points of later periods from northern Alaska (e.g.. Ford, 1959, Fig. 64), the presence of only the one example— from House 1 , from which a few other finds are reminiscent of Naknek River drainage implements of the first millennium— sug- gests rather that it is somehow derived from some earlier deposit (Fig. 46a). Arrow shafts (19) are represented by three frag- mentary specimens. The first is incomplete at both ends and has a diameter of 9 mm. The second includes the nock and, as is usual with the prox- imal ends of arrows, is flattened and oval in cross section (Fig. 1 7e). The third arrow shaft fragment is complete at the distal end, where there is a deep notch 3.5 cm long and pointed at the lower end to receive the tang of the arrowhead (Fig. 1 7k). A cast bronze gun side plate (20), slightly curved at one end, has holes at both ends to receive the screws that hold the lock to the gun stock. Initials stamped on the back are either "HD" or "HU" (Fig. 1 7g). This side plate probably was part of the lock mechanism of a shotgun. The collection contains one bullet mold half (21) made from medium-grain sandstone, rectangular in shape with a prepared flat surface into which has been ground a circular depression 0.8 cm in diameter. At one end of the depression is a groove that, when the identical other half of the mold was tied or otherwise affixed to this one, would permit the lead to be poured in (Fig. 1 7j). This stone mold may have fitted into a wood or antler handle re- 42 Part Four sembling those in ethnographic collections (e.g., Fitzhugh & Kaplan, 1982, p. 167; Nelson, 1983, PI. LXIII, 8). In southwestern Alaska similar molds have been recovered from archaeological sites at Crow Village on the Kuskokwim River (Oswalt & VanStone, 1967, p. 31, PI. 2, n), Akulivikchuk on the Nushagak River (VanStone, 1970, p. 62, PI. 11,2), and the Nunakakhnak site on Kodiak Island (Knecht & Jordan, 1985, p. 29, Fig. 1 1). The point for a boat or meat hook (22) is made of ivory and has two oval lashing slots parallel to the flat surface that would lie along the shaft (Fig. 1 7a). This hook is smaller than most of the boat hooks illustrated by Nelson (1983, pi. LXVIII, 22- 25, LXXX, 1-5) and may have been used for drag- ging large pieces of meat. Fishing Fishing was presumably as important to the Paug- vik natives as it was to most coastal peoples south of Bering Strait, where there are great salmon runs. There is evidence in the collection to indicate the use of the three-pronged fish spear and leister pronged spear and probable use of the salmon har- poon, nets, and hook and line. The collection contains a single fish-shaped lure- hook (23) and five lurehook shanks (24). The com- plete specimen is small and was presumably used for taking small fish such as tomcod or sculpin. The fish-shaped shank of ivory flattens at the prox- imal end, where there is an oval line hole cut at right angles to the small nail at the distal end that serves as a barb (Fig. 1 7i). Five objects, four of ivory and one of antler, are identified as lurehook shanks; all are unfinished, are considerably larger than the complete lure- hook, and hence were probably intended for taking larger fish such as grayling or trout. One ivory specimen is drilled at the distal end for a barb (Fig. 1 7h) and another is drilled near the proximal end for the leader or line (Fig. 1 7f). Two ivory shanks are flattened at the proximal end, and the single antler shank is narrower and flatter than the others. Similar lurehooks collected throughout south- western Alaska were described and illustrated by Nelson (1983, pp. 175, Pis. LXVIII, LXIX, Fig. 48). Surprisingly, fish-shaped lurehooks are absent from the archaeological collection from the Old Togiak site (Kowta, 1963, p. 104). The collection contains eight slender, barbless pointed objects of antler (25), which are round in cross section. Three are identified as probable cen- ter prongs for the three-pronged fish spear. Al- though no matching side prongs were recovered, this type of implement was widespread throughout southwestern Alaska in the late prehistoric and historic periods. All but two of the pointed objects are complete and shoulderless, sloping to a pointed proximal end (Fig. 1 7m-o). Similar barbless points have been recovered at Hooper Bay village (Os- walt, 1952a, pp. 54-55, PI. 2, items 9, 10) and Old Togiak (Kowta, 1963, pp. 114-121, Pis. 12-14). The 1 3 items identified here as leister prongs (26), five of which are complete, are made of ant- ler; some may in fact have been bird spear side prongs (see Nelson, 1983, Figs. 42, 44, PI. LIX). Two of the complete prongs and one nearly com- plete example were found together and thus may represent pieces of a single leister; each has nine barbs (Fig. 1 71). The other complete prongs have eight, six, and four barbs, respectively. Similar weapon points were illustrated by Nelson (1983, Pis. LX, 1, LXVII, 2, LXVIII, 1, Fig. 44), and unilaterally multibarbed prongs were recovered at Hooper Bay village (Oswalt, 1952a, pp. 54-55, PI. 2, 1 1-13) and the Old Togiak site (Kowta, 1963, pp. 1 23-1 26, PI. 1 5) and from House 1 5 at Chagvan Bay (Staley, 1990, pp. 249-250, Fig. 50a-c). A piece of low-grade steel, perhaps a spike orig- inally, was flattened and pounded out at one end to form a barb. It is identified as difish spear point (27). Because the distal end is bent slightly, it may have been intended for use with a three-pronged fish spear (Fig. 1 7p). Similar steel points were re- covered from the Nushagak site (VanStone, 1972, p. 55, PI. 11, items 2, 6). The use of nets at Paugvik is indicated by 20 recovered net weights (28), 12 of bone, seven of antler, and one of mammoth ivory. Six of the bone weights are made from the curved, unsplit ribs of large mammals, probably beluga, cut to length and drilled at each end for suspension. The holes were placed vertically, or what would be edge-to-edge of the unmodified rib, and the surface is essentially unmodified except for flattening at the ends (Fig. 18i). Two of the weights are smaller, possibly formed of caribou ribs, and are worked on all sur- faces. In one case the suspension holes were drilled laterally, or flat-side to flat-side (Fig. 1 7t), and on the other vertically. Three bone weights are made from split sections of the material worked to a rectangular shape and with laterally drilled sus- pension holes (Fig. 18h). The 12th bone specimen is different, having been worked to a flattened sur- face at one end where there is a single suspension Collections 43 hole (Fig. 17q). This last weight may have been used with hook and Une. Of the seven antler net weights, six of which are complete, all except one are made of split seg- ments. The exception is an unmodified section of antler tine drilled vertically at each end. The other three are roughly rectangular in cross section with laterally drilled suspension holes (Fig. 18b). The net weight of mammoth ivory is a piece of the exfoliated outer surface of a tusk, roughly rect- angular in shape with suspension holes drilled lat- erally (Fig. 1 8d). Net weights similar to those from Paugvik have been reported from all coastal and riverine sites in southwestern Alaska. Two incomplete net floats (29) are made of Cot- tonwood bark, roughly rectangular in outline, with a rectangular gouged line hole for attachment to the net. In cross section these fragments have a rounded triangular form, thinner at the top and thick at the bottom (Fig. 1 7u). A piece of wood that may be another incomplete net float is ap- proximately 1 7 cm long and 6 cm wide and rough- ly rectangular in shape; there are no suspension holes. Bark net floats have been recovered from Hooper Bay village (Oswalt, 1 952a, p. 55), Tikchik (VanStone, 1968, pp. 283-284, PI. 6, 14), and Ak- ulivichuk (VanStone, 1970, p. 68, PI. 1 1, 1 1). A single net mesh gauge (30) is a made-over fragment of a bone sled shoe. The gauging distance of 6 cm may have been designed for nets for black- fish or herring (Fig. 17r). Wooden mesh gauges were recovered at Hooper Bay village (Oswalt, 1952a, p. 55) and Crow Village (Oswalt & VanStone, 1967, p. 32, PI. 3, k, o, p), and antler examples were found at Akulivikchuk (VanStone, 1970, p. 61, PI. 9, items 19,20). The broad, flattened area of a caribou antler tine has been modified so as to be concave along its working edge and taper at the proximal end to form a handle (Fig. 1 7s). This is tentatively iden- tified as a. fish scaler {3 \). Similar implements from Old Togiak are made of caribou scapulae (Kowta, 1963, p. 147, PI. 20). Trapping Aside from the presence of the hair and bones of fur-bearing animals, the only direct evidence for trapping is the presence of five y^ood pelt stretchers (32). Presumably they were made locally at Paug- vik rather than obtained as trade items. Hides of freshly skinned furbearers were turned and stretched on these frames and traded when dry. Of the five stretchers, four are complete enough so that their overall shape can be determined. One is long and narrow, the upper (nose) end being extremely thin and pointed and the lower end rounded and considerably wider. Approximately 40 cm from the lower end there is a triangular perforation (Fig. 19a). According to present-day Naknek trappers, this was a stretcher for fox pelts. The three complete shorter stretchers range in length from 49 to 63 cm, are broad at the lower end, and taper slightly to a rounded point at the upper end (Fig. 19b). One specimen has a series of vertical cuts on one surface and was evidently used secondarily as a cutting board. Naknek trap- pers identified these stretchers as intended for muskrat pelts. The single incomplete specimen was apparently once about the same size and shape as the muskrat stretchers but has been cut off" at the upper or nose end. Transportation Artifacts related to travel are poorly represented. There is a kayak deck beam (33) made from a single piece of spruce driftwood, presumably a curved tree stump (Fig. 1 8a). Data concerning the construction of a modem kayak at Hooper Bay suggest that a beam of this length and curvature would be positioned directly in front or in back of the cockpit (Zimmerly, 1979, Fig. 74, p. 95). A complete antler kayak keel protector or shoe (34) has a pair of holes with antler pegs for attaching the shoe to the kayak (Fig. 451). Keel protectors were used at each end of the vessel to protect the skin cover when the boat was drawn up on the beach. Evidence for the use of the large skin boat is restricted to a single umiak rib or riser (35). The lower end is notched where the rib would be fitted to the chine, and the upper end is slightly concave to receive the gunwale. On the inner side is a notch where a stringer would be attached. Approxi- mately 7 cm from the top is a drilled hole and the remains of a sealskin lashing for the attachment of the gunwale. There is a similar hole at the lower end for lashing the rib to a chine (Fig. 20b). A poorly preserved, wedge-shaped piece of wood with a rectangular groove at the upper end that does not completely penetrate the object (Fig. 2 1 f) is tentatively identified as a sled stanchion (36). Stanchions were mortised into the top of a runner and extended to the crosspieces that made up the bed of the sled. 44 Part Four Even more tentative is the identification of a sled upright (37) for the type of sled with a railing. This piece is wedge shaped at one end, above which is a large, oval perforation, and narrows at the other end, which is broken (Fig. 20a). This upright, if the identification is correct, may have been placed toward the rear of the sled, with the perforation intended to receive the handlebar. More certainly identified is a fragment of a sled runner (38) from the front of a sled. The piece has a slight upward curve and a flat area at the front to receive a crosspiece. On the side and in the flat surface are holes for lashing to hold the crosspiece in place. Along the lower surface of the runner fragment are holes for the pegs that hold the sled shoes to the runner. A number of wooden pegs are still in place (Fig. 21a). This fragment would have been sufficiently close to the front of the sled to need no slots in the upper surface to receive stan- chions. The 15 antler and 15 whalebone sled shoe (39) fragments range in width from 1 .5 cm to 4 cm and are as much as 1.2 cm thick, although most are much thinner. There are irregularly spaced holes in the shoe fragments for pegging to the sled run- ners; in no case are there grooves between the holes that would suggest lashing rather than pegging (Fig. 1 8c,f,g). Although it might be supposed that antler would be the most satisfactory material for pegs, the only pegs in place in a shoe fragment are of wood. The sled runner described above also has wooden pegs. A single wooden snowshoe crosspiece (40) is thinned at each end for mortising into the outer frame. Along one edge are three notches to receive the webbing (Fig. ISe). The absence of holes through which webbing could be strung suggests the relatively crude type of snowshoe with coarse sealskin webbing intended for use on frozen snow or on the rough surface of the sea ice (Nelson, 1983, pp. 213-214, Fig. 64). Similar crosspieces were recovered from the Hooper Bay Village site (Oswalt, 1952a, p. 67, PI. 5, 14) and Old Togiak (Kowta, 1963, pp. 177-178, PI. 25, k,l). Maintenance Network Tools A large percentage of the tools received from the Paugvik site are traditional Eskimo forms, al- though some incorporate materials of European origin, such as metal for blades. As a group, tools include heavy woodworking implements as well as finer woodworking and antler-carving imple- ments and skin-working tools. In addition, there are several implements associated with general maintenance such as rakes, picks, and a shovel. The most abundant tool in the collection is the antler- splitting wedge (41). A total of 30 were re- covered from the houses and virtually all levels of the trenches. The typical wedge is made of a sec- tion of caribou antler cut off" square at one end and worked to a wedge-shaped bevel at the other. On most of the wedges the bevel is unifaced to take advantage of the hard outer part of the antler for the working edge. Some shaping of the opposite face is evident on most, however. The wedges vary in length from 9 cm to 24 cm and average 14 cm (Fig. 20f,g, 21b-e). Only a few show signs of ex- tensive use. One is somewhat different, having been made from the heavy base of an antler and worked to a bevel at the distal end (Fig. 45 k). In addition to antler wedges, eight steel wedges (42) were recovered. Four are rectangular sections of low-grade steel, cold hammered at one end to a bifacial bevel (Fig. 20e). Two are made from iron spikes flattened at the distal end (Fig. 20d). A single specimen is a heavy, oval steel fragment cut off" squarely at the top and slightly tapered and rounded at the other end; it may be unfinished (Fig. 20c). The eighth specimen was apparently fashioned from a section of thin steel, round in cross section, possibly a machine part. It is ffat- tened at the proximal end, where there is a rect- angular notch, and flattened to a working edge at the distal end (Fig. 22d). All these wedges are heavily rusted. For driving wedges to split logs, a maul (43) of wood was used. The single example is round in cross section with a sharp shoulder and rounded handle. It shows signs of heavy use and may, in fact, have been discarded for that reason (Fig. 22a). Similar mauls of cottonwood were recovered at the Crow Village site (Oswalt & VanStone, 1967, PI. 4b). There are two iron axe heads (44). The first is roughly rectangular, with a slightly flaring edge and a thickened poll. The eye is teardrop shaped and contains a fragment of the helve (Fig. 22b). Nine- teenth-century axes were usually made in two steps. One end of an elongated, flat plate of iron was hammered out while hot and wrapped around a pattern to form the eye (Russell, 1967, p. 257). Then a piece of steel was inserted to serve as the edge and the joints were welded by heating and hammering (Peterson, 1965, pp. 18-19). Collections 45 The second axe head is broken, so that a whole section is missing on one side from the poll to a point near the edge. This specimen has a flat poll, a widely flaring edge, and a pointed lower lip that grips the handle (Fig. 22c). It closely resembles axe heads found on Umnak Island in the Aleutians, described and illustrated by Russell (1967, p. 296, Fig. 79b), at the Nunakakhnak site on Kodiak Is- land (Knecht & Jordan, 1985, pp. 26-27, Fig. 8), and at a site on the southeast coast of the Kenai Peninsula (Schaaf, 1988, p. 20, PI. XIV). Similar axe heads have also been recovered from Russian- American Company sites in Kodiak (Shinkwin & Andrews, 1 979) and Sitka (Bamett & Schumacher, 1967). Of two stone adze blades (45), the one from House 2 is made from a slightly metamorphosed sedimentary rock that has a pronounced metallic sheen. It is roughly worked except for a finely ground working edge, which is V-shaped in cross section. The blade tapers toward the proximal end for insertion into a socketed head (Fig. 23f)- The form is that classed as Adze IV in prehistoric col- lections of the Naknek region (Dumond, 1981), where it is characteristic of the first millennium A.D. The second, from the portion of the 1973 trench that is now recognized as part of House 6, is slate, more smoothly polished and celtlike even though the thin blade is also polished only at the bit (Fig. 46s); classed as Adze II in the earlier analysis (Dumond, 1981), the form is more char- acteristic of the latest prehistoric period. The one stone skin scraper blade (46) is flaked with a finely polished bit on one end, appearing adze-like except for its overall narrow shape (Fig. 46u). This form in the earlier analysis (Dumond, 1981) was called End-Shaver II and is character- istic of the early first millennium a.d. A possible skin scraper blade blank (47) of vol- canic stone is roughly chipped on all surfaces, probably preparatory to the final grinding of a working edge (Fig. 23b). A crooked knife handle (48) is made from a slightly curved piece of antler. At one end is an open notch 5 cm long to hold the blade, which was presumably lashed in place (Fig. 23a). There are three metal crooked knife blades (49), curved at the distal end (Fig. 23c,d; Dumond, 1981, PI. XV, Fb). The collection contains three composite knife handle (50) halves, two of antler and one of wood, flat on the inner side and rounded on the outer surface. Both of the antler specimens have short, thin blade slots, possibly for metal blades, with raised lashing lips at the distal end. One has a raised lashing knob and narrow lashing grooves at the proximal end, and the other has only a single lashing groove in this position. There are three engraved circle-dot designs on one handle half (Fig. 23h). The wooden knife half has a longer, wider blade slit and a lashing lip at the distal end (Fig. 23g). Two end-bladed knife blades (5 1) of low-grade steel have long, thin tangs that narrow toward the proximal end (Figs. 44b, 45d; Dumond, 1981, PI. XV, Fa). A rodent incisor knife (52) has the bit still in place, hafted in a line with the long axis of a wood- en handle; there is a pronounced lashing lip. The handle is constricted toward the proximal end and on one side is a circular depression, which may have contained a glass bead or some other deco- ration (Fig. 23e). A complete engraving tool (53) has a badly cor- roded metal blade set into a slit in a crude wooden handle that has a pronounced lashing lip; the lash- ing is of narrow strips of baleen (Fig. 23n). Another example consists of only the distal end of the han- dle with a lashing knob and an asymmetrical metal blade (Fig. 45f). There are two metal knife or engraver fragments (54) that cannot be further identified with certain- ty. One is simply the proximal end of a metal blade embedded in part of an antler handle (Fig. 23o). The other is half of a composite antler handle, at the distal end of which is a broad slot and lashing knob. It may be part of an engraving tool (Fig. 23i). The 21 whetstones have for description been divided into five types based on the nature of the material from which they are made. The eight specimens of type / (55) are of granitic rock, with a variety of sizes and shapes represented. All are fragmentary and are worked on two or more sur- faces (Fig. 23p-r). The eight of type 2 (56) are of shale; all are fragmentary and have been worked on one or more surfaces (Fig. 231,m). The three type 3 whetstones (57) are fragments of pumice; two are small and have been worked on one surface (Fig. 22e), whereas the larger piece has a series of parallel, deep, narrow grooves on one surface and appears to have been used as a sharpener for items such as ulu blades and steel needles (Fig. 24). The single type 4 specimen (58) is of medium-grain sandstone, worked on all four surfaces (Fig. 23j). The single example of type 5 (59), of schist, is worked on the two narrow surfaces (Fig. 23k). The category stone saw (60), of which there are 46 Part Four three in the collection, is separated from other abrasive stones on the basis of form rather than material. Generally a relatively thin sandstone slab, one edge shows heavy wear on two intersecting planes (Fig. 46o). The function of the artifact is made clear by the numerous slate slabs with saw kerfs, illustrating the technique of abrasive sawing and snapping by which the plentiful slate projectile inserts, and presumably some slate ulu blades, were manufactured. The ulu or woman's knife is represented in the Paugvik collection by two complete metal imple- ments, three metal blades, one whole slate ulu blade, various fragments, and an incomplete wooden handle. The most impressive complete metal ulu (61) has a blade of low-grade steel with a semilunar edge and a large, thick wooden handle with a centrally located oval slot near the proximal surface. Narrow striations on both sides of the handle suggest that it was used occasionally as a cutting board (Fig. 25b). The other complete ulu is in very fragile condition. It also has a steel blade and a narrow wooden handle that turns upward at one end (Fig. 25a). Two of the metal ulu blades (62) are of a form that appears to be unique for the Paugvik site. Both have a semilunar edge, and extending from one end is a narrow metal strip that curves upward and over the top of the blade and ends in a tight circle or spiral (Fig. 25c,d). These blades are pre- sumed not to have been of local manufacture and could be used without the addition of a wooden handle. Although the shape of the handle of one of the complete ulus (Fig. 25b) seems to suggest that it covers such a curved appendage, at the time it was excavated the wood of the handle was wet and soft enough to permit examination of the haft edge of the blade, which was disappointingly square. The third metal blade is made from tinned steel plate of the type normally associated with the manufacture of tin cans. Flat across the top, it has a semilunar edge (Fig. 25e). The single whole slate ulu blade (63) is tabular in form and 4 mm thick, with a cutting edge nearly 60 mm in length (Fig. 46t). The type was earlier classed (Dumond, 198 1) as Ulu III. Six additional ulu fragments (64) may relate to the same type, although at least one of the fragments suggests the presence of a tang set off from the body of the blade. The unattached ulu handle (65) has a broad blade slit possibly intended to receive a stone blade with- out a tang (Fig. 45a). The collection contains a single pair of badly corroded metal scissors (66), apparently of fully modem form. The temptation is to consider these a much more recent intrusion into the site, but the provenience, essentially on the small piece of the floor excavated around the hearth of House 2A, seemed undisturbed and genuine enough at the time of excavation. Five objects of bone and antler have been iden- tified as awls (67). A seal scapula is sharpened to a point at one end (Fig. 25g) as is a caribou met- acarpus or metatarsus. The other three specimens are simply antler fragments worked to a tapering point at one end (Fig. 25h). Four retouched stone flakes apparently served as scrapers or knives (68). Three of these, retouched along one edge (Fig. 25f), presumably were used unhafted, but a fourth, from the 1 96 1 excavations, is partially wrapped with a strip of lead (Dumond, 1981, PI. XV, Cj), suggesting that it must have been wedged into a haft. Although the other three objects could belong to an earlier, prehistoric ar- chaeological horizon of the region, the fourth clearly does not. In addition, there are two scrapers made from retouched Augments of green bottle glass (69), both about 7 cm in thickness (Fig. 25i). Chipped glass scrapers have been reported from several his- toric sites in Alaska, and the form is also common elsewhere in North America. The collection contains two ivory pick or mat- tock blades (70), one of which is complete. The complete specimen is flattened along one surface, presumably for lashing to a wooden handle, al- though there are no lashing grooves. The working edge is beveled and slightly convex (Fig. 26a). The second blade is fragmentary; only the upper part is present. One surface is flattened, and there is a broad groove along one side. Approximately 1 1 cm from the distal end is a broad lashing groove (Fig. 25j). A shovel blade (71) is made from the shoulder blade of a large sea lion or walrus. The acromion process has been cut away, and a rectangular slot to receive the handle extends downward from the glenoid fossa for a distance of 9.5 cm (Fig. 26c). A similar shovel blade was recovered at the Old Togiak site (Kowta, 1963, p. 284, PI. 56a). Two rake prongs (72) are made of antlers. Oval holes for attachment of the handles have been drilled near the proximal ends, but the antlers are otherwise unaltered (Fig. 27b). Nelson (1983, pp. 74-75, PI. XXXV, 2) described and illustrated a somewhat similar rake from Sabotnisky on the lower Yukon, where rakes were used to remove refuse from the fireplace in the qasqig or men's Collections 47 house, for clearing away refuse material while building a house, and for clearing drift material from places where nets or fish traps were set in rivers and streams. Tentatively identified as an icepick or chisel (7 3) is a length of antler rounded and worked to a wedge shape at the distal end. The upper half of this implement is deeply recessed, presumably to re- ceive a long wooden handle. In the center of this recessed area is a round lashing hole, and there is a lashing knob at the proximal end (Fig. 27e). Three flattened pieces of wood, oval in cross section, are tentatively identified as snow beaters (74) for beating snow from clothing and other ob- jects. The two complete specimens taper slightly at the proximal end to form a handle (Fig. 27a,d). Somewhat similar implements from various lo- cations in Alaska were described and illustrated by Nelson (1983, pp. 77-78, Fig. 21). Unidentified metal objects (75) that are pre- sumed to have been intended as some form of tool include a section of gun barrel partially flattened at one end, possibly for use as a hide flesher (Fig. 28f), and a heavy iron ferule that has an attach- ment hole at the proximal end and tapers to the distal end, which is broken (Fig. 27c). A piece of steel with what appears to be a concave working edge may be the blade for an ulu. Numerous items of stone are apparently arti- facts in the process of manufacture. Six of these are rather clearly sawed slate blanks (76) for insert blades, completely cut to basic form, but not yet sharpened and faceted (Fig. 46p>-r). The majority of the rest are 5 1 sawed slate pieces (77) charac- terized by the presence of saw kerfs but of no ap- parent final shape. There are also 25 miscellaneous polished stone scraps (78), largely slate, that may be unidentifiable fragments of ulus or insert blades or, in a few cases, chips from the resharpening of stone adze blades. The 1 7 crudely chipped bifaces (79) of slate or shale are also presumably artifacts in process of manufacture (Fig. 46b-d). These are scattered through the site, but more than half were recovered from the 1973 trench. Of those recovered in 1985, six are lanceolate in form, 55-90 mm in length, and are presumably blanks awaiting polishing into lance heads. Although chipping before grinding is indicated for some Pavik phase artifacts, these bi- faces are also reminiscent of artifacts termed lan- ceolate biface classes I and II (i.e., those above and below 70 mm in length), which are especially char- acteristic of the Brooks River Camp phase of the early second millennium a.d. in the Naknek region (Dumond, 1981), when the basic shape of slate implements was formed before polishing by chip- ping rather than sawing. One of these from Paug- vik is a slate ellipsoid 50 mm long, again remi- niscent of the Camp phase. One slab of a fractured cobble is heavily stained with red hematite in its fortuitous basin, evidently from use as an ochre anvil (80) in crushing paint. There are eight heavily scarred pebbles that have been used as hammerstones or pounders (81). Household Equipment All domestic equipment not considered in previ- ous sections is described here, including wooden serving and storage vessels and utensils, pottery and stone lamps, other ceramics, and woven ma- terials. There are three wooden compound vessels (82) that are sufficiently complete so that their size and form can be determined with certainty. These ves- sels are of two-piece construction, consisting of a flat, oval bottom and a thin strip bent around to form the sides. The overlapping ends of the side pieces are fastened together by sewing strips of root through holes drilled for the purpose. Base pieces have chamfered edges to fit into a groove around the inner edge of the side pieces. The first of these vessels is complete except for a section of one side. The two ends of the side, which is 4.5 cm high, are lap-spliced with root through two parallel rows of slits. On the bottom of this vessel are two shallow incisions in the form of a cross (Fig. 26b). The second vessel is very shallow and complete but badly warped; a portion of the bottom is split. The two ends of the side, which is 3 cm high, are fastened together with root through a single row of slits (Fig. 28b). Much of the rim of the third vessel is missing, but it is clear that the two ends of the side piece were lashed together through two parallel rows of slits. In ad- dition to the groove on the inner edge of the side, four wooden pegs, one on each side and end, held the bottom in place. Seven fragments of compound vessel sides (83) were recovered, only one of which is complete enough to indicate the height of the vessel. This fragment is from a much larger container than the complete vessels just described and lacks a groove running around the lower edge to receive the vessel bottom (Fig. 28c). A much smaller fragment does show this groove (Fig. 28d). The rims on three 48 Part Four fragments are rounded. All fragments show lashing that held the two ends of the side together. On the two more complete fragments the ends of the side were fastened with root lashing through a single row of slits. The collection contains 17 fragmentary com- pound vessel bottoms (84), and on the basis of size and shape two types can be differentiated. The first type includes six fragments of oval vessel bottoms, none of which are chamfered to fit into grooves in the sides. Most are from relatively large vessels, the longest being approximately 29 cm in length, the smallest 8 cm (Fig. 28a,e). Three apparently consisted of two pieces of wood, probably equal halves, pegged together with wooden pegs. Both lashing and pegs were used to fasten two of these bottoms to the sides. Two bottoms were appar- ently used as cutting boards, perhaps after being discarded as vessels, and two are badly charred. The second type of vessel bottom, of which there are six in the collection, is very small. All were apparently round or nearly so and may be the bases of trinket or snuffboxes similar to those illustrated by Nelson (1983, PI. LXXXVI) rather than of household containers (Fig. 29i). It is also possible that one or more of these round, flat pieces of wood are poke stoppers associated with the storage of food or seal oil. However, they lack the deep lash- ing grooves usually found on plugs and stoppers. Nine relatively whole spoons (85), eight of antler and one of wood, and two additional fragments were excavated from the Paugvik site. Three of the antler specimens have elongated oval bowls and straight handles of various lengths (Fig. 29a,c,g). Two have deeper, more carefully shaped bowls; the handle of one widens at the proximal end (Fig. 29d), and the handle of the other has a pronounced curve (Fig. 45i; Dumond, 1981, PI. XVII, Dg). One specimen has paired parallel in- cised lines running around the edges (Fig. 29c), and another has a ribbed handle with a single straight incised line in the center of the bowl (Fig. 29a). The single wooden spoon is much more crudely made; it is a narrow strip of wood hol- lowed out at the wider distal end to form the bowl (Fig. 29e). Of the fragmentary spoons, one is the handle of a much larger wooden specimen broken off at the point where it widens to form the bowl (Fig. 30b). The other, of antler, consists of the bowl only. Running down the center is a single incised line that terminates in a Y pattern ornamented with spurred lines (Fig. 29f). Spoons similar to those from Paugvik were described and illustrated by Nelson (1983, p. 69, PI. XXX, 207) and were re- covered from the Old Togiak site (Kowta, 1963, p. 281, PI. 55, a-g). Five large spoonlike objects are identified as la- dles (86). Two antler specimens have spatulate bowls that are flat at the distal end (Fig. 30e,g). The wooden ladle consists of a bowl only, which is shaped like the bowls of the complete spoons but is larger (Fig. 300- Two ladles, one of antler and the other of bone, are simply large, irregularly shaped bowls that would have had separate, at- tached handles (Fig. 30c,d). The antler specimen has a pair of drilled holes at the proximal end for this purpose (Fig. 30c). These two objects may be small shovels. The Paugvik collection contains one complete dipper (87) and two fragments. The complete dip- per of wood is carefully made, with a bowl that has a flat bottom and sides that slope out toward the rim. The handle flares at the proximal end, which is rounded (Fig. 30a). A similar dipper, identified as a ladle, was recovered at Crow Village (Oswalt & VanStone, 1967, p. 35, PI. 5, f). A single small fragment from the point where the handle joins the bowl appears to be from a similar dipper. A large dipper fragment is made from a single piece of wood carefully fashioned into a thin han- dle at one end and thinned down to a wedge-shaped point at the other. The wood was then steamed and bent to form the sides of a circular bowl and lashed just inside the base of the handle (Fig. 30h). The bottom would have been a separate piece. This type of dipper, common throughout south- western Alaska, was described and illustrated by Nelson (1983, pp. 65-66, PI. XXIX, 6-8). A spoon-shaped water bag nozzle (88) of antler somewhat resembles similar objects from south- western Alaska illustrated by Nelson (1983, p. 74, PI. XXXIIIa, 5). These bags, made from the stom- achs or bladders of animals, were used to carry water or oil while on hunting trips at sea; they had wooden stoppers (Fig. 29j). The collection also contains two other antler nozzles that obviously are for containers of some sort (Fig. 29h,l). One of these has a projecting lip (Fig. 29h) and may have been a bladder float nozzle (89). Two large fragments of conical, loosely woven, twined grass bags (90) appear to be from those that according to Nelson (1983, p. 203) were used to hold fish. The tops consist of two parallel rows of two-strand braided grass. Bags with similar tops, although more closely woven, were illustrated by Fitzhugh and Kaplan (1982, p. 125) and Kaplan Collections 49 and Barsness (1986, p. 122). Fish bags of this type were also used on Nunivak Island (Lantis, 1946, Fig. 17 opp. p. 177). Twined work is also represented by seven mat or bag fragments (9 1 ) varying in fineness of weave. The coarser examples may be parts of sleeping maps similar to one described and illustrated by Nelson ( 1 983, p. 203, PI. LXXIV, 1 5), but all four could be bag fragments. All fragments may have been more tightly woven than they appear at pres- ent (Figs. 31-33). In addition to the twined frag- ments, the collection contains eight fragments of braided grass cordage (92) (Figs. 34, 35). The single fragment of a birch bark basket (93) indicates that the vessel was made from one piece of bark folded at the four comers and then stitched, probably with spruce root as indicated by the large and widely spaced stitching holes. Three small birch bark fragments may also be from baskets. Con- tainers of birch bark are commonly associated with interior Eskimo settlements in southwestern Alas- ka and have been recovered from the Crow Vil- lage, Tikchik, and Akulivikchuk sites (Oswalt & VanStone, 1967, pp. 47^8, PI. lib; VanStone, 1968, p. 283; 1970, p. 67, PI. 11, 13). The collection includes several metal kettle parts (94). A cast iron kettle rim fragment includes a circular lug welded to the rim and is from an ex- tremely large vessel (Fig. 30i). There are also two lugs for kettle handles of the type that was riveted to the kettle rim on opposite sides just below the lip. One lug is brass (Fig. 29k) and the other is cast iron (Fig. 45b). A brass kettle lid (94) has raised edges and a ring handle. Attached to the handle is a short strip of two-strand braided grass (Fig. 45j). A round brass box (95) has a convex top with recessed lower edges and a flat bottom (Fig. 45g). The saucer-shaped pottery lamp (96), wide- spread through southwestern Alaska, is repre- sented by four virtually complete examples and sherds that represent seven additional lamps. The complete specimens are all undecorated and are fired poorly, if at all. The temper of these is pre- dominantly grass, although some gravel can be noted in at least one. The walls are thick, and the pronounced rims are rounded. Three lamps are extremely shallow (Fig. 36a, b), while the fourth is deeper (Fig. 37a). Grass is also the predominant temper in the fragments, with one exception that appears to be tempered primarily with hair. All are poorly fired, and at least two are from lamps even shallower than any of the complete examples; on one of these fragments the lip barely projects above the surface. Oswalt (1952b, pp. 21-22) suggested that saucer- shaped clay lamps were derived from the conical- bottomed, wide-mouth clay lamp common in northern Alaska during the early phases of Eskimo prehistory. Early examples of the saucer-shaped clay lamp have been excavated from sites in the Kobuk River-Kotzebue Sound region, from which they evidently spread to the Bristol Bay-Norton Sound area. Four stone lamps (97) were also recovered at Paugvik. One of these is crudely worked from a roughly circular piece of granitic rock flattened on one side and hollowed out on the other. It is en- crusted with carbon (Fig. 29b). Two others are heavy stone spalls with fortuitous basins that ap- pear from carbon deposits to have been pressed into service as lamps. A fourth, also of granitic rock, has been carefully worked to an elongated oval shape. This lamp is shallow with a rounded lip and shows signs of use (Fig. 37b). Because it resembles lamps from phases of the first millen- nium A.D. (e.g., Dumond, 1981, PI. VI, Fc, PI. XI, De), it may have been salvaged by Paugvik resi- dents from earlier sites in the Naknek River re- gion. In addition to the bottle glass scrapers, the col- lection contains six small bottle glass fragments (98). Two of these, one green and the other brown, are approximately 0.8 cm thick; a third is a bottom fragment from a small bottle of clear glass. The remaining fragments are extremely small and thin, ranging in thickness from 1 to 2 mm. One is a fragment of a faceted bottle. Excavations at the Paugvik site in 1961 and 1973 yielded only eight nondescript chinaware fragments (99). In 1985, 40 fragments were re- covered. Most of those collected, like those from other historic sites in southwestern Alaska, are sherds of factory-made ironstone (earthenware), a utilitarian stoneware variant that was extremely popular during the 19th century, particularly in frontier areas, because of its strength and dura- bility. Chinaware sherds from the 1985 excavations were each assigned a serial number and then a potential vessel number, as nearly as such an as- signment could be made from appearance alone, for none of the sherds could be fitted to one an- other (Table 3). Looked at in this way the 40 sherds could come from no more than 3 1 vessels, but in two cases where sherds were indicated as possibly from the same vessel (nos. 3 and 24, and nos. 1 5, 34, and 37) the proveniences of the separate sherds were so widely separated that their origins in a 50 Part Fovu- single vessel seems unlikely. Thus the 40 sherds probably represent at least 35 different vessels. Factory-made ceramics are commonly the most voluminous trade goods excavated from historic sites in southwestern Alaska. Nearly 6,000 chi- naware fragments have been excavated from six published sites along the Kuskokwim and Nush- agak rivers and on Lake Clark (Oswalt & Van- Stone, 1967, pp. 52-55; VanStone, 1968, pp. 288- 292, 1970, pp. 74-81, 1972, pp. 55-60; VanStone & Townsend, 1970, pp. 75-86; Oswalt, 1980, pp. 70-73). Decorative types recovered from these sites consist primarily of plain, under-glazed lined, cut sponge-stamped, hand-painted, and transfer- printed wares. Although there is some late 19th- and early 20th-century American ironstone in the assemblages, the majority is the standard British export ware described by Jewett (1878) that sus- tained the North American market in the 19th century. British ceramics reached Alaska through the Russian-American Company, which found it cheaper and more convenient to obtain manufac- tured goods that reached the Northwest Coast on British and American ships rather than to rely on the long overland or ocean supply lines to Russia. In 1839 the Hudson's Bay Company contracted to supply Russian America with provisions and manufactured goods, and the agreement became effective in 1840. After 10 years, the agreement was not renewed (Davidson, 1941; Gibson, 1976, pp. 83, 139, 200-208). Nevertheless, the ceramic supply network, which came to include an increas- ing number of Staffordshire and other British pot- teries, continued after the sale of Alaska to the United States in 1867. The number of exotic contact-period ceramic fragments recovered from the Paugvik site is thus unusually small. There are no identifiable maker's marks, and only five patterns can be identified: "Willow," "Watteau," "Cherry Picking," "Cam- illa," and possibly "Davenport" (Fig. 38). Only the ubiquitous "Willow" pattern has been report- ed from all the other excavated sites in south- western Alaska, and the "Cherry Picking" pattern occurs at Crow Village (Table 3). The "Camilla" and "Watteau" patterns were recovered at the Nu- nakakhnak site on Kodiak Island (Knecht & Jor- dan, 1985, Table 1, p. 25). The number of decorated china ware sherds from the Paugvik site is also too small to permit mean- ingful comparison with other sites and thus at first glance seems not to provide new information re- lating to the chronological or distributional ques- tions associated with this particular trade item. It is certainly significant, however, that the residents of Paugvik apparently had restricted access to ce- ramics, as compared with the residents of other excavated village and trading post sites in the re- gion. Aside from the clay lamps, ceramics of aborig- inal type are of what has been classed as Naknek ware (100-106), one of two ware types known for the region prehistorically (Dumond, 1981). Nak- nek ware vessels are patch modeled and paddled against the hand, tan to black in color, baked in an open fire, and when found commonly have caked food residues in the interior. The range of shapes in any one period is limited. Techniques of clay treatment are poor, resulting in consider- able variation in frequency and distribution of temper, which is predominantly water-worn grav- el in such quantity as to result in a pronouncedly crumbly fracture, and varying directly in size with the vessel wall thickness. Grass may also be pres- ent, and temper fraction may vary significantly over different parts of the vessel. Naknek ware is then divided into two subclass- es, depending simply on thickness, in a division that has been shown to be temporally significant (Dumond, 198 1). Naknek thin ware has walls less than 10 mm in thickness and is often relatively hard. Naknek thick ware has walls of 10 mm or more, sometimes more than double that dimen- sion (Fig. 39). In the Paugvik collection there is no overall surface decoration, so that the only two types represented are Naknek thin plain, by far the more common, and Naknek thick plain, much of which in fact probably pertains to an earlier occupation in the vicinity. In keeping with the less than consistent manufacturing techniques, some otherwise thin vessels may have a few reinforced sections that in small sherds may be classed as thick; a few other thick sherds may actually be derived from lamps (Fig. 39L) rather than from the ordinary Naknek ware cooking vessels. Within each type, varieties are distinguished by vessel shape, chiefly indicated by rim sherds. There are four of these varieties represented at Paugvik. The Camp variety (102, 106) appears in both types but is far more common in thick plain. The variety is characterized by a globular shape with in-sloping lips that restrict a neckless opening (Fig. 39K,L). The base tends to be small, although not pointed but tapering to a flat area (Dumond, 1981, Fig. A.l). There are examples of this rim in the Paugvik collection but no examples of the base. The Pavik variety (101) is confined to the Nak- CoUections 5 1 Table 3. Chinaware sherds from the 1985 Paugvik excavations. Sherd Vessel no. Unif no. Description 1 HI 1 Base and foot of a transfer-printed cup (Fig. 38b). The letters ORT on the base are probably final letters of the word DAVENPORT, a factory at Longport in the Staffordshire Potteries. This firm, which exported widely to North America, was in existence from c. 1793 to 1887, and after 1850 their wares were normally marked with the name Davenport (Godden, 1964, p. 189). 2 HI 2 Plain (?) fragment of ironstone from a plate or saucer. 3 H2 3 Blue transfer-printed cup rim with handle junction (Fig. 38g), manufactured by the Copeland Spode factory. The pattern is "Watteau" (Sussman, 1979, p. 231) and has been recovered from Hudson's Bay Company sites in western Canada. 4 H2 4 Fragment of transfer-printed plate rim (Fig. 38c) manufactured by Cope- land and Garrett, Spode Works, Stoke, Staffordshire Potteries, between 1833 and 1847. The design is called "Cherry Picking" and dates from 1838. It was not recorded by Sussman (1979) and is not generally known to have been exported to North America (Louise M. Jackson, personal communication). However, it was recovered from the Crow Village site. Plain plate (ironstone) foot. Transfer-printed blue willow pattern border. ^ Plain fragment (plate?). -^ Transfer-printed blue willow plate soup rim with moulded ridge. Transfer-printed blue willow border of rim fragment. Transfer-printed blue willow pattern. Possibly the ball of a soup plate shoulder or rim. Plain or cream fragment, possibly from a soup plate or saucer. Blue transfer-printed cup body; staining on the inside. Plain fragment from shoulder of a soup plate. Brownware fragment of the lid of a storage vessel. Fragment of green transfer-printed cup (see comment to sherd no. 34). Plain plate body fragment with illegible impressed mark. Flake with no glaze, unidentifiable. Fragment from brownware storage vessel. Plain fragment of moulding around plate rim; ironstone. Rim fragment of porcelain bowl with plain pink band on the outside. Body fragment of brownware serving bowl. Blue floral transfer-printed cup fragment, possibly a Copeland and Garrett or W. T. Copeland piece (Louise M. Jackson, personal communication). Fragment of plate body without glaze on either side. Transfer-printed blue cup fragment (Fig. 38d) with the "Watteau" pattern (see Sussman, 1979, p. 231). Possibly same as sherd no. 3, although pro- veniences differ. Body fragment of utilitarian brownstone serving vessel. Basal fragment of utilitarian brownstone vessel with part of an impressed mark, enclosed in a circle; includes the final letters of two words, REENS in a curve at the top of the circle, and SIDE horizontally across the mid- dle. It has not been possible to identify this mark. Chip from a utilitarian brownstone vessel. Fragment of plate with blue feather edge. A hole has been drilled through the shoulder. Body fragment of blue transfer-printed plate. Fragment of blue transfer- printed plate. ' Plain plate or soup plate fragment. Plain cup fragment. Body fragment of blue transfer-printed cup (Fig. 38a) with the "Camilla" pattern manufactured by Copeland and Garrett and W. T. Copeland of Stoke, Staffordshire Potteries, from 1833 and still manufactured by Spode Limited (Sussman, 1979, p. 83). 34 T2 11 Fragment of green transfer-printed plate rim. It could belong to the same vessel as no. 15, and no. 37, although proveniences differ. The pattern design may be "Davenport IV," illustrated by Williams and Weber (1986, p. 168), made by the Davenport factory (see sherd no. 1). 5 H2 5 6 H2 6 7 H2 7 8 H2 6 9 H2 6 10 H2 6 11 H2 8 12 H2 9 13 H2 8 14 H2 10 IS H2 11 16 H2 12 17 H2 13 18 H2 14 19 H3 15 20 H3 16 21 H3 17 22 H3 18 23 H3 19 24 H3 3 25 H3 20 26 H3 21 27 H3 22 28 H5 23 29 HS 24 30 HS 24 31 HS 24 32 Tl 25 33 Tl 26 52 Part Four Table 3. Continued. Sherd no. Unit^ Vessel no. Description 35 T2 27 36 T4 28 37 T4 11 38 T4 29 39 T4 30 40 T4 31 Bowl fragment with hand-painted brown band on the outside. Plain plate body fragment with fragmentary unidentified impression. Green transfer-printed body fragment (see sherd nos. 1 5 and 34). Unidentified flake without glaze. Blue transfer-printed flake. Blue transfer-printed flake. '^ H = house; T = trench. nek thin plain type and has an unrestricted opening and sides tapering outward in flower-pot form, often with an additional slight flare at the lip (Fig. 39B-G). In some cases, this flare occurs above a very slight constriction after the manner of the so- called situla shape that has been described for his- toric-period ceramics in western Alaska to the north (e.g., Oswalt, 1955). The base of this variety is flat and relatively wide (Fig. 39M). The Brooks River variety (103) of the Naknek thin plain type has a form approaching that of a cylinder or barrel (Fig. 3 9 A). Although not gen- erally found in vessels of Naknek ware paste any- where in the region (where the shape commonly pertains to the earlier Brooks River ware with dis- tinctive fiber-tempered paste), some rims at Paug- vik cannot reasonably be assigned to any other shape. Unfortunately, the restriction at the lip of such vessels may be pronounced enough that rim sherds too small to reveal the conformation of the lower vessel walls can be mistaken for the lips of globular pots and so classed as Camp variety. This may be the case with some sherds in the present collection (Table 2). The base of vessels of this variety are indistinguishable from those of the Pa- vik variety (Fig. 39M). The exterior ridged variety (104) is represented by even fewer sherds. The total vessel shape is evidently that of the Pavik variety, of which this may be considered a variant, in which the wet clay was pinched into a pronounced horizontal ridge somewhat below the lip (Fig. 39H), as though to emphasize the thickened region that often occurs in that portion of the vessel walls (e.g.. Fig. 39B,D,E). This is the only approach to decorative treatment in the Pavik ceramic collection. In the Naknek region, the Naknek thick plain type. Camp variety, is characteristic of the period from about a.d. 1000 to 1450, the time of the Brooks River Camp phase (Dumond, 1981). The Naknek thin plain type in the same variety, in- cluding some vessels with exterior ridges, appears thereafter in the Brooks River Blufls phase. The Pavik variety of the Naknek thin plain type is then present in quantity only in the Pavik phase of historic times. In the present case, as has been indicated, the majority of the Naknek thick plain potsherds of the Paugvik site are thought to rep- resent earlier deposits located in the vicinity or in some cases to possibly result from misclassifica- tion of small fragments of clay lamps. Personal Adornment A total of 538 complete glass trade beads (107) were recovered from the Paugvik site in 1 985, with 317 recorded for the trenches of 1961 and 1973 (Table 4). Although most were probably used as items of personal adornment, it is probable that a few may have served to decorate other items of material culture. For present purposes, the 1985 sample is deemed of ample size for analysis, in- asmuch as an examination of the earlier material revealed no apparent difference in the range of types. The bead typology developed by Kenneth and Martha Kidd ( 1 970) is here applied to the Paugvik bead sample of 1985, although some problems were encountered in its use. Colors were some- times difficult to define or assign, and it was also difficult at times to separate precisely the "round" bead varieties from the "circular" varieties. As a result, there may be some mixture of such Kidd varieties as IVa6 and IVa7. A total of 1 16 beads, or 20% of the 1985 collection, could not be as- signed within the original Kidd classification, which for present purposes was expanded to accommo- date the Paugvik sample. Table 4 gives a complete list of the number and varieties of beads excavated at Paugvik in 1985. The 54 bead varieties from the site represent eight separate Kidd types. The most frequently occurring variety is Ila 1 4, circular in shape, opaque, Collections 53 ^^_-«-«cr)r-^ — — fN rf rr (N 00 fN m 1/^ mm t/f2vfc«>y32ij2(/3t/3~t/fi-lV3i-l>C/3i-l>i-l2w32i-liy5(/3C/32 i^ O a & 1» .s? 3 — ^ is aii o w 4J C U V CQ » p 5 c > C »] U u c c3 £0 aj 1) 3 «2 wi oS -^ pple gre urf greei nt light ght aqu qua blu qua blu II in's e in's e brigh ,_j 3 ♦- X) * ♦- c O" C o o n 9. ea Wi 4> .3 w w tfl « u u « 1 U U 3 (U (U U 3 3^333 ^1 U U 3 3 3 5 "en a o" ;2 o* a o" «i CO c« C C cd cd ^ cd cQ c^ 2 a a 2 ex a a o i: a a 2 o o i: Un O O t! o o o 3 " HHHHHHHHHHHHUUUttJOOiOittiftJUOiCiiUOoJoSUPiU o m ^ <<^ vo /^CT\ — -H — — — OOC«C/3"oOC/3'NjjSNjHJ-jS 3 SB. c w J5 c c « 4) . o2> bb JJ u u u u u 3 3 3 CT CT CT a Q. a o o o QiUUUUUHHh-Hf-OiOaJUUC^OiBSoJalOoJ — < «N fn O — »N <^ -JJ (N — — — — — IX S -_«-_«•'«<« «w -^^ ra cd cq cd cd cQ ^^ ^^ ^^ ^^ ^^ ^^ w^ [.H Is II s 3 ■•3 u 6 c 'S 2 — k> u (/) a 3 C C R « (u X V|" II -H HU>X> Collections 55 Table 5. Types and varieties of beads in the 1985 Paugvik sample. Table 6. Comaline d' Aleppo red beads from south- western Alaskan sites.'' Kidd no. No. of beads No. of present varieties'' la 165 11 Ic 1 2 Ila 200 25 Ilia 90 3 Illf 2 1 IVa 110 5 Wlb 14 7 Wlc 1 1 Total 583 54 " Kidd tyj)es have been expanded for present uses. See Table 4. and white. The second most common is Ia5, a tubular opaque white bead. The nine most com- monly occurring varieties account for 61% of the total bead sample. Rounded beads (represented by Kidd classes I and III) make up 53% of the sample. The greatest number of beads of any one type is 1 84 beads of type Ila, representing 20 different varieties. Type la, with 165 beads, is represented by 1 1 varieties at Paugvik. Of the 116 beads that could not be placed within the Kidd classification, 98 belong to types Ilia and IVa, tubular and cir- cular beads of opaque oyster white with a white interior. The assemblage includes only 15 wire- wound beads, or 2.5% of the sample. The total Kidd types and varieties, as expanded for the pres- ent analysis, are listed in Table 5. For a comparison of the Paugvik beads with those from other historic sites in southwestern Alaska, perhaps the most significant varieties are circular or round beads with opaque red exteriors and clear green or brown interiors (IIIa3, IVa5- 6), of which there are 82 examples, comprising 14% of the sample. These are varieties for a form known as Comaline d' Aleppo, which derives its name from association with the Italian export business with the city of Aleppo in Syria. This sort of bead was widely distributed among Indians of North America in the first half of the 19th century (Orchard, 1929, p. 87; Woodward, 1965, pp. 19- 20). Of the two primary varieties of Comaline d' Aleppo bead, the green- and brown-lined red forms occur chronologically earlier in the eastem United States and Canada than do those with white cores, a variety absent from the Paugvik sample but present in bead assemblages from other sites in southwestern Alaska. Previous students of beads from archaeological sites in this region have be- lieved that both forms were introduced into Alas- Site No. of beads "White "Green "Brown lined" lined" lined" centers Black «6 Paugvik 583 82 Crow Village 416 8 7 Akulivikchuk 537 2 4 Tikchik 407 2 18 Kijik 1,229 111 12 Kolmakov- skiyR. (2,431) Russian 75 levels U.S. levels 80 " Based on information from Oswalt and VanStone (1967), VanStone (1968, 1970), VanStone and Town- send (1970), and Oswalt (1980). * Possibly a misidentification of green-lined red. ka after extensive use elsewhere in North America but that the exact time of introduction cannot be determined (Oswalt & VanStone, 1967, p. 60; VanStone, 1968, p. 295; VanStone & Townsend, 1970, p. 97). Distribution of Comaline d' Aleppo beads at ex- cavated historic sites in southwestem Alaska is shown in Table 6. A large number of white-lined red beads were found at the Kijik site, which may have been oc- cupied somewhat later than the others. Also, at Kolmakovskiy Redoubt beads that are "red with black centers" occur in approximately equal num- bers in both Russian- and U.S.-period levels. Be- cause the clear centers of the green-lined red beads appear black unless held up to the light, the Kol- makovskiy beads may be the green-lined red va- riety. In any event, they cannot be said to be an exclusively Russian import. The white-lined red form also occurs in both the Russian- and U.S.- period levels but in much smaller numbers. South of Bristol Bay the picture is somewhat clearer. Bead assemblages from Chirikof Island south of Kodiak (Workman, 1 969), the Korov- inskii site on Atka Island in the Aleutians (Veltre, 1979), Reese Bay on Unalaska Island (Francis, 1988), and Nunakakhnak on Kodiak Island (Sha- piro, 1988) indicate that the green-lined red va- riety was imported into Alaska relatively early in the 19th century. The Chirikof Island site was abandoned by the time of the Alaska purchase of 1867, and thus all trade goods recovered there belong to the Russian period. The Chirikof col- lection includes 45 green-lined red beads and a single white-lined red example, which Workman (1969, pp. 200-212) believed had been imported 56 Part Four from the interior. The Korovinski collection con- tains 1 1 green-lined red beads but no white-lined red examples, and because the site was largely abandoned by 1872 the beads there can also be considered of Russian importation. Longhouses at the Reese Bay site were abandoned about 1806, and approximately 10% of the beads recovered from them were the green-lined form. At the Nu- nakakhnak site, apparently abandoned in the 1 880s (Knecht & Jordan, 1985, p. 21), 37% of the re- covered beads were the Comaline d' Aleppo form; those with a light green center were by far the most common. Thus, Workman (1969, p. 204) was ap- parently correct when he noted that the green-lined red form of the Comaline d' Aleppo is "a marker of Russian contact in this area." The absence of the white-lined red variety at Paugvik seems significant. Nevertheless, given the conflicting evidence at Chirikof Island, Korovin- ski, and Kolmakovskiy Redoubt, as well as the presence of both varieties at Kijik, Crow Village, and the Nushagak River sites, the precise chro- nological significance of the Comaline d' Aleppo beads in southwestern Alaska still wants clarifi- cation. Also of interest in the bead assemblage at Paug- vik are the large number of white beads that cannot be accommodated within the Kidds' classification system (Table 4). These beads are oyster white on the exterior with an opaque white core. Although a wide range of beads of various sizes and colors were recovered from the historic sites in south- westem Alaska, the aboriginal people were ap- parently partial to white beads, as this was the predominant color present in the assemblages from all the Nushagak River sites except Nushagak it- self, as well as those from Kijik, Crow Village, and Kolmakovskiy. Most of the beads are described simply as "white" without reference to exterior/ interior differences, but at Akulivikchuk (Van- Stone, 1 970, p. 84) and Kijik (VanStone & Town- send, 1970, p. 94), a large number exhibited a variation between exterior and interior color. If the white beads from all these sites were reex- amined, a considerable number probably would be found to have cores that differ slightly in color from the exteriors. A single flat native bead o^MgMXt (108) appears to represent an item manufactured in aboriginal style. With the exception of beads, few items of per- sonal adomment were recovered from the Paugvik site. There are two brass ^«^^r rings (109; Fig. 38h) and a circular band of soft binding iron that may have been wom as a bracelet (1 10). A bear's tooth has been drilled along one side at the edge and may have been wom by itself or as part of a necklace (111; Fig. 38f). An antler hair comb (112; Fig. 38j) is roughly rectangular in shape, with a series of short, closely spaced teeth at one end. The teeth appear to be too closely spaced for the implement to have been used for shredding grass or sinew. Smoking Complex A rectangular section of antler may be one side of an oval snuffbox {Wi). The edges of the fragment are omamented with parallel incised lines and in the center are a pair of incised circle designs with radiating spurs (Fig. 38i). For smoking, fungus ash was frequently mixed with tobacco to improve the taste and to make the tobacco last longer. The collection contains a single piece of birch fungus (1 14). These fungi, cut from trees in the interior, were traded to the coast by Athapaskan Indians (Nelson, 1983, p. 271). Toys There are two fragments of toy bows (115), one with a simple rounded nock and the other with a nock that is roughly diamond shaped. Both are ovoid in cross section, but the smaller is relatively wide and flat (Fig. 38k,l). Ceremonial Objects A single unfinished mask of wood (116) was re- covered from the Paugvik site. The shaping ap- pears to be virtually complete, but the nose, eyes, and mouth are barely indicated. On the reverse side the surface is nearly flat except for the area that would fit over the nose of the wearer (Fig. 40a). There are also three mask appendages (117) for the type of composite mask characteristic of southwestem Alaska. The first is fragmentary and roughly paddle shaped (Fig. 38e). The second is in the shape of a human hand with a hole through the palm (Fig. 38m). Pierced hand appendages are believed to have been associated with masks rep- resenting powerful tuunrat spirits that controlled the availability of animals on earth. The holes symbolize the willingness of the spirits to allow some animals to slip through their fingers, thus Collections 57 assuring their continued abundance on earth (Fitz- hugh & Kaplan, 1982, p. 202). Most Yupik carved hands have four fingers and no thumb, whereas this one has three fingers and a short thumb. The more problematic mask appendage is a piece of wood carved in the shape of a human leg (Fig. 45e). Two wooden human^^r/n^5 (1 18) are worked to a point at one end. On the larger, which is very poorly preserved, the head and shoulders are de- picted but the features have been obliterated (Fig. 40c). A similar figurine was recovered at Paugvik by Larsen (1950, Fig. 55a, 7). The smaller depicts only the head, but the incised features are clearly indicated. The marks of a rodent tooth tool are clearly visible on this figurine (Fig. 40b). Miscellaneous A bundle of grass wrapped with a sealskin thong may have served as a respirator ( II 9) for a person taking a sweat bath, although Nelson (1983, p. 288) indicates that these respirators were usually made of wood shavings (Fig. 40d). Protective Network Clothing Fragments of skin and commercial cloth are rel- atively scarce in the inventory of materials from the Paugvik site, and most of those recovered are too small or too poorly preserved for identification as to the type of apparel they represent. One ex- planation, of course, is that preservation in many parts of the site was poor, with frozen sections discontinuous except for those in Trench 1 and House 6, and it was only from frozen matrix that cloth and leather fragments were recovered. Cloth garments along with those of skin were probably of importance at the site; when the Korsakovskiy expedition visited Paugvik in 1818, European clothing was already among the trade objects most desired by the natives (VanStone, ed., 1988, pp. 28-29). There are six fragments of sealskin mukluk soles (120), all of which are quite small but include an area of crimping around the toe (Fig. 41c); no up- per sections were identified. Three fragments of cut sealskin, each found with numerous small, deteriorated pieces, may be gar- ment fragments (121). Two circular pieces of sealskin with a row of sewing holes around the edges are identified as patches (122), probably for parkas or boots but perhaps for boat covers (Fig. 41a). Large pieces of sea mammal intestine suggest parts of raincoats (123). Five buttons (124) were recovered. Three are four-hole buttons of wood that are obviously homemade (Fig. 4 1 f). The fourth is covered with brown wool fabric and the material of the button itself cannot be determined (Fig. 41g). Half of a plain brass button is a coin-shaped disc that once had an eye of the same material soldered to the back. Around the edges on the reverse is stamped "F Barnes & Co." (Fig. 41e). A similar complete button from the Nushagak site is stamped with the words "F. BARNES &. CO./LONDON" (VanStone, 1972, p. 64, PI. 13, 10). It has not been possible to locate this firm in lists of known button companies, but buttons of this type, with soldered eyes, were manufactured between 1812 and 1820 (Olsen, 1963, pp. 31-33). Evidently an attempt had been made to cut the Paugvik button into strips. The two fragments of factory-made shoes {\25) are too fragmentary to provide much information about the method of manufacture or to be of chro- nological significance. A single sole fragment ap- pears to have been sewn to the upper, the insole attached to it by a row of wooden pegs that run longitudinally along the center of the foot; one peg is still in place (Fig. 41b). In the United States, machine-made pegs were introduced about 1811, and a hand-operated pegging machine was pat- ented in 1829 (Anderson, 1968, pp. 58-59). The other shoe fragment is the outer section of a heel made of leather, which had been fastened to the lifts around the edges and in the center with heavy iron nails as much as 0.4 cm in diameter. The nail heads protrude on the outer surface (Fig. 4 Id). In addition to the skin clothing fragments, there are 48 sea mammal or caribou skin fragments with stitching holes (126) along one or more edges, 64 cut skin fragments ( 1 27) presumably associated with clothing, and 45 uncut skin fragments (128). Because of the fragility of the deteriorated mate- rial, these counts are approximate. There are also two fragments oi knotted sealskin line (129; Fig. 41h) and one fragment oi knotted baleen (130). Unfortunately, little can be inferred concerning the European clothing in use because of the small number and poor condition of those fabrics re- 58 Part Four covered. The fibers of the 1 3 cloth fragments (131), now brown or black in apparent color, are ex- tremely degraded but appear to be wool of plain (tabby) weave or twilled weave. Two fragments may have been originally fulled, and one has a seam with stitch holes along one edge. On another, stitch marks and circular impressions indicate that three buttons had once been sewn along one edge. On one fragment two paired sets of wefts are ev- ident, one now black and the other brown. One fragment is unusual in having two such sets of wefts, one of wool and the other spun with coarse animal hair of unknown origin. the Paugvik shards is very much in keeping with a date sometime before the sale of Alaska in 1 867. Two square-cut nails (133) were recovered, on one of which the head is missing. The complete nail is within the range of the 40d length. There are also two badly corroded screws (134) with rounded heads. A small fragment of muscovite mica (135) ap- proximately 2 cm X 1.5 cm may have been part of the covering for a window. Zagoskin (1967, p. 1 86) mentioned that mica was brought from Sitka to be used for window panes in the Russian-Amer- ican Company buildings at Nulato on the Yukon. The collection also contains one very small frag- ment of brick (136). Imported Building Materials The Paugvik natives had very little access to im- ported building materials. Some of the few pieces that were recovered may have been salvaged from driftwood, and others may be unrelated to occu- pation of the site. The 1 1 fragments ofwz>K/ow^/as5( 132) are clear and small (maximum dimension of the largest is 42 mm) and range in thickness from 1.0 to 2.2 mm, with a mean thickness of 1.51 mm; five of the 11 are between 1.1 and 1.65 mm. Although window glass was highly prized by the natives of southwestern Alaska at least as early as 1 842 (Za- goskin, 1967, p. 255), such glass was apparently available at Paugvik only in small quantities. The Hudson's Bay Company imported English win- dow glass into the Pacific Northwest from the time of its establishment there and after 1 840 can be expected to have been the source for flat glass found in Alaska at least until 1867. Before the mid- 19th century the major English production was of spun- blown crown glass, much of which was very thin (Roenke, 1978, pp. 5-6). Although large sheets of this glass show circular patterns of bubbles or im- perfections, these are almost never discernible in fragments as small as those reported here. Ac- cording to Roenke ( 1 978, p. 11 6), the modal thick- ness of sheet glass fragments found in archaeolog- ical sites in the Pacific Northwest that were oc- cupied before 1845 does not exceed 1.4 mm, be- comes thicker than the mean of the Paugvik shards only sometime after that date, and exceeds the thickness of the single thickest Paugvik fragment only after about 1870. Although variations in the thickness of such glass products are great enough to rule out the definitive dating of very small sam- ples by this measurement alone, the thickness of Unidentified Objects There are a large number of unidentified objects, most of them fragmentary. The more interesting of these are described according to the material of which they are made. Wood The following objects, complete or nearly so, were found. Eight are stakes (137) pointed at one end, ranging in length from 3 cm to 10 cm, two of them complete (Fig. 42a,c), and the rest are unidentified (138): a short, handlelike object recessed at one end, with a sharp shoulder (Fig. 42d); a flat, oval piece with a series of incised lines on one side (Fig. 42i); an oval shaft with a narrow blade slit at one end, grooves left by sinew lashing, and narrow, fringelike slits at the other end (Fig. 42b); an object wedge shaped at one end and rounded with an incised groove at the other (Fig. 42j); a piece of birch bark cut to an oval shape (Fig. 42e); and a small peg enlarged at one end (Fig. 42g). Antler, Ivory, Bone Those unidentified objects (139) illustrated include a partially exfoliated strip of antler, with incised eyes at one end as well as a number of other incised lines, possibly representing an animal or fish (Fig. 43c); a thin antler fragment with parallel incised lines on one side and a vertical series of drilled holes, probably part of the rim of something (Fig. 42f); a paddle-shaped object of antler with a pro- Collections 59 jection enlarged at the end and a line hole (Fig. 43a); a partially exfoliated antler fragment, pos- sibly a handle, with a knob at one end on which a mouth and eyes are incised (Fig. 42k); a forked handle-like object of antler (Fig. 45c; Dumond, 1981, PI. XVII, Cg); a possible spoon handle of antler, along the rear of which is a vertical row of circular depressions that once contained beads, with one bead still in place (Fig. 45h); two oval ivory tubes, on one of which are incised lines (Fig. 43d,e); two sea mammal ribs with elliptical per- forations in one end (Fig. 43f); and a caribou scap- ula with two round perforations in the blade (Fig. 43b). Miscellaneous Debris This section lists manufacturing detritus, items that appear to have been picked up and purposely brought to the site, and subsistence byproducts. Some of these items are confined to those collected in 1985 (see Table 2). Scraps of cut or broken material discarded in manufacture, not yet completely formed for use, or at times possibly only fortuitously present in- clude 107 fragments of5/a/e( 140), 31 chips of chert or quartzite (141), several pieces of pumice (142), and cut bone, ivory, and antler fragments (143, 144, 145). Metal scraps include 70 fragments of iron or steel (146) many of which appear to be cut from barrel hoops. There are also 2 1 fragments of cop- per-based alloys of brass or bronze (147). Five of these have been analyzed for content, confirming the industrial origin of four of them but suggesting that one, from Trench 1 (see Part 3) is possibly native copper (Harritt & Dumond, in prepara- tion). Although there were numerous scraps of baleen, these were not systematically collected, except for a piece cut into a leaf shape (148; Fig. 42h). Fragments of remains of Pleistocene woolly mammoth of Alaska {Mammuthus primigenius) attest to the proclivity of the Paugvik people for bringing home segments of tusk and tooth from mammoth remains that erode regularly from Pleistocene deposits around the mouth of the Nak- nek River and upper Bristol Bay. The latest ra- diocarbon designation associated with Alaskan mammoth remains is on the order of 1 3,500 years ago (see, for example, Guthrie, 1 990, p. 244), and no Alaskan mammoth remains have yet been found associated with contemporary human traces. But mammoth tusk or tooth fragments (149) were re- corded in eight excavation units at Paugvik, tusk laminae in five of them (in House 2, level A, the tusk fragment found was apparently used as a net weight), and cheek teeth or their sections in four of them. Remains of more contemporary fauna include a number of samples of //a/r (150-156), including that of canids (dog, wolf, or fox), beaver, muskrat, seal, caribou, bear, and humans. The extensive bone waste is set out separately in Appendix 1 , by the original excavation units. Samples of ash and soil were taken from six of the hearths excavated in 1985, in the hope that these samples would provide information regard- ing use of vegetal materials other than those rep- resented by scattered remains of fern rhizomes. The samples were dried and then floated in water, with preservation and subsequent drying of both heavy and light fractions. A summary of the rather disappointing results is given in Table 7. 60 Part Four Table 7. Flotation analysis of six Paugvik hearths. Hearth location Fraction HI HIA light heavy light la heavy light heavy m light heavy H6 (secondary hearth)'' H6 (main hearth) light heavy light heavy Content charcoal, roots pulverized mammal bone, bird bone, shell charcoal, rootlets, mica flecks; 1 probable fish ver- tebra fragment; grass; calcined bone fragments same as light, without vertebra; 1 white seed bead charcoal; twigs; bird bone, probably duck (Alci- / dae) small bone, pulverized large mammal bone; 1 piece green glass; salmonid vertebra; bird toes twigs; charcoal; salmonid vertebra; bird toe salmonid vertebrae; bird toe; calcined bird bone, mammal bone charcoal pulverized bird bone, mammal bone charcoal; unbumed wood; grass pulverized bird or rabbit bone; pulverized shell; mica flecks; 1 white seed bead " The small section of fire at the east edge of the house; see Figure 14. Collections 6 1 cr 11 21 3| 1\ I I I I I I I I I I Fig. 1 6. a, harpoon socketpiece; b, haipoon socketpiece; c, harpoon socketpiece; d, harpoon socketpiece; e, harpoon socketpiece;/ harpoon dart head; g, float mouthpiece; h, harpoon dart head; /, harpoon foreshaft;7, float mouthpiece; k, harpoon socketpiece fragment; /, harpoon socketpiece fragment; m. harpwon ice pick; n, lance blade sheath; o, unfinished arrowhead; p, arrowhead; q, unfinished arrowhead; r, bow fragment; s, wound plug; t, basal fragment of arrowhead; u, blunt arrowhead; v, blunt arrowhead (fmnh neg. no. A-1 10990). 62 Part Four Fig. 17. a, boat or meat hook; b, end blade; c. end blade; d, end blade; e, arrow shaft fragment;/ lurehook shank; g, gun side plate; h, lurehook shank; /, lurehook; j, bullet mold half; k, arrow shaft fragment; /, leister prong; m, pointed object; n, pointed object; o. pointed object; p, fish spear pKjint; q, net weight; r, mesh gauge; s, fish scaler (?); /, net weight; u, net float fragment (fmnh neg. no. A-1 10994). Collections 63 ''I''l''l''ll'l Fig. 18. a, kayak deck beam; b, net weight; c, sled shoe fragment; d, net weight; e, snowshoe crosspiece;/ sled shoe; g, sled shoe fragment; h, net weight; /, net weight (fmnh neg. no. A-1 10991). 64 Part Fovir Fio. 19. a, pelt stretcher; b, pelt stretcher (fmnh neg. no. A-1 10989). Collections 65 J 'II1I1I1I1 Fio. 20. a, sled upright; b, umiak rib or riser; c, wedge; d. wedge; e, wedge; / wedge; g, wedge (fmnh neg. no. A- 110986). 66 Part Four Fig. 21. a, sled runner; b, wedge; c, wedge; d, wedge; e, wedge;/ sled stanchion (?) (fmnh neg. no. A-1 10988). Collections 67 ;"i ii .1 li Fig. 22. a, maul; b, axe head; c. axe head; d, wedge; e, whetstone (fmnh neg. no. A-1 10986). 68 Part Four I I • llllllll Fig. 23. a, crooked knife handle; b, skin scraper blade blank; c, crooked knife blade; d, crooked knife blade; e, rodent incisor knife;/ adze blade; g, composite knife handle; h, composite knife handle; /, knife or engraver fragment; j, whetstone; k, whetstone; /, whetstone; m, whetstone; n, engraving tool; o, knife or engraver fragment; p, whetstone; q, whetstone; r, whetstone (fmnh neg. no. A-1 10992). Collections 69 Fig. 24. Whetstone (fmnh neg. no. A-1 10987). 70 Part Four Fig. 25. a, ulu; b, ulu; c, ulu; d, ulu; e, ulu;/ scraper or knife; g, awl; h, awl; /, scraper or knife;/ pick or mattock blade (fmnh neg. no. A-1 10984). Collections 7 1 Him^^ '■' ' v^' .''■"' ■'.'.' ' ■■.'■• '*• m nTrTiiiy : V -. • B s Fio. 26. fl, pick or mattock blade, b, compound vessel; c, shovel blade (fmnh neg. no. A-1 10982). 72 Part Four Fig. 27. a. snow beater, b. rake prong; c, unidentified ; ological Papers, no. 44. 1994. A reevaluation of the late prehistoric houses of the Naknek River region, southwestern Alas- ka. Arctic Anthropology, 32(2): 108-1 18. Efimov, a. v., ED. 1964. Atlas geograficheskikh ot- krytiy v Sibiri i v Severo-Zapadnoy Amerike XVII- XVIII vv. Nauka, Moscow, Russia. Elliott, H. W. 1875. A Report upon the Condition of Affairs in the Territory of Alaska. U.S. Treasury Department, Washington, D.C. . 1886. Our Arctic Province: Alaska and the Seal Islands. Charles Scribner's Sons, New York. Fedorova, S. G. 1973a. New data on Russian Geo- graphic and ethnographic investigations in Alaska (first half of the 1 9th century). Paper read at the 9th Inter- national Congress of Anthropological and Ethnologi- cal Sciences, Chicago, 1973. . 1 973b. The Russian Population in Alaska and Cahfomia: Late 18th Century- 1867. [Pierce, R. A., and A. S. Donnelly, transls., eds.] The Limestone Press, Kingston, Ontario. FiENUP-RiORDAN, A. 1990. Yup'ik warfare and the myth of the peaceful Eskimo, pp. 146-166. In Eskimo Es- says: Yup'k Lives and How We See Them. Rutgers University Press, New Brunswick, New Jersey. FiTZHUGH, W. W., AND S. A. Kaplan. 1982. Inua: Spirit World of the Bering Sea Eskimo. Smithsonian Institution Press, Washington, D.C. Ford, J. A. 1 959. Eskimo prehistory in the vicinity of Point Barrow, Alaska. Anthropological Papers of the American Museum of Natural History, 47(1). Franos, p., Jr. 1988. Early Russian trade beads in Alaska. Ornament, 12(1): 26-27, 82-85. FWS [U.S. Fish and Wildlife Service]. 1985. The Bris- tol Bay Regional Management Plan and Final Envi- ronmental Impact Statement. U.S. Department of the Interior, Washington, D.C. Gibson, J. R. 1 976. Imperial Russia in Frontier Amer- ica: The Changing Geography of Supply of Russian America, 1784-1876. Oxford University Press, Lon- don. GoDDEN, G. A. 1 964. Encyclof)edia of British Pottery and Porcelain Marks. Crown Publishers, New York. Guthrie, R. D. 1990. Frozen Fauna of the Mammoth Steppe: The Story of Blue Babe. University of Chicago Press, Chicago. Harrttt, R. K. 1988. The Late Prehistory of Brooks River, Alaska. University of Oregon Anthropological Papers, no. 38. Harrftt, R. K., AND D. E. DuMOND. n.d. Analysis of copper artifacts and raw materials from southern Alas- ka. Manuscript in preparation. Hemming, J. E. 1971. The Distribution and Movement Patterns of Caribou in Alaska. Alaska Department of Fish and Game, Game Technical Bulletin, no. 1. Hrdlicka, A. 1943. Alaska Diary, 1926-1931. Jaques Cattell, Lancaster, Pennsylvania. . 1 944. Catalog of Human Crania in the United States National Museum Collections: Non-Eskimo People of the Northwest Coast, Alaska, and Siberia. Proceedings of the United States National Museum, 94: 1-175. Jewett, L. 1878. The Ceramic Art of Great Britain from Prehistoric Times Down to the Present Day: Being a History of the Ancient and Modem Pottery and Porcelain Works of the Kingdom and the Products of Every Class. 2 vols. Virtue and Co., London. Jordan, R. A., and R. A. Knecht. 1988. Archaeo- logical research on western Kodiak Island, Alaska: The development of Koniag culture, pp. 225-306. /nShaw, R. D., R. K. Harritt, and D. E. Dumond, eds.. The Late Prehistoric Development of Alaska's Native Peo- ple. Alaska Anthropological Association Monograph Series, no. 4. Kaplan, S. A., and K. T. Barsness. 1986. Raven's Journey: The World of Alaska's Native People. The University Museum, University of Pennsylvania, Philadelphia. Khlebnikov, K. T. 1976. Colonial Russian America: Kyrill T. Khlebnikov's Reports, 1817-1832. [Dmy- tryshyn, B., and E. A. P. Crownhart-Vaughan, transls. and eds.] Oregon Historical Society, Portland. Kidd, K. E., and M. a. Kidd. 1970. A classification system of glass beads for the use of field archaeologists, pp. 45-89. In Canadian Historical Sites, Occasional Papers in Archaeology and History, no. 1 . Department of Indian Affairs and Northern Development, Ottawa. Knecht, R. A., and R. H. Jordan. 1985. Nunakakh- nak: An historic Koniag village in Karluk, Kodiak Island, Alaska. Arctic Anthropology, 22(2): 1 7-35. KowTA, M. 1 963. Old Togiak in prehistory. Ph.D. diss., Department of Anthropology, University of Califor- nia, Los Angeles. University Microfilms, Ann Arbor, Michigan. Krauss, M. 1982. Native Peoples and Languages of Alaska (map). Alaska Native Language Center, Uni- versity of Alaska, Fairbanks. Krech, S., III. 1989. A Victorian Earl in the Arctic: The Travels and Collections of the Fifth Earl of Lons- dale 1888-89. University of Washington Press, Seat- tle. Langsdorff, G. H. von. 1814. Voyages and Travels in Various Parts of the World, During the Years 1 803, 1804, 1805, and 1807. Part II. The Voyage to the Aleutian Islands and North-West Coast of America. Henry Colbum, London. Lantis, M. 1946. The social culture of the Nunivak Eskimo. Transactions of the American Philosophical Society, n.s., 35(3): 153-323. Larsen, H. 1950. Archaeological Investigations in southwestern Alaska. American Antiquity, 15(3): 177- 186. Lyapunova, R. G., and S. G. Fedorova, eds. 1979. Russkaya Amerika v neopublikovannykh zapiskakh K. T. Khlebnikova. Nauka, Leningrad. Mason, O. T. 1902. Aboriginal American harpoons: A study in ethnic distribution and invention, pp. 1 89- 304. In Report of the United States National Museum for 1900. Washington, D.C. McGhee, R. 1974. Beluga Hunters: An Archaeological Reconstruction of the History and Culture of the Mac- Literature Cited 103 kenzie Delta Kittegariyumiut. Newfoundland Social and Economic Studies, no. 13. Morrow, J. E. 1 980. The Freshwater Fishes of Alaska. Northwest Publishing Co., Anchorage. MosER, J. F. 1902. The Salmon and Salmon Fisheries of Alaska. Report of the Alaskan Salmon Investiga- tions of the United States Fish Commission Steamer Albatross in 1900 and 1901. U.S. Fish Commission, Washington, D.C. Nelson, E.W. 1983. The Eskimo About Bering Strait, pp. 19-518. In Eighteenth Annual Report of the Bu- reau of American Ethnology, Pt. 1 . U.S. Fish Com- mission, Washington, D.C. [Reprint of 1899 edition.] Olsen, S. J. 1963. Dating early American buttons by their form. American Antiquity, 28(4): 551-554. Orchard, W. C. 1929. Beads and beadwork of the American Indians. Museum of the American Indian, Heye Foundation, contribution no. 11. Oswalt, W. H. 1 952a. The archaeology of Hooper Bay village, Alaska. Anthropological Papers of the Uni- versity of Alaska, 1(1): 47-91. . 1952b. The saucer-shaped Eskimo lamp. An- thropological Papers of the University of Alaska, 1(2): 15-33. 1955. Alaskan pottery: A classification and historic reconstruction. American Antiquity, 21(1): 32- 43. 1980. Kolmakovskiy Redoubt: The Ethnoar- chaeology of a Russian Fort in Alaska. University of California, Institute of Archaeology, Monumenta 7 Archaeologica. Institute of Archaeology, Los Angeles. Oswalt, W. H., AND J. W. VanStone. 1967. The Eth- noarchaeology of Crow Village, Alaska. Bureau of American Ethnology, bulletin no. 199. Peterson, H. L. 1965. American Indian Tomahawks. Museum of the American Indian, Heye Foundation, contribution no. 19. Petroff, I. 1881. Population and Resources of Alaska. 46th U.S. Congress, 3rd Session, House Executive Documents, no. 40, series 1968. Government Printing Office, Washington, D.C. . 1884. Report on the Population, Industries, and Resources of Alaska. Department of the Interior, Census Office, Washington, D.C. Porter, R. P., comp. 1893. Report on the Population and Resources of Alaska at the 11th Census: 1890. Department of the Interior, Census Office, Washing- ton, D.C. RAC [Russian-American Company]. 1818-1865. Re- cords: Communications Sent and Received. Manu- scripts in the National Archives (Record Group 1 1). Washington, D.C. RoENKE, K. G. 1 978. Flat glass: Its use as a dating tool for nineteenth century archaeological sites in the Pa- cific Northwest and elsewhere. Northwest Anthropo- logical Research Notes, 12(2): part 2. Russell, C. P. 1967. Firearms, Traps, & Tools of the Mountain Men. Alfred A. Knopf, New York. Saraitan, W. L. 1970. Russian-American Company Employee Policies and Practices, 1799-1867. Ph.D. diss., Dept. of History, University of California, Los Angeles. University Microfilms, Ann Arbor, Michi- gan. ScHAAF, J. 1988. Report on the Denton Collection. Verdant Cave, Aialik Bay, Kenai Fjords National Park. Alaska Regional Office, National Park Service, An- chorage. Shapiro, E. G. 1988. Trade beads excavated from a European/Konyak site on Kodiak Island, Alaska. The Bead Forum, 13: 7-12. Shinkwin, A., AND E. Andrews. 1979. Archaeological Excavations at the Erskin House, Kodiak, Alaska— 1978. Manuscript on file, University of Alaska, Fair- banks. Spurr, J. E. 1 900. A reconnaissance in southwestern Alaska in 1 898. In Annual Report of the United States Geological Survey, part 7, 20: 3 1-264. U.S. Geological Survey, Washington, D.C. Staley, D. p. 1990. The Archaeological Analysis of House 15 at Chagvan Bay Beach Site. M.A. thesis, Washington State University, Pullman. SussMAN, L. 1979. Spode/Copeland transfer-printed patterns found at 20 Hudson's Bay Company sites. Canadian Historic Sites, Occasional papers in Ar- chaeology and History, 22. Department of Indian Af- fairs and Northern Development, Ottawa. U A [University of Alaska]. 1974. The Bristol Bay En- vironment. Report for the U.S. Army Corps of En- gineers, Alaska District. Arctic Environmental and Data Center, and Institute of Social, Economic, Gov- ernment Research, University of Alaska, Fairbanks. . 1981. Wild Edible and Poisonous Plants of Alaska. Cooperative Extension Service, Publication, no. 28. U.S. Census. 1900. Census enumeration sheets, Alas- ka, Bristol Bay enumeration district. Manuscripts in the National Archives, Washington, D.C. . 1910. Census enumeration sheets, Alaska, Nushagak enumeration district. Manuscripts in the National Archives, Washington, D.C. VanStone, J. W. 1 967. Eskimos of the Nushagak Riv- er: An Ethnographic History. University of Washing- ton Press, Seattle. 1968. Tikchik Village: A Nineteenth Century Riverine Community in Southwestern Alaska. Field- iana: Anthropology, 56(3): 2 1 5-368. 1 970. Akulivikchuk: A Nineteenth Century Es- kimo Village on the Nushagak River, Alaska. Fieldi- ana: Anthropology, 60. 1972. Nushagak: An Historic Trading Center in Southwestern Alaska. Fieldiana: Anthropology, 62. VanStone, J. W., ed. 1973. V. S. Khromchenko's Coastal Explorations in Southwestern Alaska, 1822. Introduction by J. W. VanStone. [Kraus, D. H., transl.] Fieldiana: Anthropology, 64. . 1988. Russian Exploration in Southwest Alas- ka: The Travel Journals of Petr Korsakovskiy (1818) and Ivan Ya. Vasilev (1829). Edited and with an In- troduction by J. W. VanStone. [Kraus, D. H., transl.]. The Rasmuson Library Historical Translation Series, 4. University of Alaska Press, Fairbanks. VanStone, J. W., and J. B. Townsend. 1970. Kijik: 104 Literature Cited An Historic Tanaina Indian Settlement. Fieldiana: Anthropology, 59. Veltre, D. W. 1979. Korovinski: The Ethnohistorical Archaeology of an Aleut and Russian Settlement on Atka, Alaska. Ph.D. diss., Department of Anthropol- ogy, University of Connecticut. University Micro- films, Ann Arbor, Michigan. Williams, P., and M. C. Weber. 1986. Staffordshire Romantic Transfer Patterns II: Cup Plates and Early Victorian China. Fountain House East, Jefferstown, Kentucky. Woodward, A. 1965. Indian Trade Goods. Oregon Historical Society, no. 2. Portland. Workman, W. B. 1969. Contributions to the Prehis- tory of Chirikov Island, Southwestern Alaska. M.A. thesis. Department of Anthropology, University of Wisconsin, Madison. Wrangell, F. p. von. 1970. The inhabitants of the northwest coast of America. [VanStone, J. W., transl. and ed.] Arctic Anthropology, 6(2): 5-20. . 1980. Russian America: Statistical and Eth- nographic Information. [Sadouski, M., transl.] Mate- rials for the Study of Alaska History, 15. Limestone Press, Kingston, Ontario. [Original edition 1839.] Zagoskin, L. a. 1967. Lieutenant Zagoskin's Travels in Russian America, 1842-1844. [Michael, H. N., ed.] Arctic Institute of North America. Translations from Russian Sources, no. 7. University of Toronto Press, Toronto. ZiMMERLY, D. W. 1979. Hooper Bay Kayak Construc- tion. National Museum of Man, Mercury Series, Ca- nadian Ethnology Service, paper no. 53. National Mu- seums of Canada, Ottawa. Literature Cited 105 Appendix: Fauna from Paugvik, Alaska David S. Reese'' Paugvik village is located at the mouth of the Naknek River in southwestern Alaska. It was occupied in the 1 9th century until about a.d. 1 870. Fauna was recovered by hand collection and dry sieving through a 1 -cm mesh. The appended catalogue provides details on the fauna recovered from Houses 1-6 and Trenches 1 and 2. House 5 and Trench 1 produced very small samples. Caribou was found in all contexts, with 28 in- dividuals. Most came from House 6(11 individ- uals) and House 2 (six individuals). There are two probable adult moose bones, from House 6 and Trench 4. Canids were found in all contexts except Houses 1 and 5, with most from House 2 (six individuals). There are a total of 27 individuals (15 fox, seven dog/wolf, and six fox or dog/wolf)- Beaver remains were found in House 1 , House 2, House 3 (two individuals). House 6 (four in- dividuals), and Trench 1 , representing nine indi- viduals. Bear remains were found in House 2, House 3, House 6 (two individuals), and Trench 2, repre- senting five individuals. Otter bones were present in House 3, House 6, and Trench 2, representing three individuals. Muskrat bones were found in two areas (House 1 , Trench 2), representing two individuals. Musk- rat hair was found in Trench 1 . Whale bones were found in all deposits, with four individuals from House 6 and one or two from Trench 1, for a total of 10 or 11 individuals. Seal remains were present in House 1, House 2, House 3 (two individuals). House 6 (three indi- viduals), and Trench 4, representing eight indi- viduals. " Department of Anthropology, Field Museum of Nat- ural History, Chicago. There are 210 birds bones present from 30 in- dividuals. Most of these came from House 6(11 bones, 1 1 individuals). House 2 (six individuals), and Trench 2 (56 individuals). There were no bird bones in House 5, and only one bone was found in both House 4 and Trench 1 . Fish bones were found in all contexts except Houses 4 and 5. There are probably no more than 12-14 individuals represented. Salmon is the most likely form present. There are 55 shells: 45 Macoma, five Mytilus, four whelks, and one Clinocardium. Most of the shells came from Houses 2 and 3. Unmodified mammoth tusk pieces were found in Houses 2 and 6. All of the fauna might have been eaten, but this is particularly likely for the caribou and rarer moose as well as the whale, seal, birds, shells, and fish. Over 30 of the terrestrial mammal individuals ( 1 5 fox, nine beavers, five bears, three otters) may have been killed for their fur rather than for meat. These same species were represented in the smaller terrestrial vertebrate sample from the ear- lier work at Paugvik: six caribou individuals, four foxes, two or three other canids, two beavers, one bear, one otter, four whales, five seals, and 27 birds (mainly ducks and geese). The aquatic faunal com- ponent was somewhat larger than that of the pres- ent sample, with 37+ fish individuals and 248 shell valves (Dumond, 1981, p. 177). The modem fauna in the Paugvik area has pre- viously been described (Dumond, 1981, pp. 10- 11). Catalogue of the Fauna from Paugvik Village Abbreviations used: F = fused, JF = just fused; L == left; MNI = minimum number of individuals; mt = metatarsus; R = right; UF = unfused. Alces = moose; Canis familiaris/C. lupus — dog/wolf; 106 Appendix Castor canadensis = beaver; cetacean = whale (probably beluga); Lutra canadensis = otter; On- datra = muskrat; Phoca vitulina = harbor seal (rib- bon seal could also be present); Rangifer tarandus = caribou; Ursus arctos = brown bear; Vulpusfulva = fox; Clinocardium nuttalli = Nuttall's cockle; Macoma balthica = Baltic macoma; Mytilus edulis = mussel. House 1 Rangifer: 7 antler fragments (7 worked), skull frag- ment, scapula fragment (cut down tuber spinae), 2 proximal femora (2 F, 1 L, 1 R [butchered through trochanter majus]), 3 distal femora (3 F, 2 L, 1 R, 2 MNI), proximal tibia (F, R), distal tibia (F, R), 2 pelvis acetabulum fragments (IF, L), 2 astragali (1 L, 1 R, 1 badly worn, 2 MNI), calcaneus (F, R), naviculocuboid, 8 carpi/tarsi/sesmoids, proximal metacarpus, proximal metatarsus, proximal me- tapodial fragment, distal metacarpus (F), 3 pha- langes 1 (3 F), 2 phalanges 2 (2 F), 2 phalanges 3, 4 vertebrae, 1 4 ribs (2 MNI). Castor. 2 mandibles (1 L, 1 R), incisors, ulna. Ondatra: ulna. Cetacean: 4 vertebrae, 3 vertebral centra epiph- yses, phalanx, rib. Phoca: pelvis. Bird: 6 bones (1 MNI). Fish: 1 7 vertebrae, no toothed elements Shells (MNI): 5 Macoma, 1 Mytilus MNI), 10 proximal tibia fragments (3 L [ 2 UF, 1 F], 2 R [2 UF], 3 fragments [1 UF]), 6 distal tibiae (3 L [1 UF, 2 F], 3 R [3 F], 4 MNI), 2 patellae (2 MNI), 6 astragali (3 L, 3 R, 2 badly worn, probably 4 MNI), 3 calcaneus fragments (1 L [F], 1 R [bro- ken], 2 MNI), 2 naviculocuboids (L, R), 8 carpals/ tarsals, 5 distal metapodials (5 F, no proximal ends), 5 phalanges 1 (5 F), 1 phalanx 3, 5 vertebral fragments, 32 ribs, 35 fragments (6 MNI). Ursus: scapula (F, L, very large), radius shaft (UF proximal and distal, L), radius shaft fragment, distal radius/ulna (F, R, proximal radius probably UF), sternum, 2 tibia shafts (2 UF, L, R), 4 carpi/ tarsi, 4 metapodials (2 UF [ 1 mt IV]), 2 vertebrae. Vulpes: various bones (3 MNI by mandible [1 subadult, 2 adult]). Canis: various bones (2 MNI by femur, tibia). Canis/Vulpes: hair (1 sample). Castor: scapula, hair (2 samples). Cetacean: 4 teeth, scapula (R), humerus, ?pelvis fragment, 24 metapodials/phalanges, 22 vertebral centra, 22 vertebral epiphyses, 1 5 ribs. Phoca: 2 distal humeri (?UF), 2 pelvis, femur (UF distal), metapodial. Bird: 50 bones and 1 feather (6 MNI, 4 of one species, and one each of two other forms). Fish: 23 vertebrae, no toothed elements. Shells (MNI): 18 Macoma, 1 Mytilus, 1 Clino- cardium, 2 whelks (1 fragment). Mammoth tusk fragment. House 3 House 2 Rangifer: 2 1 antler fragments (2 1 worked), 2 skull fragments (1 temporal, 1 occipital), 2 mandibles (2 adult, 2 R, 2 MNI), 14 isolated premolars/mo- lars, atlas, 8 scapulae (8 F, 6 R, 2 L, 6 MNI), 3 proximal humerus fragments (2 UF, L, R), 8 distal humeri (8 F, 6 R, 2 L [ 1 butchered]), 4 proximal radii (4 F, 2 L [1 worn], 2 R), 3 distal radii (2 R [1 UF epiphysis, 1 F], 1 L [UF epiphysis]), 2 distal radius shafts, 5 proximal ulnae (3 R ([1 UF, 2 broken], 2 L [1 butchered, proximal]), 3 pelvis fragments (1 acetabulum, 1 MNI), sacrum frag- ment, 5 proximal femora ( 1 UF head, 1 UF tro- chanter majus, 1 L F, 1 R F, 1 F head; 3 MNI), 3 distal femur fragments (1 UF, 1 F, 1 butchered, 2 Rangifer: 8 antler fragments (8 worked), atlas, scapula fragment, 2 ulnae (F, R, butchered down shaft; broken, L; 2 MNI), distal radius/ulna (F, L, worn), ?ulna shaft, proximal femur (?UF, L, worn), distal tibia (F, L), 4 carpi/tarsi, phalanx 1 (F, worn), 5 vertebrae fragments (1 caudal), 12 ribs (2 MNI). Ursus: palate fragment, 4 metapodials, vertebra, claw sheath, cf Ursus hair (1 sample). Vulpes: several bones. Canis: several bones. Castor: several bones (2 MNI by humerus). Lutra: several bones. Cetacean: skull fragment, jaw fragment, atlas, 1 metapodials/phalanges, vertebra centra, 3 ver- tebral epiphyses, 3 other bones. Phoca: several bones (2 MNI by pelvis, ulna). Bird: 1 2 bones (4 MNI by humerus, 2 species). Appendix 107 Fish: 57 vertebrae, 1 toothed element (2 MNI). Shells (MNI): 9 Macoma. 2 Mytilus. 1 whelk (fragment). House 4 (tested only) Rangifer. 2 antler fragments (2 worked), atlas (adult), distal radius (F), 2 ulnae (1 L, 1 R, 1 F, 1 broken adult; F has butchered radius shaft), prox- imal tibia (UF, opened for marrow), distal tibia (F, fresh, opened for marrow), carpus/tarsus, pha- lanx 1 (F, worn), 4 vertebral fragments (2 UF cen- tra), 1 1 ribs. Canis/Vulpes: mandible (adult), scapula (no gle- noid, young), humerus (proximal UF, distal F), radius and ulna (articulate, but separate), distal tibia (F), 5 vertebrae (2 MNI). Vulpes: mandible (adult), humerus (UF proxi- mal), radius and ulna (articulating). Cetacean: skull fragment, jaw fragment, meta- podial/phalanx, vertebra centra, vertebra epiph- ysis. Bird: 1 bone. Shells: 1 Macoma. House 5 (tested only, N17/E110) Rangifer. 2 antler fragments (2 worked), distal fe- mur (UF), vertebra (UF), rib (butchered). Cetacean: jaw (no teeth). Shells (MNI): 1 Macoma. House 6 (excavated as Trench 2 and Trench 1, E20-26) Rangifer. 20 antler fragments (20 worked), 8 skull fragments, 2 palate fragments ( 1 L, 1 R, 2 MNI), 4 atlantes (1 butchered, 4 MNI), 3 axes (1 butch- ered through and behind articular surface, 1 worn), 19 mandible fragments (all adult, 7 MNI L, 6 MNI R), 23 isolated premolars/molars, 1 8 scapula frag- ments (7 L [7 F], 7 R [1 smaller and broken, 6 F], 8 MNI), 16 proximal humeri (11 R [4 UF, 4 F], 5 L [3 UF, 2 F], 8 MNI), 22 distal humeri (12 R [1 UF, 10 F], 10 L [1 UF, 9 F]), humerus shaft (young, possibly UF, R), 4 proximal radii (3 L [3 F], 1 R [F]), 6 ulnae (5 L [3 UF, 2 broken], 1 R [UF]), 1 1 distal radii (7 L [5 UF, 2 F], 2 R [1 UF, 1 F]), distal radius shaft fragment, 39 pelvis frag- ments with 1 8 acetabulum fragments ( 1 L [ 1 UF, 6 F, 3 broken], 8 R [2 UF, 5 F, 1 broken], 9 MNI), 1 1 sacrum fragments (3 + MNI), 6 proximal fem- ora (4 R [2 UF, 2 F], 2 L [1 JF, 1 F]), proximal femur epiphysis (head, R), 6 distal femora (3 UF, 2 MNI; 3 F, 2 MNI), 3 distal femoral epiphyses (2 MNI), 1 small femur (all UF), proximal tibia (1 1 R [5 UF, 4 F], 7 L [3 UF, 1 JF, 2 F, 10 MNI], 13 distal tibia fragments (8 L [1 UF, 7 F], 5 R [3 UF, 2 F], also 1 shaft and 2 epiphyses, 7 MNI), 2 patellae, 9 astragali (6 L, 3 R, 6 MNI), 8 calcani (9 fragments, 5 L [3 UF, 2 F], 3 R [3 UF]), 4 naviculare (3 MNI), 24 carpi/tarsi, 5 proximal me- tacarpi (3 MNI), 5 proximal metatarsi (2 very large, 4 or 5 MNI), 5 metapodial fragments, 14 distal metapodials (13 F [1 mt], 1 broken mt), 6 pha- langes 1 (6 F), 4 phalanges 2 (4 F), 102 vertebral centra and 49 fragments, 165 ribs (11 MNI). ^Rangifer. phalanx 1 (UF), burnt black. lAlces: proximal radius (very large, F, R). Ursus: scapula (F, R), proximal humerus (F), distal humerus epiphysis (R), 2 radii (2 all UF), proximal ulna (F, R), proximal tibia epiphysis (L), tibia (F, R), 2 astragali ( 1 L, 1 R), calcaneus (UF, L), 3 carpi/tarsi, 6 metapodials (5 F), 1 phalanx 1 , 1 phalanx 2, 2 phalanx 3, 3 phalanx 3 sheaths, 19 vertebrae (2 MNI), cf. Ursus hair (9 samples). Canis: various bones, adult (3 MNI tibia; 2 MNI mandible, atlas, axis), 1 upper shaft of a femur has an incised slit (3 MNI). Vulpes: various bones (9 MNI mandibles [2 sub- adult, 7 adult], 3 MNI atlas, femur; 2 MNI axis, tibia) (9 MNI). Canis/Vulpes: hair (8 samples). Castor, various bones (4 MNI by humerus, 2 MNI by mandible and pelvis). One individual is very large based on a humerus and femur (4 MNI). Lutra: scapula, femur (UP). Cetacean: 6 teeth, 6 jaw fragments (1 looks burnt, 4 MNI), 4 scapula fragments (1 glenoid butchered, 3 MNI), femur head, 25 metapodials/phalanges, 44 vertebrae, 7 vertebral spine fragments, 67 ver- tebral epiphyses, 29 ribs (4 MNI). Phoca: 8 humeri and 3 epiphyses (5 MNI), 3 ulnae (3 MNI), 2 pelves, 2 femora (2 MNI), 20 metapodials/phalanges, 1 other bone (3 MNI), hair (2 samples). Bird: 117 bones and 13 feathers (11 MNI by humerus). Fish: 43 vertebrae, 7 toothed elements (3 or 4 MNI). Shells (MNI): 8 Macoma (1 burnt), 1 whelk (fragment). Mammoth tusk fragment (from Trench 2 back, N10/E20). 108 Appendix Trench 1 (E27-36); mixed midden Rangifer. 5 antler fragments (4 worked), 3 molar fragments, 3 proximal humerus (3 F, 2 L, 1 R, 2 MNI), 2 distal humerus (2 F, 1 L, 1 R, 1 MNI), proximal radius (F, R) attaches to proximal ulna (broken), distal radius/ulna (F, R), pelvis frag- ment, proximal femur (maybe UF, R, worn), fe- mur head epiphysis, proximal tibia (young adult/ adult, R, broken proximal), astragalus (R), 2 cal- caneus (1 L [F], 1 R [butchered through distal end]), naviculocuboid, 7 carpus/tarsus, 2 proximal metatarsus ( 1 butchered), distal metapodial (F, very worn), phalanx 3 (very large), 10 vertebrae [1 UF epiphysis], 18 ribs (2 MNI), hair (1 sample). Canis: palate fragment, 2 mandibles (1 MNI), distal humerus (F). Canis/Vulpes: hair (2 samples). Castor: 2 pelvis (1 L, 1 R, 1 MNI). Ondatra: hair(l sample). Cetacean: jaw and 5 jaw fragments (no teeth, 1 or 2 MNI), humerus head, large shaft (UF), me- tapodial/phalanx, 5 vertebral epiphyses, 3 small vertebrae, 2 ribs (1 or 2 MNI). Bird: 1 bone. Fish: 23 vertebrae, no toothed elements (1 or 2 MNI). Shells (MNI): 2 Macoma. Trench 4 (midden near the entrances of Houses 2 and 3) L), 2 proximal radii (2 F, 1 R, 1 L [butchered down center of shaft]), 3 distal radii (1 UF, 2 F [1 butch- ered toward distal], 3 L, 3 MNI), 9 pelvis frag- ments (2 L, 2 R, 1 butchered fragment, 2 MNI), 4 proximal femora (2 L [1 UF trochanter majus, 1 F], 2 R [F], 3 MNI), 2 distal femora (1 UF epiphysis, 1 F (distal, worn), 2 MNI), 2 distal tibiae (1 R [UF], 1 L [F], 2 MNI), astragalus (R), cal- caneus (UF, R), naviculocuboid, phalanx 1 (F), 1 3 vertebrae, 36 ribs (3 MNI). lAlces: posterior mandible (condyle process). Ursus: axis, 2 carpi/tarsi. Canis/Vulpes: various bones, all adult (4 MNI by mandibles, 2 MNI by scapula and pelvis), 1 burnt mandible, hair ( 1 sample) (4 MNI). Vulpes (adult): skull, 2 mandibles, atlas, axis, pelvis, sacrum, 1 astragalus, 1 calcaneus, most ver- tebrae and ribs; missing other limbs (Animal 1, Feature 14, N24/E65). Lutra: complete skeleton except for 1 astragalus and 1 calcaneus (second skeleton from Feature 14, N24/E65). Ondatra: mandible. Cetacean: tooth, jaw lacking teeth, scapula, fe- mur, 3 metapodials/phalanges, limb bone, 5 ver- tebrae, 3 vertebral centra epiphyses. Phoca: 2 humerus, 2 humeral epiphyses (UF), ulna (UF, L), pelvis, metapodial. Bird: 32 bones (6 MNI; 2 each of 2 species). Fish: largest sample 41 vertebrae (4 MNI by skull bones). Shells: 1 Macoma, 1 Mytilns. Rangifer: 3 antler fragments (3 worked), 2 man- dible fragments (2 L, 2 MNI), atlas, scapula (F, '0 Appendix 109 A Selected Listing of Other fleidiana: Anthropology 1 itles Available I Ethnographic Collect! - Sakhalin iropolog:^ ■' o. 8, l')88. 67 pages, 3;) iilus lidii Trade Ornaments in the Colle».iicii)a oi i leid Museum ui Fieldiana: Anthropology, n.s., no. 13, 1989. 40 pages, 32 illu: Material Culture of the Blackfoot (Blood) Indians of Southern Alberta, n.s., no. 19, 1992. 80 pages, 53 illus. . ; ; , Publication 13SJ PublicatioD 1404, $il.00 Publication 1439, $19.00 Material Cultureof the ChilcotinAthanaskansofWest Central Bri.'ish v'"nlnmhi:i Rv l.ir Fieldiana: Anthropology '■'■ pages, 25 ilhn Fui)UCJJK>!l 14 »(^-, ^iZ.Ut) less and Warfare Among the Kayenta Anasazi of the Thirteenth Century A.D. By Jonathan Haas and Winifred Creamer. Fit'l/iiono- 4>ith>-nnnJn\y\: n.s no. 71 iQQ'^ ?M ngp'^'i 74 illus. Publication 1450, $32.00 oUection of Copper Inuit Material Culture. By James W . VanStone. Fieldian c '^94. 71 pages, 44 illus. Publication ir^ -:' 7 00 chaeological Research at Tumatumani, Juii, Peru. By Charles Stanish and Lee Steadman, . : v .:ontribution by Matthew T. Seddon. Fieldiana: Anthropology, n.s., no. 23, 1994. 1 1 1 pages, 17o p u-, Publication 1457, $23.00 rder by publication number and/or ask for a free copy of our price lis inois residents add current destination tax. All foreign orders are payable m i_.o. <.il n-rt/cii^^^RS v^jiavwi 1 any U.S. bank or the U.S. subsidiary of any foreign bank. Prices and terms subject to change without Address all request' '^ FIELD MUSEUM OF NATURAL Hie Library— Publications Division Roosevelt Road at Lake Shore Dri^ c Chicago, Illinois 60605-2498, ' seum of Natural History CENTRAL GRCULATION BOOKSTACKS The person charging this material is re- sponsible for its renewal or its return to the library from which it was borrowed on or before the Latest Date stamped below. You may be charged a minimum fee of $75.00 for each lost book. Ifcsft/ mvHlotten, and W H ii»r t ltilnfl mi books oro rooaom for disciplinary actloa and may r w o i t in dJindiioi froa tlM Unlvorslty. TO RENEW CALL TELEPHONE CENTER, 33> . » 400 UNIVERSITY Of ILLINOIS LIBRARY AT URBANA-CHAAAPAION FEB 1 9 1997 When renewing 1^ phone, write new due date bdow previous due date. Ll^ UNIVERSITY OF ILLINOIS-URBANA 3 0112 028183785