4 Zi Yi a aes ELLE LEE a ee > LL a SAY WY . ON SAAN Wy NY SY y Li Lie Zy LE yyy LLL, Z Li Z LL: Vo, Z Die CEE Li ee GEL. Z Z ti Z tyyjyy RE EEE te tL LIB tj Yj GI: : Yip; yy Ze tj) FIO Se PD ED ey I PIT EET IITA AIT L TTT PRET DET aT Le RN Pag ee Ry i ‘i : Nu “ og SN I a i eo, coe ae Pe Se mn ote ee) Dh) OU aes Hyak eat) i, oa .aay Fi ie . ; | i. i Tig : Pale ; te \. ih : na 4 4 “t , “al t ~ ae tea) f, i ap) bby ie Py j iG Ly ( ¢ wit : he ed Van J wee ¥) - r - “& BUREAU OF AMERICAN ETHNOLOGY BULEEETIN 192° PEATE 1 ae ¢ yaar. “b tnt Tre. a eet Sites in the Yakutat Bay area. (For explanation, see p. 218.) SMITHSONIAN INSTITUTION BUREAU OF AMERICAN ETHNOLOGY BULLETIN 192 ARCHEOLOGY OF THE YAKUTAT BAY AREA, ALASKA By FREDERICA DE LAGUNA, FRANCIS A. RIDDELL, DONALD F. McGEEIN, KENNETH S. LANE, and J. ARTHUR FREED, with a chapter by CAROLYN OSBORNE U.S. GOVERNMENT PRINTING OFFICE WASHINGTON : 1964 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C., 20402 - Price $3.25 (Cloth) LETTER OF TRANSMITTAL SMITHSONIAN InsTITUTION, Bureau or AMERICAN ETHNOLOGY, Washington, D.C., June 28, 1963. Str: I have the honor to submit the accompanying manuscript, entitled “Archeology of the Yakutat Bay Area, Alaska,” by Frederica de Laguna, Francis A. Riddell, Donald F. McGeein, Kenneth S. Lane, and J. Arthur Freed, with a chapter by Carolyn Osborne, and to recommend that it be published as a bulletin of the Bureau of American Ethnology. Very respectfully yours, Frank H. H. Rossrrs, Jr., Director. Dr. LronArp CARMICHAEL, Secretary, Smithsonian Institution. gL f) N l4p> ; 17 198 _ LIBRARY_2 . aT CONTENTS PAGE TRIS FR VLAN ER Te a see ee aan Ix InrRopuctTIoN, by Frederica de Laguna-_----------------------------- 1 fine probleme. 23-50. S225. 5 soso Je ee ee =A 1 Traditional history of the Yakutat area__._..-.-.--------=------- 3 Srninay with bilrOpeuis 2. ssesoo 226 ooo sso ees aes See eee 10 Tue Yaxurtat ArRzEA, by Frederica de Laguna-_---_-------------------- 13 Geurrap hye ses 525 o05 55S So SSS 8 UL ie seas ee Se eels 13 (Goolgricaehanves 60 S20 cls 2. ste 5 4S Lae REE es 15 Seihement On WY aKUtAl DaYosaso52-.sSss~s— 3 Ss See Seale = 20 Sethioments in the Anka area... .> 2-25 2-- = See ees es = 23 Settlements, Lost River to Italio River. o2v- --2 22225-5226 252SL25 24 Battlements im the Dry bay area. 2525 23-2. >=. -Steeeas ee = 28 Oup Town, Knieut IsLanp, by Francis A. Riddelland Frederica de Laguna_ 31 Native: tradiions. 02) 222 2.3 Se ee ee bass pee seks ae 31 RNRIRN LOM fps 0 ain bao soe a ee 33 Shaman’s grave, Koicht Islandas 3 seu 2c ee. gel2ee 5-25} eee ees 35 Excavation and mapping techniques - -...—.---------+-5-=-------- 36 Stratigraphy of the trash mounds___.._...----------------------- 36 Hovusrs AND Cacuss, by Francis A. Riddell and Frederica de Laguna---- 43 MAIR BSG ae eae oe hg rs Si SS aaa Sa sees Se eee 43 Smaller surface and subsurface. pits... =.-L0 22/2825 vet so lsslSee 45 Sic Sten ne es BOUSG Sos oo 2 Ss Sa Ss a SE SLOSS 2 48 irieiyen = OCIS BS 2 oo 45 so baw ee ae sees oj ES SSL PS 51 GUNek Biro 2 SS on ee re eee eee a= 2 SOS 52 ROO fee oe TS BER os EA a BS SBSH SEE BAe eS. = 54 igloos a ae ee a Eee 54 She t bath r OGRE Sh see ee ot aa Sa SO SS 54 Are | Sees epee te Oi LS by Rt ee Fn Fs de eer a Oe BOE 23 55 PEM GRU ey thd te So en Bo ea tan tals SR Se BS 2 58 ELOUINE:. ONS PU OS sg oo FS ee Ole Dee OOO BOE E 61 \ CCL LE Se ee Seo ereme me Lie ies ese tell. ae 61 PRR gh he en al Pa Bae Say hs se A ES 63 BOG lS Pe, BS. Se eee HUGS ENS 63 Bear ihe ease bo nee esa pe oeee see See cece 63 Imirusive featuresic 2.220. A ele eo MUI AS eles 63 Rest. Pits ee soe te eo le BO Se 2. OPE Oe oe 64 Mcgee Et irag Se ee eee eee 82, BA ae oe eh Save Ch be BONING Ee 65 BP BTIGEONCH el oho semi fe ee ee st Ie RA 66 Gorninarisonae cs: 22.8. UPL bP Sure. oo ee 2 re SIO A AeRea FE . - 73 ANALYsIS OF FAUNAL REMAINS FRoM Otp Town, Knicut Istanp, by J. Arthur treed. and Kenneth §..Lane.-.2-.22 2-224 =. 22220 Foe esse. 77 Artrracts, by Frederica de Laguna, Francis A. Riddell, and Donald F. McGeein- 85 IntroGuction. 2. sel. 2) Jo eee 2 SUSE TS Be DARL 85 @byects or copper-ue = 2) toot ae 2b aOR? BNO ie Lek 87 Objecte aharOnew ows e sey soos a SR LON BAe = 88 Adzes, axes, and small woodworking tools-_----------------------- 90 Splitting adzes=. 2-52 fe 2 PE TO e NAS Sele ele oes 90 TV: CONTENTS Artiracts, by Frederica de Laguna and Francis A. Riddell—Continued Adzes, axes, and small woodworking tools—Continued PAGE MA ice 2 Te sere am ea a a ne, ope 92 Planing adzes <2 2-2 eee ee eee ere es 8 Se eee 93 Adz irarments 2255.2 oo ees eee ete ees Se ek ee 95 Small svoodworkingtoolss 2234 23-2 es eee 95 Rubbing; tools2... 2222 = 4. 2 A ees ei Marae 99 Bone; burins;andchisels:...2.5-25-555 aesee e 99 Knives,scerapers, and, choppers- 23.2 See oie ee ee ores 99 Wl0se.o2os6n See a a ee ta a 99 Ulo-with. lateral handles #2222. 22a. Salus Oe eee eee 103 Ulo-shaped bark. scraper or knife=2 73255. 5 103 Crooked knives. (?)i2... 22255255 Sees Se ee ee 104 Stone iscrapérs.. 2.22 252 ee ee ee ee ave 105 Bone scrapers. 22-225. 0 8 Se eee ee ee Se ee 108 Hammerstoneés,,anvils,,(?),, and maulse sae 2s eee ee eee 108 Hammerstones. 222-2 442 283s oe ee ee ei a tee 108 Anvil (Biol tied. (RG, Aoi ge eh Hela ae Sine ae 110 Unhaftedvhandimanisiorspestles seas ee 111 Hafted maulthesds: S27 26 Cee ee ee ee ee eee ee 112 Stone saws (?), grinding slabs, whetstones, and paint_____________-_- 113 Stone/sawsn(?):.< 22-23-2552) See ee ee le eee 113 Grinding slabs_....2.2.+.42...5) eee eee 114 WhetstoniesS..2uiyeco2fb eo {onli 2 ee eee See ee eee 115 Paing 2 oe le Be Ss ee Sr 116 Stone Jamps:and fire making... <- 4 hye ee Se eee lal Stone lamps <2-.--2...<---2c. 5 eee eee ee ee 117 Wire sMmakine 62 = uj ook 2 See ee eee 122 WieabOns® 22.355 55 26s
sine cow ARIE tps == ae Paes «Mga 4h) jyemty al mA oe wie. 1 = ooh bsetraheanaig 9 29 ieee by eee sneer mo @ & ees ~f ny ae ae * ch A < - Als f pA? | vie 4 7 = i = = < - > aa f Fong <— = tal . am ek os s 4 A ye a oe ieges _prplapaad ees if ~' ogee Pal at @ 01) 4¢se04h thm Agas os half 2 Sclaeee ian ibaa eee : abel (ie mero LST eee Sap pe: ta labtas oF ae=ies te ae os ei So aneabort , ’ weet ee - Ee ee i Cet Sac? WA itp ._dgnbane tombe lavith aicfheaney Wy » ig aveliore aetse tw | + alt Lie ne we wecltya Avie oy » : F 7 Ue\ Cr vlad), Gan 5 —_— ibe Hiatal De ae mau et a > oH wrap niet ianed Gale w ences a nap 2 E “see bys potibes gus ola Hu genase © tas) Da \es-s= | Set he pee — a ewile--'s . aida unio) ud er i + =F cape ipamranypesl a Wa ise eir ‘ pe prebiaah laine ‘anaes > 9, 5g Say Par « = (hee ee SON OO Pp WH VII ILLUSTRATIONS PLATES (All plates except frontispiece follow p. 226) . Sites in the Yakutat Bay area. (Frontispiece.) . Knight Island, Yakutat Bay. . Lamp and petroglyph. . Objects of iron. . Splitting adzes, axes, and war club. . Planing adzes. . Small woodworking tools like miniature adz blades. . Small woodworking tools like chisels and burins. . Stone scrapers and choppers. . Rubbing tools, whetstones, and hammers. . Stone lamps. . Stone lamps. . Barbed bone harpoon heads. . Copper arrowheads, knife blades, and pins. . Bone arrowheads and small weapon points. . Bone chisels, awls, and small tools. . Ornaments. . Twined weaving. . Fragments of Chilkat blanket from shaman’s grave, Knight Island, Yakutat Bay. TEXT FIGURES PAGE Penchematice profile of Mound By Old) Towns 2422.52.22) se2.2-c2hece 37 SeeLorace House. and Subsurface Pit: $8..-...22.....-2---&~-.-2-4--- 49 BEGERGIIBE DUCGIS ues hep Mert sc a5. ns a cee ae 52 TBLECONNULIC HON, OF VOUS Gu. o2Gs co. toe t ooo Kae cae ead eeee oe 53 5. Cross section of Trench 51, through House 8 and House Pit 1_____- facing 54 6. Cross section of Trench 53, through House Pit 1 and House 9______ facing 58 epi ECOR louse 0. 22 lek Fee Bee he et gh Be 60 BELL OCOMSTEUCHIOINGL ELOURG 9) foo ete ee ee oe 62 9. Cross section of Trench 33, through Mound C and House Pit 7___-__- 65 SOMES SCT AGN ee eect ce oe eb cee aes gee a an Wee RE ee 102 Inn V CH ARG BUEADETS.. <2 sae ee 2 Se Dk es oo SAA ey ee 106 Pere nuod OF StonG lampas: - 22. Seta) oe ee oR ee bee ee 119 13. Blades and points for large weapons____.._..._--_-..------.--------- 126 14. Ground slate blades for weapons or knives____.____..-___-_----__-- 128 Heat ped Nesan ANd, WOOMEM DIN. 2-2. 2 fe ee es oe 132 EEN NSIT OM IEC hd nee we et at ote ee en ek 140 ie Garbed points and ‘arrowheads... ==. 2 222 So) Se A ba ee BO 144 iS.) Devices. used. in fishing and trapping....22.2.5 2.22.22 2. 24222522 152 Why | 2 Sherr Se ee eae ye ee Pe oe oe eee eee 156 BRMIEHOPAUCCLODICUISES seer a ee ee EE eS ee the a 166 geearved and mcised stone objects: _.._2 2 2-228 170 22. Wooden figurine, floor of Storage House, Old Town II (194)_________ 173 2a box irapments Ana baud Of prass. 2-2 So) 2 ee ee 176 eee COTE Oe tet enero 2a oI As Re ie 182 25. Diagrams of weave of Chilkat blanket from shaman’s grave, Knight LIVES OM EESTI o 1, 13h ee ge ne a a ae Se 190 VIII ILLUSTRATIONS MAPS PAGE Le he: Gulf Coastiof Alaska. >. 2 Set @eee 3 oe eer eee See facing 2 2. The Coastal and Inland Tlingit and their neighbors_____-_-_------- 5 Sh, Srey mel arey Meany IBID pba alten = oe ee eee 14 A. Sites in southeastern Yakutat-Bay ot 25> _ S238 22s Sete eee facing 22 5. Map of Knight Island, Yakutat Bay, with detail indicating the site of Old Towne 2.2252 5. en ose eee ee eee ae 32 6. Map of Old Town, Knight Island, Yakutat Bay____-__-_.---------- 34 i. Map of Mound B, Old Town, Knight Island==22222_223) === facing 36 FOREWORD This report deals with archeological investigations in the Yakutat Bay area, Alaska, which were undertaken as part of a larger program of coordinated archeological and ethnological studies of the northern Tlingit. The ultimate objective of the program, as originally formu- lated, was to gather materials on the history of northern Tlingit culture and to analyze as far as possible the factors and forces respon- sible for the development and decline of Tlingit cultural patterns. These studies were begun in 1949, when Frederica de Laguna, senior author of the report, made an archeological and ethnological reconnaissance to select some area or areas for the proposed research. On this trip she was assisted by Edward Malin and William Irving, then students at the University of Washington and the University of Alaska, respectively. As a result, the Yakutat region on the Gulf of Alaska and the territory of the Angoon Tlingit in southeastern Alaska were chosen. A report on the fieldwork at Angoon in 1949 and 1950 has been published by the Bureau of American Ethnology as Bulletin 172. In the summer of 1952 combined archeological and ethnological fieldwork was carried out at Yakutat. The archeological investiga- tions were continued in the summer of 1953, and the ethnological work in the winter and spring of 1954. While Frederica de Laguna was in overall charge of this research, the archeological parties were led by Francis A. Riddell (now State archeologist for the California Department of Parks and Recreation). He was assisted in 1952 by J. Arthur Freed, Kenneth S. Lane, and Donald F. McGeein, and in 1953 by Lane, McGeein, Albert H. Olson, Jr., and Robert T. Anderson (now assistant professor of sociology and anthropology at Mills College), then all students at the University of California or recent graduates from that institution. Dr. Catharine McClellan (now associate professor of anthropology at the University of Wisconsin) collaborated in the ethnological work at Yakutat in 1952, and Mary Jane Downes (now Mrs. Benjamin Lenz, then fellow in anthropology at Bryn Mawr College) served as ethnographic assistant in 1954. In connection with this program, a study of Eyak linguistics was made Ix x FOREWORD at Yakutat in 1952 by Dr. Fang-Kuei Li, Department of Far Eastern Studies, University of Washington. When it became apparent that the ethnographic investigations should be extended to the neighbors of the Yakutat Indians, Frederica de Laguna and Catharine McClellan collaborated in studying the Atna of the Copper River during the summers of 1954, 1958, and 1960. The field researches at Yakutat were supported by the Arctic Institute of North America, with funds from the Office of Naval Research, in 1949 and 1953; by the Wenner-Gren Foundation for Anthropological Research in 1949 and 1952; by the Social Science Research Council and the American Philosophical Society in 1954; and by the Department of Anthropology, University of California at Berkeley, in 1952 and 1953. ‘The University of Pennsylvania Museum and Bryn Mawr College were also sponsors. A faculty research fellowship from the Social Science Research Council and the hospitality of the Berkeley campus during the senior author’s sabbatical leave have permitted her to finish this report. We wish to express our gratitude to the above-named organizations and also to acknowledge the assistance generously given by the U.S. Geological Survey, the National Park Service, the U.S. Fish and Wildlife Service, the Alaska Native Service, and the U.S. Coast Guard. We are also grateful for the valuable information furnished by Dr. J. Louis Giddings, director of the Haffenreffer Museum, Brown University; by Elizabeth Ralph, Radiocarbon Laboratory, University of Pennsylvania; by Dr. Charles E. Borden, University of British Columbia; by Dr. William O. Field, of the American Geo- graphical Society; by Dr. Calvin J. Heusser of the Osborn Botanical Laboratory at Yale University; and by George Plafker and the late Don J. Miller, of the U.S. Geological Survey. In preparing this report, Francis A. Riddell was responsible for the original descriptions of the artifacts and of the archeological features at the Old Town site, although his preliminary draft was later revised. Donald McGeein collaborated with Riddell and also drew all the maps, diagrams, and text figures, except a few prepared by HE. F. Chapman, Mrs. Arlie Ostlie, Irene Brion, Richard A. Gould, and Frederica de Laguna. The photographs of specimens were taken by the late Reuben Goldberg, of the University of Pennsylvania Museum, with the exception of a few by Kenneth Lane and the Campus Studios of the University of Washington. Arthur Freed and Kenneth Lane prepared the analysis of faunal remains. We are indebted to Carolyn Osborne for the description and interpretation of the Yakutat blanket. While the senior author has been responsible for the organization and FOREWORD XI editing of the report in general, her specific contributions have been the ethnological and comparative sections, and the historical and theoretical speculations. For assistance in the final editorial preparation of the report for publication, we wish to thank Mr. Edward G. Schumacher, illustrator for the Bureau of American Ethnology, and Mrs. Eloise B. Edelen, editor for the Bureau, and her assistants. FREDERICA DE LAGUNA, Bryn Mawr College; Francis A. RIDDELL, California Department of Parks and Recreation; Donawp F. McGere1n; Kennetu S. LANs; J. ARTHUR FREED. * * * * * * * NOTE The system used herein for transliterating native words is essen- tially that employed by Boas (1917) for Tlingit, except that digraphs are used for affricatives, and A, 5, 1, and U are substituted for Greek letters. The archeological specimens are deposited in the University of , Pennsylvania Museum in Philadelphie, although for the most part they are designated by their field catalog numbers (in parentheses). 6 We r spat Oy A gay rire dodaetecd? wh sahiidne oP hae SUSE. ae esto nib bid 5 tle paereaarl his Rotter aie BT's aay Re are See Vi 384 iw tegs Sh! PH» * ay (uy i; anes Viney oie ; d yh =v Het Des Be im “sible ee ~ Med A wre eee _— eres eats Is = ja ihe +, « i i * 2 ; re me Pit 0; fwe a | ry ASIA BE alin ay omikact ‘UME ans start ead) gil aye Jeyethy dais Jodd dere OS, OR, esiytiord errohe rope yoni Gol tine Lah h Sxie Ul hig t aia isin als r > Vee rele i, Charbavidl) Bs ik! -tlouggali: Thi a8 Suit: fomisoind ib tact pe eae) bd hist scent i ae ! ‘hs ae ae pry renner ay isancatinornd al) i} olga! abbas fF un Saag i dnieeainds eed ¥a Se ae eat bie . + oh? hunt id n ie ' Oyao Sogn) ae pl om = oe Cape St. Elias ayak I. ky Cope >) “ \ can ened KY, , \ 4 Glacier_- Dy ¢ 41 eri ns ae nig ht Point Maney, Be) Ocean Cape o x iF Ig ( nig, ote SCALE Cape Fairweather’ KS = MI,CRILLON, Lituya Bay > = o 5 50 100 MILES ve Poi 2X4 50 0 50 100 KILOMETERS at ae nS wd Cape Spencer | | a | = ee ‘g a mS a a Yakobi I 16 144° 142° 140° 138° 7 WILLIAM 5G SyA Z }? 4 inchinbrook I .. le be Aishihiké? yee *%, } 2S Aishihik = 7 tp { X Keene Yee bs \ we. 23 CT 1g Oren Pley keefield S ier( AN wy Te AMT. HUBBARD, GL TF \)\\) se Ng oe ag MT AUS oN ee os id Katalla _Chilay A Poi Si u; ay (fy wigs: <% i per ) arroller ae Q Z LOS e "2 — XS Dy ); OXF Os SF D) “= A i ¢ ‘4 ae De Sve Ip ee ‘a m i) i 4/ = yy is WY) SS \\ | | Map 1.—The Gulf coast of Alaska. Redrawn by Richard A. Gould from U.S.G.S. Topographic series, “Alaska, Map B.” i : % i all ¢ “7 oa a sak \' > . My Be * i aA de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 3 Cordova on the eastern edge of Prince William Sound to the Italio River, alittle over 30 miles east of Yakutat. Still farther east, the inhabitants of the Akwe River and Dry Bay area are reported to have spoken Athabaskan (Tutchone?). Already in the 18th century was being felt that movement of Tlingit from southeastern Alaska which intro- duced Tlingit speech and culture to Yakutat Bay, and some Tlingit were apparently even then living at Lituya and Dry Bays. The Russians in 1788 and Malaspina in 1791 met Tlingit in Yakutat Bay; Colnett (MS., 1788) noted that the natives there spoke different languages. We do not know when Eyak was completely abandoned in favor of Tlingit. Some items of material culture, notably the hunting canoe with forked prow, link the Yakutat with the Eyak of the Copper River Delta. According to Birket-Smith (Birket-Smith and de Laguna, 1938, pp. 530 f.), Eyak culture represents a very ancient phase of northern Northwest Coast culture, somewhat modified by more recent influences from the Eskimo and the Tlingit. Although it may be impossible to trace Eyak speech southeast of the Italio River, it seems likely that all the northern Tlingit area was occupied until relatively recently by small scattered populations with a simple form of Northwest Coast culture, quite possibly one similar to that of the Eyak (de Laguna, 1953). Nowhere in northern Tlingit territory are there large or numerous archeological sites comparable to those of the Koniag or Chugach or to those on the southern British Columbia coast (Drucker, 1943). Tlingit sib traditions would indicate a very recent expansion of population, owing in part to immigration from the south, perhaps under pressure from the Tsimshian and Haida, in part to immigration of Athabaskans from the interior, and in part to local population growth. This expansion probably accompanied the development of classic Tlingit cultural patterns. An important factor may have been contact with the European traders in the 18th century which made possible a richer life on the coast. The same processes by which the coastal Tlingit of southeastern Alaska absorbed and acculturated originally non-Tlingit elements presumably operated in the Yakutat area, where some of the events and changes are remembered in oral traditions. It was expected, therefore, that archeological research at Yakutat would reveal a rather simple type of culture, resembling the culture reported ethnologically from the Copper River Eyak, and that the more recent sites might document the growth of Tlingit influence. TRADITIONAL HISTORY OF THE YAKUTAT AREA According to informants at Yakutat, their ancestors once occupied all of the Gulf of Alaska from Cape Martin, east of the Copper River, to 4 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 Cape Fairweather between Dry Bay and Lituya Bay. The Eyak of the Copper River Delta farther west were not a distinct people; indeed, the inhabitants of Cape Martin were grouped with them in Yakutat thought, although the Cape Martin people are said to have been originally an offshoot of those living in Controller Bay. Vague traditions suggest that the area between Cape Fairweather and Cross Sound may have been occupied by the same groups that lived at Dry Bay, though mixed with Tlingit of southeastern Alaskan derivation. In any case, this Lituya Bay region is now deserted, though claimed as hunting territory by a Tlingit sib of Hoonah, in southeastern Alaska. (Map 2.) All of the Gulf of Alaska Indians are said to have been divided into exogamous matrilineal moieties (Raven and Eagle), like those of the Tlingit, and the legendary history of the area is told in the form of sib traditions. The interior Athabaskans—the Atna of the Copper River and the Southern Tutchone of the upper Yukon and Alsek Rivers—also have a similar social organization, and the Yakutat people felt that they were related to them through migration and intermarriage. Excluding the Copper River Delta on the west and the Lituya Bay- Cape Spencer area on the southeast, the Gulf of Alaska may be divided into the following four districts, according to native thought: (1) Controller Bay and the shore almost to the Icy Bay area is claimed by the Galyrx-Kagwantan, an originally Eyak-speaking Eagle sib who settled at the Kaliakh River (from which the first part of their name is derived), after “the Flood.” There is, however, strong evidence that a branch of Chugach Eskimo may have occupied, and certainly frequented, Controller Bay during the 18th century, until they were driven from it by the Tlingit or Tlingitized Eyak from the east. This is attested by Chugach and Copper River Eyak traditions, by Eskimo place names in Controller Bay, and by the observations of Steller and other 18th-century explorers (Birket-Smith and de Laguna, 1938, pp. 341-354; Birket-Smith, 1953, pp. 19, 20). Before their expansion westward, the Galy1x-Kagwantan were presumably living between Cape Suckling and Cape Yakataga. The distribution of the Eyak language suggests, of course, that at a still earlier period the Indians lived in Controller Bay. The division into Copper River and Yakutat dialects may have been caused by the subsequent intrusion of the Eskimo. Indian tradition tells of Chugach raids on the village at the Kaliakh River, and even as far east as Yakutat Bay. The versions of some informants that Yakutat Bay was originally occupied by Eskimo may reflect Chugach occupation of Controller Bay and their warlike excursions into Yakutat territory, as well as the fact that skin boats, like those of the Chugach, were once used on Yakutat Bay. ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA de Laguna] (‘Es6T ‘A8ojoaeyoiy uvoswy Joy Aqa1o0g “6 loweuw ‘; °*S WOl,J) ‘eunse’y] op voleapelg Aq suonvoyipow yiIM URIPIOIN euLlvy ed JO uoNdelIp oy Jopun ‘ueuldeyD “yy AqumeIq = ‘sioqysiou Joy} pur WSU, purjuy pue [eiseod oN —Z AVI as) St) ’ * b: . ie : “~\ ‘ Tan x N s \ => ‘ ‘ ‘ . 2 ’ . . ae H oe \ BNINILS ANIC ; ie . ae f : nonduna yf RS se a a atk ae Se , = AN SS. NOOONY we os . of s AVL ; " Vises vas Sy “. LIONTLL Ky ‘IWLSVOO ; i NILV \ = an oor iene oS = i H j 0. H \ + NY SY - AM y, C nid . ‘ C s Ww we Y= LOOWTIHD ” ~LVHIHD i\ Ww A zw =p as =— (7 Ave “Sg AVLN HVA - oa . Bue a eri face HSIow “6 «) ee oo \ ou) NINS3L el. ge HSID VL fs ~4 LION dort rts dNVINI 2 693-818—64 6 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 (2) Icy Bay, Yakutat Bay, and the coastal plain as far east as the site of the Yakutat airfield, but excluding Russell Fiord at the head of Yakutat Bay, compose the territory of the K”ackqwan, a Raven sib who trace their origin to the middle Copper River near Chitina. (3) The district from Lost River near the Yakutat airfield to just east of Italio River, including the head of Russell Fiord, belongs to the Teqwedi, a Tlingit Eagle sib from southeastern Alaska. The Bear House lineage owns the western area of Lost and Situk Rivers; the Drum House lineage claims the eastern lands on Ahrnklin, Dangerous, and Italio Rivers. (4) Akwe River and Dry Bay area belongs to the Tluk’axadi, a formerly Athabaskan-speaking Raven sib, and to the TY uknaxadi, a Raven sib from southeastern Alaska, with whom the remnants of the original inhabitants have merged. Also resident in the area, although they have never established full territorial claims, were the Kagle Tcukanedi, associated chiefly with Lituya Bay, and two other Kagle sibs from southeastern Alaska: the Kagwantan (proper) and the Cankugqedi. The latter came via an interior route from Lynn Canal, up the Chilkat and down the Alsek, and are said to have inter- married with the Southern Tutchone Athabaskans. Although there were once many settlements in these regions, they are now deserted, and the population of about 250 to 300 natives is now concentrated in the modern town of Yakutat. Only a handful of Eyak are reported to be at Cordova, and the little tribe is almost culturally and linguistically extinct. Some of the Dry Bay and Lituya Bay people emigrated to Hoonah and Sitka. In 1880, there were 170 ‘‘Thlinket” in Controller Bay, 150 near Cape Yakataga, and 500 at Yakutat and on the mainland as far south as Cape Spencer (Petroff, 1884, pp. 29, 32), making a total of 820, if Petroff’s figures are to be trusted. Although this count was taken after the disastrous smallpox epidemics, there is no reason to suppose that the Indian population on the Gulf Coast was ever very large. The legendary history of Yakutat begins some ‘“‘ten generations” ago, when the ancestors of the K”ackqwan emigrated from Chitina on the Copper River because of an intrasib quarrel. At that time they spoke Atna Athabaskan, and are referred to as the Ginexqwan or “people of Ginex’’ (Bremner River, an eastern tributary of the Copper River). The emigrants are said to have ascended this river and crossed the glaciers. Part of the group that became separated from the rest eventually became the Ganaxtedi Raven sib of the Kyak at the mouth of the Copper River. The main party traveled across the ice, past Mount Saint Elias, which they now claim as a crest, and reached the coast somewhere west of Icy Bay, which was de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 7 then filled with ice. Here the Gmexqwan met and intermarried with some Galyrx-Kagwantan who were moving eastward by canoe. The extensive icefields of Bagley and Bering (or Guyot?) Glaciers which the emigrants had to cross, as well as mountain ridges up to 5,000 feet high, might seem to present an insuperable barrier and so cast doubt not only upon this tradition, but also upon the report that copper from the interior was carried to the coast via a “shortcut” to the mouth of the Duktoth River at Cape Yakataga. Don J. Miller, of the U.S. Geological Survey, who knew this whole area thoroughly, assured us, however (letters of October 30 and December 6, 1957), that not only was the route possible, but that it had actually been followed by prospectors in the early 1900’s on the basis of the Indian legend. The natives would have come up the Tana River (a southern tributary of the Chitina), up Granite Creek or Tana Glacier, and then over Bagley and Bering Glaciers, and down the Duktoth River to the coast. Miller furnished some details of the prospectors’ journeys, the first of which were made in 1905 and 1906. Crossing the glaciers took from 3 days to almost 3 weeks. One of the men found a piece of split wood, 2 feet long, on a moraine (of Bering Glacier?), apparently left there by the Indians (cf. also Moffit, 1918, p. 77). Miller reports the ice along most of the route as “relatively smooth and little cre- vassed—treally good traveling, as glaciers go.”’ Native accounts vary as to how long the G@mexqwan and their spouses stayed near Icy Bay, but eventually they came to Yakutat ‘Bay, which was then largely covered by a glacier. They crossed the bay, walking over the ice according to some informants or using skin boats according to others. The islands in the bay and the eastern shore were already owned by a group or groups, variously identified as Chugach or as Indian. Our most knowledgeable informant called them the Hinyedi, a Raven sib (presumably Eyak-speaking), although there may have been other small tribes in the area. From them, the Copper River immigrants acquired by purchase the territory along the shores of Yakutat Bay, including the stream, K”ack (‘Humpback Salmon” in Eyak), from which the sib takes its present name. Payment was made in copper which they had brought from the Copper River. After selling their lands, the Hmyedi are said to have emigrated to southeastern Alaska, although we suspect that some of them merged with the K’ackqwan. One group with hunting camps on Yakutat Bay and settlements along Lost and Situk Rivers to the east were called the Tlaxayik- (“Yakutat Bay’’)-Teqwedi. They were an Eyak-speaking group closely related to, or possibly a branch of, the Galyrx-Kagwantan, and were responsible for the destruction of the Russian post at Yakutat in 1805. Slightly prior to this, the first of the Tlingit 8 BUREAU OF AMERICAN ETHNOLOGY [BuLu. 192 Teqwedi from southeastern Alaska were marrying into the Yakutat Ravens. Shortly after 1805, the Tlaxayik-Teqwedi were nearly exterminated by Tlingit parties from Akwe River, and the survivors were apparently absorbed by the true Teqwedi. Another group, now extinct, were the Duxedi, also an Eagle sib, possibly a branch of the Tlaxayik-Teqwedi. Their name refers to the muddy water of the Situk River, which at that time drained the ice-dammed lake at what is now the head of Russell Fiord. Meanwhile the Tlingit had been moving up from southeastern Alaska and were established in the Dry Bay area. This is reflected in the story of the man from the Hoonah district who taught Tlingit arts to the Dry Bay Athabaskans and who became rich by trading with them (see Swanton, 1909, Tales 32 and 104). Thesibs that moved in at this time seem to have been the Kagwantan, T?uknaxadi, and Teqwedi, although some of the Bear House lineage of the last sib were already living near Yakutat. The Cankugedi were presumably established by that time at Dry Bay through intermarriage with the local inhabitants. The Drum House branch of the Teqwedi pur- chased the Ahrnklin-Italio district from the Staxadi, a branch of the Hmyedi, and the Bear House lineage of the Teqwedi acquired the Situk-Lost River area by preemption. The latter were probably not secure in their holdings until the T?uknaxadi from Dry Bay, eager to get the Russian loot held by the Tlaxayik-Teqwedi, made war on the latter and nearly exterminated them. The Ti’ uknaxadi apparently did not enjoy their wealth for very long because they became embroiled in a war with one of the Chilkat Raven sibs. One of their war parties was lost when a number of their canoes capsized in Lituya Bay under mysterious circumstances. This occurred about 1850, and the disheartened relatives at Akwe River abandoned their town. Some moved to Hoonah and Sitka, and others came eventually to Yakutat. More important than the many wars and intrasib quarrels as causes for the abandonment of settlements were the various epidemics, of which the smallpox epidemic of 1836-39 was the most disastrous. The establishment of a trading post, of the mission, and of the cannery concentrated the scattered population at the modern town of Yakutat early in the present century. One of the most important legendary figures in the native history of Yakutat was Xatgawet, a Tlingit Teqwedi of the Bear House lineage. He is said to have been born on the Akwe River and to have traveled all over, even as far west as Katalla, marrying the daughters of local chiefs and acquiring great wealth from the gifts customarily bestowed upon brothers- and sons-in-law. This is, of course, a device used by the Tlingit to establish profitable “trade” with the Atha- de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 9 baskans (Olson, 1936, p. 214). He is said to have ‘organized’ the coastal groups and to have given them the names of Tlingit sibs, such as Ganaxtedi, Kagwantan, and so on. He also gave the maximum number of eight potlatches, a feat not afterward equaled at Yakutat. All of these stories suggest that while the earlier inhabitants of the Gulf of Alaska may have had matrilineal sibs and moieties (perhaps rather loosely organized, like those of the Copper River Atna), it was the Tlingit immigrants to Yakutat who introduced the fully devel- oped patterns of Tlingit social and ceremonial life. According to some informants, Xatgawet was also a shaman and acquired one of his familiar spirits from a Tsimshian colleague, a story which suggests the northward diffusion of shamanistic practices. Some informants say that Xatgawet bought Knight Island for his Gmexqwan wives and children, and that he assisted his brother-in- law in founding the village on that island and named it T’ukwan, or TPak*’an “Old Town,” after the famous Chilkat village (Klukwan). Our best informant maintains, however, that Xatgawet lived much later, after the Russians had been expelled. He is said to have been the grandfather of a woman who died shortly after 1900 and the ereat-grandfather of a woman who was born in 1874. Furthermore, it is denied that he had anything to do with Knight Island, but lived on Lost River. Giving him a post-Russian date would place him in the period in which Tlingit had replaced Eyak speech at Yakutat. It is possible that the traditions are confused because there were ‘several persons with the same name. In any case, the first Tlingit trade with Yakutat antedated the visits of the first Kuropean ex- plorers in the last quarter of the 18th century, for it was already well established at the time of the explorations of Ismailov and Bocharov to Yakutat in 1788, witness the arrogant behavior of the Tlingit chief, Yelxak (‘Iichak’’), from Chilkat (Shelikhov, 1793, pp. 228-229, 233-237; and in Coxe, 1808, pp. 324-325, 329-332). There is no doubt that the fur trade in the late 18th and early 19th centuries stimulated the northward expansion of the Tlingit. Asso- ciated with the diffusion of Tlingit trading patterns and the Tlingit language, many other aspects of Tlingit culture must have been spread, probably including the style of potlatches, of peace ceremonies, and of shamanism and witchcraft. All of these would have been reflected in such items of material culture as the large multifamily lineage house with totemic crests on carved house posts and painted screens, and ceremonial regalia of all kinds. The Tlingit also intro- duced the Haida-derived style in secular songs, and shamans’ spirit songs in Tsimshian. Tlingit trade also brought to the Yakutat people large canoes of southern manufacture (Haida and Nootka), Tsimshian-made dance headdresses, dentalia, abalone, and flat-headed 10 BUREAU OF AMERICAN ETHNOLOGY [BuLu. 192 slaves, in return for which the Yakutat traded their magnificent baskets, copper from the Copper River, and furs obtained locally or from their relatives in the interior or farther west along the Gulf Coast. CONTACT WITH EUROPEANS According to the natives, their first contact with Europeans occurred some time before the Russians established themselves at Yakutat. A ship was wrecked on the shore near Malaspina Glacier. Two men and a woman survived, but the men fell down a crevasse and only the woman was alive when the Indians found the wreck. The latter, through ignorance, spoiled most of the treasures they took from the ship. Thus, they put the guns into a fire and pounded up the barrels with stones to make spears. They could work iron because they already knew how to shape copper. At that time an iron spear point was worth a slave, and so the men became rich. One of them married the White woman, who lived to old age. From written sources we may infer that the first direct contact between the Gulf of Alaska Indians and the Russians was in 1783 when Potap Zaikov led an exploring party into Prince William Sound and Controller Bay. Other hunting parties, consisting of several hundred Aleuts and four or five Russians, apparently went down the coast, perhaps as far as Lituya Bay, but of these we have no details. Lituya Bay was visited by LaPérouse in 1786, where he met natives who may be taken as typical of the expanding northern Tlingit. The Indians there had iron tools and beads. One of our informants told about the coming of the first ship to Lituya Bay, and a fuller version was obtained by Emmons (1911) from a chief at Douglas or Juneau. Dixon visited Yakutat in 1787, and the following year Ismailov and Bocharov explored Controller, Yakutat, and Lituya Bays, as reported by Shelikhov, and Colnett also traded with the natives in Controller and Yakutat Bays and in inlets farther southeast. In 1788, Douglass anchored off the Ahrnklin and Dangerous Rivers, or off Dry Bay, where he traded with the natives, but he failed to discover an anchorage in Yakutat Bay. Malaspina’s more thorough exploration of Yakutat Bay was made in 1791. Brown traded in this area in 1792, 1793, and 1794. In 1793 a war party from Yakutat went to attack the Chugach in Prince William Sound, but, to their mis- fortune, fell in with Baranov and were defeated. That same year the Russians sent a party of Aleuts to Yakutat under the leadership of Shields, and in 1794 a large flotilla of bidarkas under Purtov and Kulikalov. This party met Lieutenant Puget with one of Van- couver’s ships (the Chatham) at Yakutat, and also the English trader, Brown, in the Jackall. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 11 The Russian post at Yakutat (actually on a lagoon inside the southeast point of the bay) was begun in 1795 and fortified the following year. In 1800 a second post, or blockhouse, was built on Monti Bay, near the present mission or near the entrance to the Ankau lagoons where the first post was located. In 1802 the In- dians attacked an Aleut hunting party at Dry Bay and accused the Russians of robbing graves, a charge which is still remembered. Our informants also listed other grievances: the failure of the Russians to pay for the land they occupied; closing the stream (with a fish weir?) between the Ankau lagoons and Summit Lake to the east, which seriously interfered with the natives’ supply of fish; taking children with promises to educate them but actually using them as slaves; and, lastly, appropriating native women at their pleasure. As a result, the Russian post was finally destroyed in 1805, and all but a few of the occupants were killed. The same year the Yakutat again invaded Prince William Sound, but this war party was annihilated by the Chugach. (For the Chugach version, see Birket-Smith, 1953, pp. 141 f.) In 1806 Campbell rescued an Aleut man and his wife whom the Yakutat had captured, and took two Indians to Kodiak as hostages. Our informants also told how the Tlaxayik-Teqwedi leader of the attack on the Russian fort was taken to Kodiak. It was not until the following year, however, that the widow and children of the Russian commander were liberated, together with a few other ‘survivors. After this a period followed in which there were few close contacts with Europeans, except when trading parties went from Yakutat to Nuchek in Prince William Sound, or to Sitka, or even to Prince Rupert and other distant, southern trading posts. We have few records of European visitors to Yakutat until about 1880, except for the Russian cartographers, Boolingin in 1807 and Khromchenko in 1823, the British navigator, Belcher, in 1837, and the U.S. Coast Survey in 1874. Although the ocean off the coast was a famous whaling ground, vessels seldom put in to shore. The first American traders began to appear at Yakutat shortly before 1880. At that time, a White man was killed and his Indian slayer taken to Portland on a gunboat. Later, the U.S.S. Adams landed a party of prospectors at Yakutat, and between 1883 and 1886 there were goldminers working the black sands of Khantaak Island and the ocean beach. Trading schooners began to call regularly, and parties attempting to climb Mount Saint Elias stopped at Yakutat to recruit porters with almost equal regularity. A Dr. Ballou ran a 12 BUREAU OF AMERICAN ETHNOLOGY [Buin 192 trading post where the mission is now located. During this same period, parties of Tsimshian and Tlingit from southeastern Alaska came to hunt seals in Yakutat Bay and sea otters in Icy Bay. The mission was established in 1888, the ‘Old Village” at Yakutat was founded shortly afterward, and the cannery was built in 1904. THE YAKUTAT AREA By FREDERICA DE LAGUNA GEOGRAPHY The region in which our archeological investigations were pursued includes Yakutat Bay and the coastal plain to the southeast as far as Dry Bay. Yakutat Bay has its entrance between Point Manby (59° 41’ N., 140° 19’ W.) and Ocean Cape (59° 32’ N., 139° 51’ W.), where it is 16 miles wide. It extends northeastward about 33 miles, narrowing to a width of 3 miles or less, then turns southeastward toward the ocean for a distance of 28 miles. The coastal plain between Yakutat and Dry Bays is about 50 miles long, and from 5 to 14 miles wide between the open Gulf of Alaska and the snow-covered peaks of the Saint Elias Range. The southern shores of Yakutat Bay, including the chain of islands (Khantaak to Knight Islands) along its eastern edge, and the fore- shores northwest and southeast of the bay, are all low-lying terrain, less than 250 feet above sea level. Most of this land is composed of alluvial gravels, sands, and silts, and is studded with lakes and swamps. The western shore of the bay is covered by the terminal moraine of Malaspina Glacier, and similar outwash deposits are found along the opposite shore. All of the permanent native settlements are in these lowland areas. (See numbered sites on map 3.) Steep rocky shores are encountered at Eleanor Cove, on the eastern side of Yakutat Bay near Knight Island, about 15 miles above the mouth of the bay, and on Bancas Point to the northwest. Here the land rises sharply to altitudes of over 4,000 feet, leaving only small areas at the mouths of streams where the natives camp in the spring when fishing for halibut or hunting bears and seals. Disenchantment Bay, the “heart” of Yakutat Bay, north of Point Latouche and Bancas Point, is filled with floating ice discharged from the glaciers that descend to tidewater from the high mountains of the Saint Elias Range, here over 14,000 and 15,000 feet in elevation. These icy waters are the principal seal-hunting areas of the region, and the natives also gather sea gull eggs from the rocky cliffs of Haenke (‘Egg’’) Island in Disenchantment Bay. From Disenchantment Bay, Russell Fiord continues southeast for about 10 miles, where it sends out to the west a 7-mile-long arm, Nunatak Fiord, at the head of which is another glacier. The bare 13 14 BUREAU OF AMERICAN ETHNOLOGY [BuLu. 192 slopes above Russell and Nunatak Fiords were formerly hunting grounds for mountain goats. From Nunatak Fiord, Russell Fiord turns southward, extending so far that its head, “‘ Mud Bay,” protrudes into the coastal plain only 14 miles from the ocean. The Alsek River rises in Yukon Territory close to the headwaters of the Yukon, Tanana, and Chilkat Rivers, and cuts through the high barrier of the Saint Elias-Fairweather Range to reach the sea at Dry Bay. The mudflats at the mouth, about 10 miles wide, are covered only at high tide. Formerly there were a number of villages or camps on the lower Alsek and at the mouths of smaller streams and sloughs that enter Dry Bay. Until fairly recently the Dry Bay people used to ascend the Alsek to hunt, fish, gather berries, and trade with their interior neighbors and relatives, while the Southern Tutchone of the upper Alsek used to visit Dry Bay. Interior trails connected the settlements of the Southern Tutchone with those of the Interior Tlingit around Tagish Lake and with the villages of the Chilkat Tlingit on Lynn Canal. A chain of sloughs, streams, lakes, and salt-water lagoons until recently provided an inland waterway for canoes going between Yakutat and Dry Bay, although some portages were necessary. It was on these streams and lakes that most of the earlier settlements were located, and here the natives obtained their supplies of salmon. Until 1875 or 1880 there were few permanent houses on Yakutat Bay itself; the most important village site was Old Town on the southern point of Knight Island. The eastern shore of Yakutat Bay as far north as Point Latouche, a distance of about 23 miles, is heavily timbered with spruce and hemlock. The former provided the natives with most of the wood used for houses, boats, and implements, while the sweet inner bark of the hemlock was used for food. In addition, cedar drift logs were sometimes found on the ocean beach and utilized. Dense stands of trees extend in narrow belts parallel to the ocean, but most of the plain between Yakutat Bay and Dry Bay is open country. The natives complain that the trees have recently been encroaching on areas where they used to gather strawberries, salmonberries, blue- berries, highbush cranberries, elderberries, Kamchatka lilies (‘wild- rice’’), wildcelery, wild ‘‘rhubarb,” and “‘Indian-potatoes,” as well as a variety of medicinal plants. Also there is a tradition that originally there were no trees on the islands in Yakutat Bay, and that even as late as 1850 or 1860 Krutoi Island was not wooded. The Ya utat area enjoys a fairly equable climate, since the thermometer rarely drops below zero or rises to 80; but there is very heavy precipitation, averaging about 130 inches a year. Over 4 feet of snow may accumulate at one time on level ground, and mountainous ’ HBO O O oo PT. MANBY (v1 ‘d a0e%g) F9—-O 818-€69 693-818 O—64 (Face p. 14) PT. MANBY Gy el \ (,6La R os oh a or * es 4 OCEAN i8(Nessuda ‘ \9(Diyaguna’et) FISHING CAMPS OST RIVER r ae la f ; Q Ne L lv 4 UD D o eS ty «Le . Y . ° CA, fo )) ] 4 Gyr if Wid BLACK SAND |. ie, OO FORMER AKWE CHB ‘lie USTAY R. jl mS ° 5 _gisning _——— orAMP NAUTICAL MILES o oe AFTER U.S. COAST & GEODETIC SURVEY 8402 WITH CORRECTIONS FISHING CAMP EAST OR EASTING R. 7 BO DOHN R~ Map 3.—Sites in the Yakutat Bay-Dry Bay area. Drawn by Donald F. McGeein. tha e = Pad Homg $13 gaant?e wOMJAS ZOeTOMUN” sa eali-yweter i» PA swe %, o a VW ip? canbes roi = > or Lrinares Were Ne . ) aa SI the catlicr sot@ ockuinel ibe a kee pupplies ¢ 15 As rg <= im vasso : ah idee on Tek ae | Re tas Ane) ae aa ae : wa Oh i ee i TRY _— Mabe isoitt ~ 4 + : 2 7 j Sere E. J 54 | sewao om Ls gt ~ [p : ; ‘ > ver F lites tin VGeD ge ale 2 £0Au0 af F 3 aus QALY nee”, Pa ae a “omar . tack TES of we levee ate Hing | Wavaue oItsdoae # TEACO ay sose —t enbi¥agnnod | wriw ———— nn i Wout th ore Sy. te li » halt ade Ar easia—-vb wall de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 15 drifts form in the forest in winter. Most of the offshore winds blow from the south and east, but on land there are often storm winds from the mountains to the north. The tidal range at Yakutat is about 10 feet. The most important animals of the area that were utilized for food or skins were: fur seal, harbor seal, sea otter, porpoise, sea lion, whale (eaten when found stranded), black bear, brown grizzly, land otter (until recently avoided as supernaturally dangerous), mountain goat, wolf, fox, wolverine, beaver, muskrat, and marmot. Moose, rabbit, and deer have come or been introduced into the area during the present century. The most valuable fish are the salmon—king, sockeye, humpback, and coho, to list them in the order of their runs; there are also halibut, eulachon, herring, steelhead, etc. Bird life is particularly abundant, especially in the Ankau system of streams, lakes, and lagoons between Ocean Cape and Lost River. The most significant for the natives were swans, geese, salt- and fresh-water ducks, terns, gulls, and other aquatic birds. Cockles, clams, mussels, chitons, crabs, and sea urchins were gathered in the sheltered waters along the eastern part of the bay or in the salt lagoons of the Ankau area. Edible seaweed was obtained on rocky points off Ocean Cape or on the outer shores of the islands. Icy Bay was an important area for hunting mountain goat, seal, and sea otter, and the Gaty1x-Kagwantan territory west of Cape Yakataga was noted for its fur bearers, especially beaver and sea otter. GEOLOGICAL CHANGES There have been a number of geological changes in the Yakutat Bay region since it was first settled, and recently discovered evidence of these changes (Plafker and Miller, 1958) have tended to confirm native traditions (de Laguna, 1958). The whole area was probably buried under ice during the Pleistocene, and while human occupation may have been possible during the recession which followed the Wisconsin glacial period, we have no evidence of it. After this, there was another advance of the ice, so that Malaspina Glacier and the two lobes that filled Icey Bay and Yakutat Bay formed a continuous front of ice along the sea. The coastal plain east of Yakutat Bay and west of Icy Bay was apparently unglaciated. The eastern edge of the glacier filling Yakutat Bay ran northeastward from Ocean Cape, covering the area around Lake Redfield, and a smaller ice lobe extended south from the head of Russell Fiord, although the land around Lost and Situk Rivers was not glaciated (Tarr, 1909, map p. 106; Plafker and Miller, 1958). The culmination of the glacial advance in Icy Bay was roughly between A.D. 600 and 920 (A.D. 756 + 160 years), 16 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 and in Yakutat Bay was between A.D. 970 and 1290 (A.D. 1127 + 160 years), according to radiocarbon dates obtained from wood in end moraines near Icy Cape and Ocean Cape (Plafker and Miller, 1958). During the subsequent recession, the ice retreated as far as or even farther than the present glacial fronts, permitting the growth of trees well behind the present timberline in Icy Bay, in Russell Fiord, and in Disenchantment Bay. ‘This retreat began somewhat before A.D. 1400, to judge from the age of living trees near Yakutat. These geological changes are further discussed on pages 204-206. Native traditions seem to imply that this recession was in progress when the ancestors of the K”ackqwan came from the Copper River; in fact they are said to have caused it by throwing a dead dog down a crevasse. Icy Bay was then completely filled with ice, and the immigrants crossed Yakutat Bay on the glacier that extended from Point Manby on the west to Eleanor Cove on the east. The native name for Yakutat Bay, Ttaxayik, is derived from the Eyak ta’ (glacier), xa’ (near), plus the Tlingit suffix -yik (place inside). The glacier was melting back, exposing the bay, and Yakutat, yak”dat, is supposed to be an Eyak expression meaning ‘‘a lagoon (or bay) is already forming.” It was presumably during this same recession that a village was founded on Guyot Bay, just inside the northwest point of Icy Bay. It was eventually overwhelmed by a readvance of the ice, which culminated during the 18th century. Tarr and Martin (1914, pp. 46 f.) quoted a version of this tradition, recorded by Topham in 1888, and believed that the glacial advance took place between 1837 when Belcher sailed into Icy Bay and 1886 when Schwatka saw a solid wall of icein the bay. Plafker and Miller (1958) have advanced convincing evidence that the “Icy Bay” of the explorers from Vancouver to Schwatka was really the former outlet of the Yahtse River, east of the present Icy Bay, and that the latter was already full of ice by 1790. Malaspina Glacier probably advanced at the same time as the glaciers in Icy and Yakutat Bays. A radiocarbon date of wood from its moraine indicates that the climax of the advance was about A.D. 1750 + 150 years. The advance in the Yakutat area was much less extreme and affected only the glaciers in Disenchantment Bay and Russell Fiord. The maximum extent of the ice in Yakutat Bay may have been to Blizhni Point and to a corresponding locality on the eastern shore, midway between Knight Island and Point Latouche (Plafker and Miller, 1958). Although the glaciers were already in retreat by 1791, Malaspina was stopped in June of that year by ice that filled Disenchantment Bay as far south as Haenke Island. Tarr (1909, p. 20) believed that this was only floating ice; if so, glacial conditions may not have been de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA Ly very different from those of today. Even now floating ice in early summer may prevent travel above Haenke Island, and we have seen bergs drifting down to Knight Island. On the other hand, there has, in general, been a great retreat of glaciers near Yakutat since the 18th century, and other geologists (Russell 1892, p. 172; Tarr and Martin, 1914, pp. 108 f.; Plafker and Miller, 1958) believe that Malaspina encountered a solid wall of glacial ice in the vicinity of Haenke Island. The natives told us that before they had guns (which they did not acquire until the end of the 18th century), they were unable to camp above Point Latouche because of the floating ice. The main sealing camp was then 38 miles south of the point, at a place called Tlaxata, an Hyak word referring to the proximity of the glacier. In the early 19th century, after the destruction of the Russian post in 1805, the natives made a fortified camp at Wuganiysz, about 2% miles above the point. At the end of the century the sealing camps just above Point Latouche were great centers (Grinnell, 1901, pp. 158-165). Remarks made by some informants suggest, however, that there may have been a period in the middle of the century when these places were little used because of the ice. Until the middle of the last century, Russell Fiord was blocked by glaciers which dammed up a fresh-water lake at the southern end of the fiord. This barrier, undoubtedly due to the 18th-century ad- vance, extended from Beasley Creek to Cape Stoss. The name for the latter was an Eyak word, meaning “‘it has the glacier in its mouth.” This lake was drained by the Situk River (see Tebenkov, 1852, chart vit; Davidson, 1904, map v1). At that time it was possible to travel by canoe from the lake, through a series of lakes and streams, to Yakutat Bay just below Knight Island. The ice barrier broke some time between 1850 and 1875, according to our informants, when the dammed-up lake waters were discharged into Russell Fiord, reducing the Situk River to a small stream. Tarr and Martin (1914, p. 230) estimate that the vegetation on the old lake beach at the head of Russell Fiord was not over 50 years old in 1909-13. This change in the size of the river must have affected adversely any settlements on the upper Situk, where a fortified village had been built shortly after 1805. We do not know, however, whether any attempt was made to reoccupy this site after the original owners, the Tlaxayik-Teqwedi, were massacred at their camp at Wuganiyz. Native tradition also refers to a breaking of a glacier bridge across the Alsek River probably about the same time or toward the end of the century. Prior to this, the river had flowed out under a tunnel of ice. The Ttuk”axadi from Dry Bay, when making their annual trips to the interior, had to carry their canoes overland through a gorge on the west side of the river in order to bypass the glacier, and 18 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 on their return downstream would paddle fearfully under the ice. The collapse of the ice is said to have created a great flood that drowned many people in Dry Bay. There were also ice-dammed lakes formed at the headwaters of the Alsek River, and a Southern Tutchone informant reported that his mother had twice seen the ice break and the water rush out in a flood, the first occurrence being about 1842, according to his estimate. Retreat of the Icy Bay glacier did not begin until about 1904 (Plafker and Miller, 1958). Our informants said that this was be- cause a dead Tsimshian sea otter hunter had been eviscerated (to preserve his body) and his entrails buried at Guyot Bay (‘“Tsimshian Bay’’), just inside the northwest point of Icy Bay. This happening was evidently after 1890 and before the death of Yakutat Chief George in 1903. Although we have no Yakutat traditions concerning the topography at Lituya Bay, Tarr and Martin (1914, p. 10) estimate that the glaciers advanced about 3 to 3% miles between 1786, the time of LaPérouse’s visit, and 1906. About 1850 a flotilla of canoes, said to have come from the Akwe River, were overturned in Lituya Bay and all the oc- cupants drowned. Possibly this disaster was caused by giant flood waves, evidence of which could be dated at 1853 or 1854 through a ring count of trees that had sprung up on the devastated area (Miller, 1960, pp. 67 f.). Still farther south, the ice in Glacier Bay on the north shore of Cross Sound has retreated about 55 miles since the latter part of the 18th century. Prior to that, there was a long period of recession when the glaciers were even smaller than they are now (Field, 1932, p. 371). The advance of the ice which destroyed a Tlingit town in this area recorded in one story by Swanton (1909, pp. 337 f.) may be the move- ment of the ice to its maximum extension in the 18th century. A very important recent geological event was the Yakutat earth- quake, which lasted for 3 weeks during September 1899. Although the center of this disturbance was 15 to 30 miles up the bay, waves washed away the native graveyard at Point Turner on Khantaak Island and threatened the mission at Yakutat. Avalanches fell all along the shore between Knight Island and Point Latouche, and giant waves in this area destroyed forests over 40 feet above sea level. The earthquake also produced considerable changes in sea level, although apparently not at Yakutat itself or on the foreland to the east. At the extreme western end of Phipps Peninsula there was subsidence of 7 feet, and some stretches of shore north of Knight Island were similarly depressed. Other areas along the eastern shore were raised from 1 to 7 feet; Haenke Island rose 17 to 19 feet, and the west side of Disenchantment Bay reached a maximum elevation of 47 feet. The de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 19 axis of tilting ran directly through the village site of Old Town on Knight Island, but fortunately the main portion of the site was not damaged. ‘The effects of the earthquake are described by Tarr and Martin (1906). Tarr and Martin (1906, pp. 52 ff.) believe that there had been previous changes of sea level, and they cite raised beaches now covered with forests on Krutoi and Otmeloi Islands, and an elevated beach south of Point Latouche (and Tlaxata) with trees only 75 years old in 1906. We observed similar old beach lines on the south point of Knight Island. In fact, according to Don J. Miller (letter of October 9, 1957), there is evidence of very recent emergence of land areas, from beneath both the sea and the ice, all the way from Copper River to Icy Point beyond Lituya Bay. The earthquake of 1899 is credited with shaking down so much snow on the glaciers in the Yakutat area that the latter were stimu- lated to renewed activity between 1905 and 1910 (Tarr and Martin, 1914, p. 37). Since then most of them have been in retreat, except that we observed that Turner Glacier in Disenchantment Bay had recently advanced farther south and east of the position shown on the U.S. Coast and Geodetic Survey chart No. 8455 (1945, 5th ed., chart of 1901). This observation was confirmed by the natives, who said that they no longer dare to camp on Osier Island nearby because of the danger of waves from calving bergs. Other recent changes in the Yakutat Bay area have been the drying up or shallowing of the sloughs connecting the Ankau lagoons with Lost and Situk Rivers, and shifts in the sandbars at the mouths of the Situk and Ahrnklin Rivers that have resulted in some disturbances of the salmon runs. On July 9, 1958, Yakutat again felt the effects of a severe earth- quake, when waves drowned three persons on Khantaak Island. The southwest end of the island near Point Turner was, according to reports, “lifted forty to fifty feet in the air[!]” and then submerged (New York Times, July 10, 1958). My local correspondents do not make clear whether the old village site on the island was affected. At the same time, a landslide in Lituya Bay raised giant waves which denuded the mountain slope near the head of the bay to the prodigious height of 1,720 feet. All the shores of the bay were lashed by waves that stripped them of vegetation and that overrode the three habi- tation sites near the mouth visited by LaPérouse in 1786. Trim lines in the forest growth indicate that similar giant waves, but of lesser extent, had previously devastated the shores of the bay: in late 1853 or early 1854 (although no major earthquake was reported), about 1874, in 1899 (probably associated with the Yakutat quake), and in 1936 (caused by a landslide) (Miller, 1960). Native traditions re- 20 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 count the complete obliteration of a village near the mouth of the bay, from which the only survivors were the men who had been out hunting sea otters and a lone woman who had been picking berries on the hills (Williams, 1952, p. 137). Possibly it was thas village, and not that on the Akwe near Dry Bay, which our informants should have associated with the Lituya Bay drownings, since the mourning song commemorat- ing the tragedy is supposed to have been composed by a woman who actually saw her relatives drown. Probably the Akwe village, occupied by the same or related sibs, was deserted about the same time, leading to an association of the two events. Although there had been no great waves in the bay for some time prior to 1786, in the opinion of Miller (1960, p. 56), who studied these phenomena, the Tlingit were well acquainted with the treacherous character of the bay, as evidenced by the traditions recorded by Emmons (1911). The dangers they feared were probably something more than ordinary storms or tide rips at the entrance. The whole southern part of Alaska is subject to seismic activity, and there were probably other earth movements which may have affected former village sites, but of which we have no direct or clear evidence. SETTLEMENTS ON YAKUTAT BAY A number of former villages or camps within the Yakutat Bay area were reported by our informants, but we were able to investigate only a few of these sites. The available information about them is sum- marized below. Unfortunately the early explorers are not very definite or specific about the location or nature of the native habita- tions they saw. Probably Colnett (MS., 1788) was correct when he estimated that the 200 natives he met at ‘“Koggy Harbour,” or Port Mulgrave, had their homes to the southeast and came to Yakutat Bay only to hunt, fish, or trade. He believed the huts he saw were only temporary summer dwellings. It is to be noted also that Beres- ford with Dixon in 1787 (1789, p. 169) noted ‘‘several huts scattered here and there in various parts of the sound.” Malaspina in 1791 seems to have found a village on or near Port Mulgrave, that is Khantaak Island, but does not indicate its location on his chart (1802; 1885, p. 156), although the latter shows the cemetery inside Ankau Creek, on the north shore, near the site of the present Alaska Native Brotherhood Cemetery. 1. On Ankau Creek, in the vicinity of the cemetery mentioned above, there was evidently a village, according to Dixon’s chart of 1787. The grave monuments which he and Malaspina describe evidently stood nearby. Vancouver in 1794 (1801, vol. 5, p. 396) also noted a village about 2 miles ‘‘within cape Phipps.” This site de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA Di was only half remembered by one of our informants. We searched for it without success. (For sites 1-6, 14-17, cf. map 4.) 2. The village on Port Mulgrave, Khantaak Island, was called Suska. The modern village, remembered by our informants from the 1880’s, was perhaps founded in 1875-80 to take advantage of the visits of trading schooners, but was located at the same spot indicated on Dixon’s map of 1787. ‘This village was abandoned by 1893. Exca- vation is impossible because the whole site is covered by a graveyard. 3. The “Old Village” of Yakutat, about three-fourths of a mile north of the cannery, was founded about 1889 when the mission was built nearby. It is still occupied, although most families moved to the present town of Yakutat in 1919 to build permanent homes near the cannery that had been established in 1904. Part of the lowland where the original houses stood at the Old Village has been washed away. 4, Sites were reported on both sides of Canoe Pass, a channel leading east from Johnstone Passage, but we were able to find only a very small shell heap on the island forming the north side of the pass. The deposit consisted of 3 inches of humus, 12 inches of stones, ash, and shells, and 7 inches of concentrated shells (clam, cockle, mussel, and sea urchin) at the bottom, forming a total depth of 22 inches. Only a cut bird bone (C/3), a rectangular slab whetstone of shale (C/2), and a quartzite hammerstone (C/1) were found, which gave no clue to the age of the site. 5. Although a former village site was reported on the east side of Dolgoi Island, we were unable to locate it. However, a site was discovered near the mouth of a small stream on the south end of the island, about 100 feet from the beach. The site is perhaps 300 feet long and 100 feet wide. Where tested, the deposit consisted of sticky black soil and fire-cracked rocks, only 6 to 10 inches deep, and con- tained a cobble hammerstone (D/2) and a ground slate tool (D/1). That the site may be of considerable age is the fact that a huge tree, fallen in 1952, once grew on the cultural deposit. 6. Various former camping places were reported on the north end of Khantaak Island, on ‘‘Crab Island” nearby, on Krutoi Island, and at the mouth of Humpback Salmon Creek opposite Krutoi Island. Some of these camps were mentioned in stories of Chugach raids, but the site at the stream was said to have been occupied by the Hmyedi, the orig- inal owners of the area. An unsuccessful search was made for it. We also failed to locate the settlement on the north end of Khantaak Island, although it is marked on Dixon’s chart near the location of a White man’s home which we visited. 7. The important site TY ak*-’an or ‘Old Town,” on the southern tip of Knight Island, was reported to have been originally settled before the Russians came, but informants differed as to whether 693-818—64__3 2? BUREAU OF AMERICAN ETHNOLOGY [Burn. 192 it was first a Hinyedi or a Chugach camp, or a KYackqwan village. There is also uncertainty as to why or when it was abandoned. In 1791, Malaspina (1885, p. 164) noted grave monuments here or in the vicinity, similar to those at Ankau Creek. Since Russian days it has been used as a camping place, and a White man has a cabin nearby. It was here that our archeological work was concentrated ; a full discussion and description of the site is given later. 8. Another site was reported at the mouth of a stream about half a mile east of Old Town on the south shore of Knight Island. A White man and his native wife now live here. We could find no trace of the site. (Numbers for sites 7 and 8 are transposed on map 3.) 9. The small rocky island close to the mainland east of Knight Tsland is called ‘Little Fort” in Thngit (pl. 1, a). It is supposed to have been fortified “in the days of Xatgawet”’ as a protection against Chugach raids. In the clearing on top of the island the outlines of the fort walls are preserved, and they confirm native traditions that the foundations of fortifications were often of stone. The walls of rough cobblestones can be traced for a distance of about 70 feet along the east side, and seem to enclose a rectangular area, 112 by 225 feet, within which shelters of some kind were presumably built. In a 6- inch cultural deposit of dark-brown rocky soil, outside the east wall, a piece of copper, probably a knife (ulo) blade, was found (F/1). The island would now be hard to defend, but it had more precipitous sides before it was elevated about 12% feet during the earthquake of 1899. 10. A former Chugach (?) camp was reported at the mouth of the stream opposite the north end of Knight Island, but was not visited. The land here rose about 5% feet during the earthquake. 11. The old sealing camp, Tlaxata, is said to be back in the woods on the north bank of the large stream about 3% miles below Point Latouche. We were unable to land there. Malaspina (1885, pp. 162-164) found natives camped here in early July 1791. 12. Three sealing camps used in post-Russian times, and described by Grinnell (1901, pp. 158-165), were at the mouths of streams approximately 1%, 24%, and 3% miles above Point Latouche. They were called, respectively, ‘‘Burned Down” (in Eyak), Wuganiyr (meaning the same in Tlingit), and “Big Valley” (in Tlingit). Wuganiyr was said to have been surrounded by a stone wall with loopholes for guns, but the Tlaxayik-Teqwedi defenders were mas- sacred by the Tlingit TYuknaxadi from Dry Bay. Although this place was visited in May 1954, the snow was too deep to permit exploration. The three camps were visited by the Harriman Alaska Expedition in the spring of 1899, when they were occupied by 300 to 400 natives from Yakutat, Sitka, and Juneau. During the earth- 693-818 O—64 (Fame p. 22) LAGOONS = WITH CORRECTIONS si 3 GS PT. CARREW PT. TURNER i 8 OLD YAKUTAT ' ) 4y Oa pS, U.S. COAST & GEODETIC SURVEY es 8455 N eieateeiieiatttttinsianeneneempercnaticmmauced . “ - A he U.S.C.G. LORAN STA. “SUMMIT LAKE” aa \ $> NAUTICAL MILES Map 4.—Sites in southeastern Yakutat Bay. Drawn by Donald F. McGeein. pectabtl oO a ey 7 wad theo iva ne. The . 9 @8tence of pbout 70 fee he tanpuler | uypa, 112 by 2 ay pee wee audbty bull Ie: soelny wel, ovbside thes ye Stade, eed found Fit). at noche jane oh asaaut 4 Me Taltuxey O50 Mh “~ Poem te aae ~eab a aur “suas saeere’, ae | eai:68 2aartwae Pid ‘Coos t : . ~ (a ao a bo eapar teguisY nveteaodiboesad 491182 aa de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 23 quake that fall, the shore was elevated from 7% to 12 feet. Modern camping places are on the south shore of Haenke Island where a flat was raised above sea level by the quake, and on the mainland opposite. 13. The Tlaxayik-Teqwedi are said to have had a camp on high ground near Bancas Point on the west side of the Bay. SETTLEMENTS IN THE ANKAU AREA The Ankau lagoon system of Phipps Peninsula is entered from Monti Bay via Ankau Creek, and consists of several salt-water lagoons and lakes. These are connected by streams with Rocky Lake, Aka Lake, and Summit Lake to the southeast. From Summit Lake, near the U.S. Coast Guard Loran Station, Lost River flows southeast- ward to enter the ocean about 11 miles from Ocean Cape. 14. The site of the Russian post, Nova Rossiysk, (‘‘New Russia’) (1796-1805), was on the narrowest part of the barrier beach between the ocean and the largest of the Ankau lagoons (‘‘Russian Lake’’). It is supposed to have contained “seven buildings defended by a stockade, and five others outside” (Dall and Baker, 1883, p. 207; see also Tebenkov’s map vu, 1852). Although we visited this spot several times with native guides, we were unable to find any trace of the fort. An Indian had a fishing camp and smokehouse at this spot; the ocean has evidently washed away much of the land. The natives also reported that the Russians fortified a small island in the lagoon and erected a “‘gate’’ (fish weir?) across the stream, T’awal, that drains Aka and Rocky Lakes into the lagoon. In 1948, on or near this stream, a Yakutat native, since dead, found a limestone rock, carved in typical Northwest Coast style to represent a bear (pl. 3, 6). We were unable to discover just where it had been found, and although this is the only known petroglyph in the Yakutat area, the natives believed that it commemorated the defeat of the Russians by the Tlaxayik-Teqwedi. The rock was taken to Yakutat, where we saw it, but it later disappeared. 15. There was a former K*ackqwan village, “On the Lake,” at the middle of the ocean side of Aka Lake. The occupants died in the smallpox epidemic of 1836-39. The site was afterward used as a fish camp until fairly recently. It is now a clearing, but our brief exploration revealed nothing more than broken crockery and scraps of iron. 16. The stream connecting Aka and Summit Lakes was said to have been a canal, “dug by slaves,’’ which probably means that they deepened or widened it at some point. Moser (1901, p. 383, map on pl. xtvi1, Yakutat to Dry Bay) reports that: ‘The rocks and boul- ders have been removed from the bed,and piled along the side, forming a Shallow channel up which canoes are tracked at low water, but may 24 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 be poled at high water.” He locates this area on the lower stretch of the creek, just above the lagoon. A village at the junction of the stream and Aka Lake was occupied by Eyak-speaking Indians who were killed by the Tlingit Teqwedi. We were not able to visit this site. 17. Summit Lake drains both into the Ankau system and into Lost River. On the ocean side of the outlet toward Lost River there is said to have been a village, called in Tlingit, ‘“Town on the Hill,’’ because of its situation on a sandhill. It was occupied first by Eyak- speakers and later by the K*ackqwan, who all died in the smallpox epidemic of 1836-39. We did not explore this locality. SETTLEMENTS, LOST RIVER TO ITALIO RIVER The stream which flows southeast from Summit Lake is erroneously designated as ‘“Tawah Creek” (U.S.C. and G.S. chart No. 8402), or as “Ankau Creek” (U.S.G.S. topographic sheet, ‘‘Yakutat’”’). The natives call it “Lost River,” and give the same name to the lower part of the stream which it joins, referring to the upper part of the same stream as ‘Little Lost River.’”’ We shall use the term ‘Lost River” to refer to the lower part of the stream; designate the stream from Summit Lake as its western branch; and retain the native name “Tittle Lost River’ for the small northern branch. 18. Before the Russians came, there was a small settlement on the western branch, approximately opposite the ‘‘“Number Two” runway of the Yakutat airfield. When the Russians were expelled, this became the principal village of the K*’ackqwan, with at least four large lineage houses and other smaller homes. The inhabitants were virtually wiped out by smallpox. Later the surviving K”ackqwan moved to Khantaak Island, and the site was used as a fish camp (only?). It is called Nessudat. Moser in 1901 (p. 384) noted three houses and some fish racks at this locality. The site occupies a fairly large clearing on the ocean side of the stream and is marked by at least three house pits and several cache pits, although none of the former could be identified as the ruins of any of the particular houses mentioned by our informants. We made test excavations along the cut bank of the stream, in the cultural deposit at the western end of the clearing, and in a house pit at the eastern end; but found only objects suggesting recent occupation. The house pit measured about 18 feet square, with postholes and remains of the corner posts some 5 inches beyond the end walls. The cultural deposits in the midden and in the house pit consisted of light- brown or gray sandy soil, containing ashes, flecks of charcoal, and occasional fire-cracked rocks. Maximum depths were about 24 inches in the midden, and 38 inches within the house pit, where cache (?) de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 25 pits were encountered below the floor level. The clearing extends about 300 feet along the stream and is about 160 feet wide. Unfortu- nately, it was only as we were leaving that we noticed what appeared to be a much older house pit in the dense wood west of the clearing. Material from Nessudat included: A cylindrical beach cobble used as a pestle (pl. 10, k), a hammerstone (N/10), a fragmentary siltstone whetstone (N/2), a small iron ball, 2.7 cm. in diameter (N/1), an iron spike and an iron nail (N/5), some small green, small white, and large white glass beads (N/6), and some fragments of English soft paste porcelain. These all came from the test hole in the house pit. From the midden on the riverbank were: a fragment of copper sheeting (N/14), a broad flat piece of iron, possibly from a can, folded and shaped into a knife or scraper (N/20), and an iron knife blade (pl. 4, 7). Other items from the midden were: a clear glass liquor bottle, remains of a tin can, a large white bead (N/7), a blue glass bead (N/8), and an iron spike. On the ocean beach near Nessudat was found an iron spearhead which we were able to borrow from the finder for sketching (see fig. 13, d). 19. On the ocean side of the west branch, one-fourth of a mile above its confluence with Little Lost River, is the site of Diyaguna’st, an Eyak word meaning ‘‘Salt water comes in here” (pl. 1, 6). It origi- nally belonged to the L’uxedi or Muddy Water People, and after changing hands several times, was finally acquired by the Bear House lineage of the Tlingit Teqwedi. It became their principal village under Xatgawet, although the latter is reported to have lived in his own house farther upstream at a place called “Strawberry Leaf” in Kyak. ‘The village was visited by smallpox in 1836-39, but a number of inhabitants survived. During Teqwedi occupation, the village is supposed to have consisted of three or more houses, surrounded by a palisade. We were told the names of seven houses, but since one lineage house might have several names, we do not know how many actual buildings were implied. The village was inhabited up to about 100 years ago. One of the houses apparently had a carved bear figure above the door, or on a post that served as the doorway. The site is on a sandbank about 100 yards long, 50 yards wide, and 20 feet high, which is now being undercut by the stream. ‘The midden deposit of humus, charcoal, and fire-cracked rocks is from 4 to 18 inches deep in most places, but in one spot the bank has caved away to expose an old house pit containing a cultural deposit about 48 inches deep. In this fill were found a broken barbed slate blade (see fig. 14, 2) and a fragment of a slate ulo or scraper (49—25-108), at depths of 24 and 30 inches. On top of the bank were three house pits, measuring 32 by 32 feet 26 BUREAU OF AMERICAN ETHNOLOGY [Buny. 192 and 5 feet deep, 30 by 20 feet and 4 feet deep, and 30 by 20 feet and 2 feet deep. Two other rectangular depressions and a circular de- pression, about 20 feet in diameter, may also be house pits. Two other circular pits, 12 and 8 feet in diameter and 5 and 8 feet in depth, respectively, were probably for caches. None of these surface house remains appears to be very old. A rotted post was found in the corner of one of the smaller pits, and test excavations in the largest house pit uncovered one of the rotted roof beams and a piece of commercial copper sheeting with nail holes. The slumped bank and the bed of the river are littered with fire- cracked rocks, pieces of commercial copper, iron nails, fragments of china, and scraps of burned bone. Some of the china (49—25-114) has been identified by Arnold R. Pilling, then a graduate student at the University of California, as probably of English manufacture between 1830 and 1875. Other pieces are of true Chinese porcelain, later than Canton ware, but similar to that introduced into California about 1850. The following objects of native manufacture were found on the bank of the river: a planing adz blade (pl. 6, d), a sandstone slab possibly used as a saw to cut stone (49-25-109), a hammerstone (-110), a piece of worked greenstone (—111), and half of a round sandstone lamp like a complete specimen previously found here by one of the Yakutat natives (pl. 3, a). The same man also found two stone blades for splitting adzes. There was also a small whetstone (pl. 10, d) a scraper made of copper (see fig. 11, f), and an iron dagger (see fig. 13, 5). This material indicates habitation in both prehistoric and modern times. 20. A small site on the west side of Little Lost River, about one- half mile above its confluence with the western branch of Lost River, is called “Shallow Water Town” (in Tlingit?), and is supposed to have been the oldest village of the Luxedi. It was acquired by the Tlingit Teqwedi, and finally given by a Teqwedi chief to his K¥ack- qwan brother-in-law, who planted native tobacco here. The site is a clearing about 200 feet long, about 75 feet wide, and some 20 feet above the streambed. There are no cultural deposits in the cut bank. An indefinite depression, about 20 by 20 feet, may be a house pit. Our two test holes indicated cultural deposits to depths of 10 and’20 inches, consisting of humus, ash, charcoal, and a few fire- cracked rocks. Three blue glass beads of the kind seen by Captain Cook in Prince William Sound in 1778 (S/2), an iron arrowhead (S/1), and a lump of red ocher were found. These may date from late protohistoric times, but fail to corroborate the native claim for great antiquity. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 27 21. Several houses and tent frames mark the modern Indian fish camp on the west side of the mouth of Lost River. Two native houses on the east bank of the river, at the end of an abandoned railway spur from the cannery at Yakutat, opposite the mouth of the west branch, were built in 1919 but are no longer occupied. One house at the end of the railway was formerly ornamented by two carved wooden brown bear paws, from which it derived its name, but these were stolen in the summer of 1952. 22. Situk Village was on the east bank of Situk River, and extended between the railway trestle and the U.S. Fish and Wildlife Service station. It was founded by the Teqwedi about 1875-80, and was abandoned about 1916 (?). The site is marked by a few graves and the remains of collapsed framehouses. In the clearing opposite the Fish and Wildlife station there are some cache pits, about 3 feet in diameter and 8 inches deep. Those explored contained only ashes, rocks, and an iron bolt. In the river above the Government weir, men at the station found a grooved maul head, carved to represent an animal (see fig. 21, d). We were unable to discover any trace of an earlier site, although there may have been one. 23. ‘“Hagle Fort,” reported to consist of four houses connected by tunnels and surrounded by a palisade, was built by the Tlaxayik- Teqwedi shortly after 1805 because they feared Russian retaliation. Here they repulsed an attack by the T?uknaxadi from Akwe River, but were defeated the next spring at Wuganiys (site 12) in Yakutat Bay. We were unable to visit the site of this fort. The name for the locality is known by both Eyak and Tlingit werds. 24. A single TY uknaxadi house was built about the middle of the last century on Johnstone Slough, about 11 miles above the end of the railway from Yakutat. The fish camp at the mouth of Johnstone Slough and Situk River is modern. 25. The main village of the Drum House branch of the Tlingit Teqwedi was reported on the Ahrnklin River, about 2 miles above the mouth, apparently near the confluence of the two main branches. The village and also the river were named “Big Town of the Animals’ (Tlingit), referring to the rich hunting in the area. There were said to have been four houses there, and the river, now undercutting the site, exposes charcoal to a depth of 4 feet. The village was abandoned when most of the inhabitants died, either in a feud or from smallpox. A Teqwedi settlement on the Ahrnklin, called ‘Wolf Cave,” may be the same place, or possibly a lineage house at this town. There are said to have been no villages on Dangerous River, but the Teqwedi are supposed to have lived on Italio River before they pur- chased the Ahrnklin area from the Staxadi. There were some 28 BUREAU OF AMERICAN ETHNOLOGY [Buu. 192 fishing cabins on the Italio in 1909 (Robson, 1910, photograph on j Oem 7a) ye SETTLEMENTS IN THE DRY BAY AREA Our information about former settlements in the Dry Bay area is unsatisfactory, owing in part to the shifting stream channels which render available maps inaccurate or confusing to our informants and in part to the great difficulties of travel encountered by Riddell and Lane in 1953. Just west of Dry Bay, the Akwe and Ustay (Akse) Rivers have a common mouth. The Akwe or western stream drains the lake at the foot of Chamberlain Glacier and also ponds and swamps near the Italio River. The Ustay (Us-tay of Moser, and Akse of Tebenkov) is formed of two tributaries, the one draining the lake at the foot of Rodman Glacier to the west, and the Tanis, draining Tanis Lake to the east. Near the confluence of these two branches, the Ustay puts out distributaries draining eastward and southeastward into Dry Bay: Gines (‘‘William’’) Creek to the north, and farther down the Ustay, the Kakanhini (‘Muddy’) Creek and the much smaller Stuhinuk (“Stickleback,” or ‘(Cannery Creek’’), both of which enter Dry Bay near the mouth of the Alsek. According to Tebenkov’s map (1852, map vu; cf. also Davidson, 1904, map vi), there were villages on both the Akwe and ‘‘Akse”’ Rivers, designated, respectively as ‘“‘Nearer’”’ and “Farther Village to the Mili- tary Post” (at Yakutat). The Coast Pilot of 1869 (Davidson, 1869, p. 136), relying on Russian sources, reports in roughly this locality (59°14’ N., 138°45’ W.) the common mouth of two streams, each with a village on it, some 6 to 12 miles (by winding channel?) from their confluence. 26. The westernmost of these two villages was probably the principal town of the TY uknaxadi, called Gusex. This was reported to have been on the Akwe River, apparently at the confluence of the main or northern with the western branch. Tebenkov’s ‘Nearer Village” is west of the river and north of a large slough. According to the Coast Pilot of 1883 (Dall and Baker, 1883, p. 206), about 1870 (or earlier?), one of the several villages between Yakutat and Dry Bay was visited by the captain of a whaler anchored at Yakutat. He reported this as ‘‘the largest, finest and most clean Indian village he had seen in all his experience on the coast. The population was large, the houses well built, solid, and adorned with paintings and carvings of wood, and expressly adapted for defense.”” It was per- manently inhabited, parties leaving to trade or to hunt seals in Disenchantment Bay. The description would fit what was told fof Gusex. Unfortunately, the chart (ibid., opp. p. 204) shows a village de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 29 at what would appear to be the confluence of the Ustay and a western tributary, but it is so inaccurate that no reliance can be placed upon it. According to our informants, Guésex had originally been an Atha- baskan Ttuk*axadi settlement before the Tlingit TYuknaxadi built houses here. ‘The names of five houses were mentioned, and the posts of one were said to have been visible 40 or 50 years ago. It was from this town that the TP uknaxadi sent war parties against the Tlaxayik- Teqwedi on Situk River and Yakutat Bay, and to it they brought the Russian cannon and other treasures taken from their defeated enemies. The town was abandoned about the middle of the last century, after many of the men from it were drowned in Lituya Bay when going to make war on the Chilkat. According to one native informant, this site was located about 4 miles (on a direct line) from the coast (see above); another placed it at the entrance to a former channel some distance down the Akwe and about 2 miles (on a direct line) from the shore. The latter described this place as having four or five house pits and one earth-covered house. The only site which it was possible for Riddell and Lane to reach was still farther downstream. 27. This was a site in a clearing on the west bank of the Akwe, opposite its (present) confluence with the Ustay River. There are several pits which may indicate former houses or caches, but digging failed to expose any cultural deposits. Tebenkov’s “Farther Village,” is located farther upstream, on the west bank near the junction of the Tanis and Gines. Natives re- ported a site on Gines (‘‘Williams”’) Creek. 28. A modern settlement was near the mouth of Stuhinuk Creek on the west side of Dry Bay. Here were native houses, some built as recently as 1909 or 1910, and the remains of a cannery, built and abandoned between 1901 and 1912. Moser’s map (1901, pl. xn1m) indicates a village here in 1901. About half a mile northeast of Kakanhini Creek and 2 miles north- west of Dry Bay, aerial photographs indicate an almost circular pattern in the heavy forest growth, according to Don J. Miller (letter, Oct. 30, 1957). ‘The photographs give the impression that trees were cut down around the circumference of a circle about 1, 200 feet in diameter, but were left untouched or only partly thinned out within the circle.”” Miller did not visit this locality, nor did our party, so we do not know what this might be, although it is suggestive of a fortified town. 29. Other sites reported in the Dry Bay area are on the north shore of the bay on the west bank of the Alsek, and a town farther down- stream called “It repeatedly shakes” (Tlingit). There were also settlements on Easting River, or possibly on Dohn River, streams 30 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 that cross the flats of Dry Bay east of the main, western, mouth of the Alsek. The first two settlements are said to have belonged to the Tiuk*axadi, but it was impossible to secure any clear information about their location or the period when they were inhabited. OLD TOWN, KNIGHT ISLAND By Francis A. RippEtt and FREDERICA DE LAGUNA NATIVE TRADITIONS Knight Island (map 5) is roughly 3 miles in diameter and lies close to the eastern side of Yakutat Bay, about 13 miles northeast of the town of Yakutat. The island is low lying, composed of sands, silts, and boulder clay. It is heavily timbered, and studded with small lakes, swamps, and creeks, none of which has salmon runs. The native name for the island, Ganawas, is not Tlingit, and we suspect it to be an Eyak word, although some informants think it is Atna or even Chugach. The island was the first piece of territory acquired by the ancestors of the K*Yackqwan when they came to Yakutat Bay from Copper River. At that time there were no trees on the island, and it was covered with strawberries. The owners caught a KY*ackqwan girl picking berries here and cut the basket from her back. Her father, a Gaty1x-Kagwantan chief, then bought the island for her and her people. Shortly afterward, a similar inci- dent involving the fishing rights at the Humpback Salmon Stream, ‘KYack (see No. 6 on map 3), induced the K*ackqwan chief to purchase territory on the mainland, thereby acquiring the present name for the sib and title to all the lands around Yakutat Bay. Some informants say that the original owners of Yakutat Bay and Knight Island were Indians (Hmyedi, Yrenyedi, or Qusqedi) ; others call them ‘‘Aleuts’’ (Chugach Eskimo), but agree that they had no permanent settlement on the island. The site which we excavated is usually described as a K¥ackqwan town, founded and abandoned before the Russians came. Its correct name is said to be “Raven Falling Down” (yet ada qutciyr), because smoke from the many houses would asphyxiate any raven attempting to fly across. The same name has also been applied to the reported site (No. 7) on the cove about half a mile farther east. A few informants said that the village was founded by a K*ackqwan chief who built the first lineage house here, “Fort House,” and that his Tlingit brother-in-law, the famous Xatgawet, built ““Bear House” and named the village “Old Town” (TPak*-’an), after the famous town (Klukwan) on the Chilkat River, to suggest that this also was the home of high-class people. While it is tempting to associate this tradition with the oldest house pit at the site, we must remember that our most reliable informant 31 32 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 a GULF OF ALASKA 4000 5000 NN mag. , K\ Decl. 29°E Oo 100 200 300 400 J00 Goertr iri tiriitiristiriss (PISN Sa Map 5.—Map of Knight Island, Yakutat Bay, with detail indicating the site of Old Town. Drawn by Donald F. McGeein. insisted that Xatgawet was a post-Russian chief associated with Lost River, and that the Knight Island site was pre-Russian. Furthermore, we found no clear evidence of post-contact occupation of the site. Although no definite tradition explains the abandonment of the site, some informants suggest that it may have been due to smallpox. This is not unreasonable, since the epidemic of 1775 is reported by de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 33 some informants to have spread to Yakutat from southeastern Alaska before the arrival of the first Europeans. After Old Town was abandoned, the main K*ackqwan village was at Nessudat on Lost River, and Knight Island was used only as a camping place for hunting parties. There are also confused stories of raids on Knight Island by the Chugach and by the mixed Athabaskan-Tlingit Indians from Dry Bay. It was impossible to secure details, and while some raids were apparently prehistoric, others were later, but we do not know whether Old Town was involved in any of them. In these stories, Teqwedi are mentioned as living or camping on Knight Island, and as having fortified the rocky islet, “Little Fort,’ nearby, but it is not clear whether Tlingit Teqwedi or Eyak-speaking Tlaxayik-Teqwedi are meant. THE SITE The site of Old Town (map 6), on the southernmost point of Knight Island, consists of four trash mounds, seven house pits, and numerous smaller pits (for caches, bathhouses, etc.), scattered over an area of 3 or 4 acres. A small stream that presumably supplied drinking water to the inhabitants flows along the northeastern border of the site and enters a little cove 500 feet to the east. The major portion of the site is an open grassy flat, bordered by a spruce-hemlock forest. In the clearing, besides ryegrass, there is a luxuriant growth of sphag- num moss, wildcelery, salmonberry and elderberry bushes, short grass, ‘and patches of nettles. In historic times the area is said to have been covered with wild strawberries, but the forest has encroached within the memory of the older natives and there are now many young trees in the clearing. Part of the site (Mounds C and D, and House Pit 7) lies within a mature forest growth apparently several hundred years old. The underlying soil is composed of banded beach sands, evidently elevated above sea level within relatively recent geologic times. Several low sand ridges that traverse the site from southwest to northeast, or that lie southeast of it, seem to represent former beach lines. The most prominent of these, just southeast of the site, was probably the shoreline at the time of habitation, and seems to have been raised during the earthquake of 1899. The axis of tilting ran directly through the site, so that the shore to the east of datum B (map 6) was elevated, while the land immediately to the west was depressed. A maximum elevation of 7 feet was noted by Tarr and Martin (1914, p. 106) at the cove three-fourths of a mile east of the site, and there was a subsidence of 5 feet on the south shore about an equal distance to the west. Fortunately the quake and the waves that accompanied it did no serious damage to the archeological re- mains, except at the southwestern edge where an indistinct depression 34 BUREAU OF AMERICAN ETHNOLOGY [Buy. 192 TS; above MLLW MOUND "D" HOUSEFIT “YT MAGNETIC N. TRUE N) 0.10 50FT. SS (PIT MOUND Tet Tree @ ---- Pit f —----— Excavated area (Elevations in feet above mean lower low water) Cabin eS EL TES TP it (Midden deposit is 18"deep in this area) Trees > Y HOUSE PIT 4 y Ly ha po) @ PIT iT © DATUM "A" DATUM Da (16.9' above MLLW) Map 6.—Map of Old Town, Knight Island, Yakutat Bay. Surveyed by Francis A. Riddell and Donald F. McGeein; redrawn by Irene Brion. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 35 (House Pit 3) has been partly washed away and filled with beach gravel. Beyond this is a low, swampy area with a few dead trees, evidently killed by salt water after the subsidence (pl. 2, 6). The earthquake may also have affected the course of the little stream, for just north of the site there is a depression that suggests a former channel. In the woods northeast of the site is the cabin owned by a White resident of Yakutat, Gil Sensmeier. SHAMAN’S GRAVE, KNIGHT ISLAND On the north bank of the stream, about 400 yards above its mouth and a quarter of a mile north of the site, a grave under an overhanging boulder was found just prior to our first visit in 1949. The following description is based upon the account given by the discoverer and members of his family and upon our own observations in 1949 and 1952. Although the skull had been removed by the finder, and a number of other bones were missing when we saw the grave, the vertebral column, the ribs, the bones of the legs and of the upper arms were still in position. The skeleton, that of an adult male, was lying on the back, head to the west. The body had evidently been placed in a coffin, but the latter was not interred. Although almost entirely disintegrated, fragments of wood indicated that the coffin had been originally 72 inches long and 30 to 32 inches wide, made of 1-inch planks put together with square-headed nails. With the skeleton were the remains of a blanket of mountain goat wool (pls. 18, 5; 19; see pp.187-192), and also a number of little blue and white glass beads (49-25-59). The latter are similar to those found with early historic (late 18th century or early 19th century) burials on Glacier Island in Prince William Sound (de Laguna, 1956, p. 211). A tiny iron pot, a painted shell, and disk-shaped shell beads are also said to have been found in the grave. Although one informant denied that the grave could have been that of a shaman because the beads were not the kind worn by shamans, it seems more probable that it was a medicine man who was buried here, for all ordinary persons were cremated at Yakutat until after the mission was founded in 1888. One informant believed that the grave was that of Daxodzu, sister of the principal K’ackqwan chief, a female shaman who foretold the arrival of the Russians. Another said that it was the grave of a Tlaxayik-Teqwedi shaman, uncle of the man who led the successful attack on the Russian fort. One reason for this attack was because the Russians accused the natives of having stolen the nails used to make this coffin. It is perhaps significant that both of these traditions associate the burial with a 36 BUREAU OF AMERICAN ETHNOLOGY [Buin. 192 shaman who died during the Russian occupation. There is, however, no reason to connect this late 18th- or early 19th-century burial with the occupation of Old Town, for shamans were often taken a long way from the nearest inhabited village for burial. EXCAVATION AND MAPPING TECHNIQUES In 1949 a rough sketch map was made of the site and some of the house pits were measured. Test excavations in Mounds A and B and in House Pit 1 indicated that the site would repay intensive investigation. Before beginning excavation in 1952, the entire site was laid out in 5-foot grids or squares from an arbitrary point (datum A) beyond the southwest corner of the occupied area (map 6). All horizontal distances were measured north and east of datum A, additional datum points were established for convenience in mapping, and compass bearings on prominent landmarks were taken from datum B. Elevations were determined for the corners of the 5-foot squares, measuring from mean lower low water, and these were translated into a contour map of Mound B (map 7). Most of the intensive excavations were concentrated in Mound B and the adjacent House Pit 1, because this mound appeared to be the largest and deepest midden deposit on the site, and because it was hoped that the house pit would yield information on house con- struction. In addition, tests were made in Mounds A, C, and D, with one end of a trench running into House Pit 7. Small test holes were made at various places in the site to determine the extent and composition of the deposits. These excavations explored a number of pits, visible on the surface or buried beneath it, and also uncovered the remains of a storage house, and of two additional houses, 8 and 9, the existence of which had not previously been suspected. With few exceptions, profiles of the midden deposits were drawn for the walls of each square ex- cavated (see figs. 5, 6, and 9). Identifiable animal bones and samples of shell were kept for each 6-inch level of each square excavated. Most of the shells and un- modified bones were identified and recorded in the field, and then discarded. Those which could not readily be identified were retained for further study in the laboratory. A more careful analysis was made of faunal remains from certain test areas in each of the trash mounds (pp. 77-84). STRATIGRAPHY OF THE TRASH MOUNDS Figure 1 is a schematic profile of a typical portion of Mound B, in which the following squares are shown from left to right: 48-57 693-818 O—64 (Face p. 36) 6! “462 63 64 65 66 67 68 60 99 T ose eis 58 57 56 55 54 53 / \ | SURFACE PITH | \ \ / \ oe 5| 50 49 48 47 46 45 61 62 63 64 65 66 67 68 d F. McGeein. 693-818 O—64 (Face p. 36) 42 | 43 | 44] 45 | 46 | 47 | 48 | 49 ALL. Wf Cay WO SURFACE PIT 6 ( pele > NN = \ ~ . oy Pa 53 ~@ |, a 2 peas 51 f ¢ - 166 170 i172 1 180 P 4 1e2 1186 164 #160 i768 | 1 ~ S 182 Dotum 0 . 5 Oj eter i708" es ‘ si STORAGE HOUSE = | HOUSE 8 = 4 “ ! 4 \ pe 49 7 a ere 8s | surface pit 5s ! —"- oe " \ ! 180 184 186 182 80 48 \ ' 74 172 170 70 tz 74 «17% «178 \ / \ Pog eee tate eae N mog 47 ohne _ e \ 4 ’ \ ° 5 10 ‘ \ 46 ! \ Scale in teet | SURFACE PIT 4 | Contour intervol 0.2 9. ‘a \ / MOUND B 41 42 43 44 45 46 47 48 49 50 5! 52 53 54 55 56 57 58 59 60 6| 62 63 64 65 66 67 68 Mar 7.—Map of Mound B, Old Town, Knight Island. Surveyed by Francis A. Riddell and Donald F. McGeein. de Laguna] 1 i} } Ih, WL ) | yyy Ay 1111 Wi I UA ae M11) Miya i / t \ Mt liyp yd Hyp My yl aT tal, | yi aptly | | VW/iy TAAL ZAIN vit / atl Me yw NM H eC ‘Se o ie ST aerea ves" |” Sf ee: 4/1) 1h y/ “ty Melee ‘Test'.< B03 194 ud DU rat =) (= o a. as = > x rs) ° — x rs) ny fea) 693-818—64——_ e o A) 5 = > U Cc iS) w” (= o io = CT) Ao} Be! E > ao) c 5 w iS = } i ea) Shell midden [~—] Charcoal ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA (feet) Drawn by Donald F. McGeein. Bark LUMT Ficure 1.—Schematic profile of Mound B, Old Town. 37 38 BUREAU OF AMERICAN ETHNOLOGY [BuLn. 192 (south and west faces), 47-58 (south and west faces), and 46-58 (south face). Figure 5 is the actual profile along the south faces of Squares 61-51 to 47-51, as exposed in a trench which passed through the middle of House 8 and the southern edge of House Pit 1 (see map 7,a-e). Figure 6 shows both faces (ff? and g—g*) of Trench 53, which ran through House Pit 1 and House 9. Figure 9 gives profiles of Trench 33 through Mound C and into House Pit 7. The maximum depth of Mound B was 90 inches. However, in areas where no pits had been dug into the underlying sterile sand, the deposits were between 24 and 36 inches thick. The entire trash mound was covered with a dense layer of moss, turf, and plants, about 2 inches thick, which has not been recorded on the profiles. The only stratum generally distributed over the mound is the black midden layer with fire-cracked rocks (see Blk. RM below), which occurs unmediately under the turf. A similar black rocky layer was found in the upper part of Mounds A and D. In the deeper sections of Mound B, strata of shell midden are predominant. Most of the layers are of sandy midden, ranging in color from tan or light gray to brown and black, depending upon the amount of organic materials contained. Because of the complex nature of the stratigraphy, owing in part to the aboriginal digging and subsequent filling of pits, and to vari- ations in the thickness of the various layers, the depths below the surface at which artifacts were found do not necessarily indicate their relative ages. Although horizontal and vertical position within the square was recorded for each artifact found, a relative chronology for these specimens had to be based upon their association with upper (later) or with lower (earlier) strata of Mound B. In general, except when specimens were found at the bottom of deep pits, there was a tendency for artifacts to be concentrated in the upper layers. The following types of deposit were distinguished in excavating Mound B, and similar materials were found in the other mounds. Black rocky midden (Blk. RM) composed of many thermal-fractured rocks, charcoal fragments, and black stained sand. Larger rocks, 8 to 10 inches long, predominate. Mammal bone was scarce and so poorly preserved that it resembled wet, mushy cardboard. Black sandy midden (Blk. SM) is medium-fine beach sand, stained by decomposed organic material and considerable quantities of char- coal. This usually contains some very fine fragments of shell, and some fire-cracked rock. Gray sandy midden (@SM) is found in two distinct shades, light gray (Lt. GSM) and dark gray (Dk. GSM), but they are identical in composition. These consist almost exclusively of medium-fine beach sand, stained by organic material and charcoal, and often de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 39 contain small flakes of mussel shell. Mammal bones and fire-cracked rocks are relatively scarce. Brown sandy midden (Brn. SM) consists of beach sand, stained light brown. A darker phase of the same midden (Dk. Brn. SM) was also recognized. It contains occasional flecks of shell, small charcoal fragments, and a few fire-cracked rocks. Tan sandy midden (TSM) is light tan in color, and contains only an occasional speck of charcoal or shell. Shell midden (SM) of almost pure shell: mussel (Mytilus) by far the most common, with clam (Saridomus), cockle (Protothaca), and sea urchin (Strongylocentrotus purpuratus) also common. These layers contain a high percentage of animal bones, mostly well pre- served, and also a good deal of thermal-fractured rock and charcoal. Considerable amounts of rotted bark were also found, often at the bottom of the layer. The shell midden is distinctive in lacking sand. Orange-brown midden (OBM) is almost exclusively associated with burned wooden structures such as houses and storehouses. It is made up of minute fragments of calcined shell, with occasional flecks of charcoal mixed with ash and sand. Clean sand (CS) occurs in lenses or layers of unstained beach sand, ranging from light gray to light tan in color. It is similar to the sterile sand which underlies the midden deposits. The maximum depth of Mound A is 40 inches. It is capped by the same type of black rocky midden that covers Mound B. No predominantly shelly strata were encountered, and the stratigraphy is not as complex as that of Mound B. While an upper and a lower layer could be distinguished, there was no evidence to suggest that this indicated different periods of cultural importance. All of Mound A would appear to be of the same age as the upper part of Mound B. The maximum depth of Mound C is 32 inches. The deposit is not capped by black rocky midden, and there are no prominent shell layers, although occasional shell lenses occur. The second layer from the top is composed of black rocky midden, but this does not extend into the adjacent House Pit 7 (see fig. 9) except in one part of the fill. This layer contains a lower percentage of fire-cracked rocks than that overlying Mound B, and is much less extensive. Mound D has a maximum depth of 30 inches. There is some black rocky midden in the top levels, and it also contains shell strata, so that in most respects it resembles Mound B, although it lacks the intrusive pits so characteristic of the latter. Upper and lower levels could be distinguished in both Mounds C and D, but there appeared to be no cultural differences between them. ‘Together with the fill in House Pit 7, they probably represent the oldest part of the site. 40 BUREAU OF AMERICAN ETHNOLOGY [BULL. 192 On the whole, the trash mounds at Old Town are not unlike the Tlingit middens of the Angoon area in southeastern Alaska (de Laguna, 1960), or the larger middens explored by Drucker (1943) in British Columbia. They differ, however, from middens in the Pacific Eskimo-Aleut area in containing far more earth or sand, and far less shell. The proportion of artifacts in all middens of the Northwest Coast is also much lower than in sites within Chugach, Koniag, and Aleut territory. This last may be due in part to the ereater use of perishable wood for artifacts on the Northwest Coast and also to the fact that bone rots quickly in the more acid, less shelly deposits. Other differences may be due to cultural factors, such as greater reliance on fish and less on shellfish on the Northwest Coast, less carelessness in losing and discarding possessions (a trait for which the Eskimo are noted), or greater neatness in disposing of rubbish at village sites. Such care does not seem to have been taken at temporary camps or forts of the Tlingit (de Laguna, 1953, p. 55). Although our informants doubtless exaggerate the neatness of aborig- inal housekeeping, the custom of frequently replacing the sand or gravel on the floor around the fireplace, which had to be done in any case before a shamanistic seance, might account for the high proportion of sand in the middens. The concentration of fire-cracked rocks in the upper layers of the deposits, in contrast to their relative scarcity in the lower layers and in Mounds C and JD, is paralleled at sites on Prince William Sound, Cook Inlet, and Kodiak Island. This has been interpreted (de Laguna, 1934, p. 162; Hrdlitka, 1944, pp. 30, 133, 394; Heizer, 1956, pp. 23 f.; de Laguna, 1956, pp. 49, 266) as indicating the rela- tively late appearance among the Pacific Eskimo of the steam bath. This type of bath has a more limited distribution and so is presumably more recent than sweat bathing in heated air without steam. It also necessitates far greater use of hot rocks than does stone boiling of food. The steam bath seems to be older on Prince William Sound than on Cook Inlet and Kodiak Island, which suggests diffusion from the Chugach or the Northwest Coast Indians. No clear proof of this hypothesis has yet been established. Thus Drucker (1943) noted no great concentration of fire-cracked rocks in the Tsimshian and Kwakiutl middens he explored, although such rocks were common and appear to have been fairly evenly distributed through the middens. We do not know, however, if there were so many as to suggest the steam bath. Ethnologically, sweat houses on the Northwest Coast are confined to the Tsimshian, Haida, and Tlingit, although the steam bath was also taken by the northern Kwakiutl (Drucker, 1950, Traits 375 and 667). Fire-cracked rocks are common in Tlingit sites in the Angoon area. Although Birket- de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 4] Smith (Birket-Smith and de Laguna, 1938, pp. 369 f.) has raised the question whether the steam bath may not have been introduced by the Russians, the evidence would indicate that it is prehistoric, although probably not very ancient. The fire-cracked rocks in Mounds C and D would suggest that the sweat bath with steam was already known to the oldest inhabitants at Old Town, while House 8 in the lowest layers of Mound B appears to have been a bathhouse. However, the concentrations of fire- cracked rocks in the upper levels of Mounds A and B may indicate increased popularity of the steam bath in later times. minal ah Slee: wine He, bontho to) noael eictact Wert, CAL a t pansion. wt diy, si }' ‘eluate tid ON tre ‘ ee - wht igdihlessayre Iie, Fal bus, at rnc?) of tee wtidaplesr io babe.. silt od avousl “quale ante i ariacaegee AT aad, Beh arsaiviek heey malinh) abieae “veh in eeotte tl nce: ned. clarrerygekh shaoibat, capt $L, bine Ay etcntot4. to men 190 we ee Evi tlt satnh fj rad, shegha ait pa | “fy i ’ ot lye ’ as Ey) coe HOUSES AND CACHES By Francis A. RippELL and FrREDERICA DE LAGUNA HOUSE PITS There were seven large rectangular depressions at Old Town that seem to have been the remains of houses. In addition, three more houses were found and excavated: House 8 and the Storage House, buried under the debris of Mound B, and House 9 which had been erected inside House Pit 1. The houses seem to have been scattered over the site without reference to any regular plan or alinement, except that the ends, where presumably the doors were located, faced the beach. Except for the two largest pits, the dimensions of the others are similar to house pits at Nessudat and Diyaguna’st on Lost River, although the latter are on the whole deeper. House Pit 1 is one of the largest at the site (pl. 2, a), with maximum measurements of 50 feet by 50 feet, not including what appears to be the entranceway toward the southeast. It was dug when Mound B was about half its present height, and some of the excavated sand was thrown back onto the mound, covering the stratified deposits above House 8, a structure already abandoned and filled with midden when House Pit 1 was excavated. (See Tan Sandy Midden in the profile of Trench 51, fig. 5.) After the abandonment of House Pit 1 and the structure which it presumably contained, a much smaller building, House 9, was erected in the southeastern (front?) end. Still later, after House 9 had been destroyed by fire, a third structure was built over the ruins. This last house evidently contained a place for sweat bathing. These three buildings in House Pit 1 (see figs. 6-8) may have been the last permanent houses to be erected at the site. What is known about them is described in later sections. House Pit 2, about 50 feet southwest of House Pit 1, was about 30 feet long (northeast-southwest), 25 feet wide, and 2 feet deep. A small test excavation in the bottom revealed no trace of timbers or floor planks. These had probably rotted. House Pits 8 and 4, lying close to the present beach and from 220 to 230 feet southwest of House Pit 1, are completely overgrown by a stand of small spruce trees (the grove in the center of pl. 2, 6). House Pit 4 is about 30 feet long (northeast-southwest), 28 feet 43 44 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 wide, and from 1 to 2 feet deep. House Pit 3, a rather irregular depression on the southeast side of House Pit 4, is about 18 feet long, 15 feet wide, and is somewhat deeper. It may have been a store- house, smokehouse, or bathhouse associated with a dwelling in House Pit 4, but since no excavations were made in either pit, their functions are unknown. House Pit 5 lies between House Pits 1 and 4. It is about 20 feet long (northeast-southwest), 10 feet wide, and 2 feet deep. The small mound southwest of the house may be sand removed in ex- cavating the pit. Neither the house pit nor the mound was excavated. House Pit 6 is an indistinct depression about 240 feet west of House Pit 1. It is roughly 30 feet long (northeast-southwest), 18 feet wide, and less than 1 foot deep. The slight elevation around the pit may be earth removed from it and refuse from the house built in it. A test hole in the mound revealed only very shallow midden. House Pit 7 lies in the forest about 180 feet north of House Pit 1. It is a shallow, indistinct depression, about 50 feet square, bordered on the south and west by Mounds C and D. Test excavations made in Mound C and the house pit (fig. 9), indicate that the latter was dug through the lower levels of the mound. Mound D was abruptly cut off as if it had lain against the wall of the house, although no trace of timbers was found. Mound A probably represents the debris from House Pits 3, 4, and 5. Mound B presumably contains the rubbish from Houses 1, 2, and 9, and from House 8 within it, and probably from other houses not yet discovered. Mound C antedates in part the digging of House Pit 7, although the fill of that pit also forms the upper levels of the mound. Mound D seems to represent the trash from House 7. Although House Pits 1 and 7 are of approximately the same size, and are the largest at the site, the latter is probably the older. Thus, large trees are growing init and on Mound C. The artifacts removed from this area, including Mound D, were not very different in type from those found in other parts of the site, except that no specimens of iron were present, although pieces of iron were recovered from Mounds A and B, and from the houses associated with the latter. It is reasonable to assume that the rubbish in Mounds C and D and House Pit 7 came from that house or from other dwellings in the vicinity, rather than from houses nearer the beach, and that this material is the oldest recovered at Old Town. The Storage House and the sweat-bath house (House 8), described below, were probably later. They were contemporaneous and were the oldest buildings discovered in Mound B. Since both of these were excavated through the lowest layers of the mound, it is evident that these accumulations of rubbish must have been derived from de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 45 some dwelling which we did not discover. Both the Storage House and House 8 were destroyed by fire, probably at the same time, and it is tempting to suppose that this may have been set by a raiding party. The fill in both belongs to the lower levels of Mound B. At a considerably later period, when Mound B had accumulated over the ruins and was about half its present height, House Pit 1 was dug, presumably for a large house. Still later, after the destruc- tion or abandonment of the latter, a small dwelling (House 9) was erected at one end of the large pit. Finally, after this in turn was burned, another more substantial structure, which seems to have in- cluded a room or place for sweat bathing, was built in House Pit 1, over the remains of House 9. The fill in House Pit 1 and House 9 (and over the third structure) is apparently contemporaneous with the upper levels of the midden in Mound B, and probably also with all of the deposits forming Mound A. Unfortunately, the relative ages of the other structures at the site could not be determined, although there was nothing to suggest a long period of occupation of the site as a whole. House Pit 5 would indicate a structure about the size of House 9; House Pits 2, 4, and 6 were probably for larger buildings, but smaller than those for which House Pits 1 and 7 were presumably intended. House Pit 3 may have been a bathhouse like House 8. This evidence, as well as the house pits at Nessudat and Diyaguna’rt, suggest that in the late prehistoric and early historic period most Yakutat houses were in- tended for occupancy by not more than four families, that some much larger multifamily dwellings were built, as well as small houses suitable for only one or two couples and their children. SMALLER SURFACE AND SUBSURFACE PITS In addition to the larger house pits, a number of depressions, about 5 to 12 feet in diameter and from 6 to 18 inches deep, were scattered over the site (see map 7). These were called “Surface Pits,” only because they were visible before excavation, not because they had been originally dug from the present surface. Thus, Surface Pit 1 was excavated when the top of Mound B was about 2 feet below its present level; Surface Pit 6 when it was 1% feet lower; and Surface Pit 7 when it was about 1 foot lower. Some of these surface pits had evidently been dug through earlier pits that were already filled with rubbish, perhaps because this facilitated digging. In turn, they be- came partially filled with midden material, generally the black rocky midden, and also with some gray sandy midden, together with scat- tered lenses of shell, charcoal, and rocks. Unbroken strata over the tops of most of these surface pits indicate that these were older than the last period of occupation of the site. Some pits contained 46 BUREAU OF AMERICAN ETHNOLOGY [BULL. 192 materials belonging to both the lower and upper levels of Mound B. About 40 completely buried pits were also discovered (see fig. 6). These had been dug into the sterile sand at the bottom of Mound B, when the top of the mound was much lower than it is at present. Subsequently they became filled with trash, mostly from the lower levels, so that no surface indications of their presence were visible. They were therefore called ‘Subsurface Pits,” although they differ in no respects from the ‘‘Surface Pits.” Some pits were rectangular, ranging from boxlike holes, 3 by 1% feet, to larger structures over 12 feet long and 6 feet wide. There must have been some kind of support for the walls, probably a lining of planks or bark, since there was usually no sign that the sand or midden had slumped in. In most pits no trace of such lining re- mained, but in several cases planks were preserved by carbonization. Probably there was also a wooden superstructure over the pit. These structures seem to have been underground or partially underground caches, like the Storage House described below. Surface Pit 3 was marked by a depression 11 feet in diameter and 18 inches deep. When excavated, it proved to have been originally a rectangular pit, about 10 feet long and 8 feet wide. It contained some burned and unburned fragments of planking, probably the remains of the lining. There were also bowl-shaped and basin-shaped pits, generally cir- cular in plan, less often oval, and differing from each other chiefly in depth relative to the diameter. Both forms differed greatly in size, ranging from 16 inches to over 5 feet in diameter. A number of both kinds had bark linings. The fill ranged from relatively clean sand to layers of shell and rock with strata of pure shell and bone, apparently representing the refuse of individual meals. Some pits contained artifacts (see below); others did not. The functions of these pits are hard to determine. ‘Those covered with bark or wood may have been caches, or pits where food was buried to become slightly rotten, as required for some native recipes; others may have served as bath- houses. Uncovered pits may have been ovens for roasting food or for heating rocks; some may have been dug simply to hold refuse. Special mention should be made of Subsurface Pit 38 (see fig. 2), a bowl-shaped depression, 2 feet in diameter and 15 inches deep, under the south end of the Storage House. Since it may have been a cellar of the latter, it will be described with it (see pp. 48-51). There were also depressions or pits in the floor of House Pit 1 and House 9, and a box in the floor of House 8. Such features, as well as fireplaces and pits with sweat-bath rocks, will be discussed in the detailed descriptions of the houses with which they were associated. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 47 The articles recovered from the various surface and subsurface pits are listed below (field catalog Nos. are in parentheses). As may be gathered, most of these holes contained nothing other than midden fill. In each case, it is indicated whether the material belongs to the upper or lower levels of Mound B. Surface Pit 1 (upper levels): Broken head for sea otter harpoon arrow, pl. 13, f Chipped green chert (29) Surface Pit 3 (upper levels): Copper fragment (17) Surface Pit 6: Upper levels: Small woodworking tool, pl. 8, l Schist drill (?) (14) Rubbing tool, pl. 10, b Copper ulo blade, fig. 10, d 2 rectangular stone scrapers, fig. 11, c (and 83) Broken double-edged slate blade, fig. 11, g Green chert flake (92) Slate fragment (85) 3 hammerstones (75, 76, 77) Lower levels: 2 hammerstones (104, 122) 2 hammerstone-abraders, one with red paint (36, 103) Beaver tooth chisel, pl. 16, b Bone gorge (51) Surface Pit 8: Level unknown, probably lower: Bird bone awl, pl. 16, m Tooth pendant, pl. 17, g Lower levels: 2 small woodworking tools (114, 121) Broken lamp (138) Broken harpoon head, pl. 18, 7 Sea otter harpoon arrowhead, pl. 13, d Barbed bird bone point, pl. 15, k Bone gorge, fig. 18, b Tooth pendant, pl. 17, h Cut bone (214) Cut wood (213) Subsurface Pit 9 (lower levels): Whetstone (9) Subsurface Pit 11: Upper levels: Small woodworking tool (15) Lower levels: Fragment of bone point (36) Subsurface Pit 14, containing traces of fire (lower levels): Broken splitting adz (40) Small woodworking tool, pl. 7, 7 Broken barbed bone arrowhead, fig. 17, e Cut bone (48) 48 BUREAU OF AMERICAN ETHNOLOGY [BuLn. 192 Subsurface Pit 15—see Storage House (below) Subsurface Pit 23 (lower levels): Broken bird bone point (211) Subsurface Pit 24—see House 8 (below) Subsurface Pit 31 (lower levels): Tooth chisel, pl. 16, c Subsurface Pit 32 (lower levels): Drift iron adz or scraper blade, pl. 4, k Subsurface Pit 36 (small cache house) (lower levels) : Small woodworking tool, pl. 7, m 2 whetstones, pl. 10, e (and 377) Unbarbed bone arrowhead, pl. 15, v Bone shaft fragment (398) Cut wood (385) Subsurface Pit 37 (lower levels): Hammerstone-abrader, pl. 10, f Toy lamp (394) Subsurface Pit 39 (lower levels): Cut wood (692) THE STORAGE HOUSE One of the oldest plank-lined cache pits uncovered during the exca- vation of Mound B was a structure in the southwestern portion of the mound (see fig. 2). It was probably a storage house that had been burned down by a fire that started at its southern end and which con- sumed all but the floor planks and the lower ends of the wall planks. The Storage House (originally designated as ‘‘Subsurface Pit 15’’) was 7 feet 9 inches long and 4 feet 6 inches wide. It had been built in a pit, sunk about 18 inches into the sterile sand below the midden. The floor level was between 3% and 4% feet below the present uneven surface of the mound. Above the remains of the house were about 2 feet of undisturbed stratified deposit, consisting of black rocky midden, shell midden, and gray sandy midden belonging to the upper levels of Mound B. The fill inside the house, belonging to the lower levels, consisted of a fairly homogeneous deposit of stained sand, char- coal fragments including remains of the wall and roof, fire-cracked rocks, ash, and bits of charred bone. There were small lenses of clean sand in the fill, and about 3 inches above the floor planks was a thin layer of light-gray sand. Below the floor planks were 1 to 3 inches of midden that had probably sifted under and between the boards; at the south end this deposit deepened into the fill of Subsurface Pit 38. The latter contained brown midden, rotted bark, bits of charcoal, bone, shell, and fire-cracked rocks. This stratigraphy may suggest that the pit was older than the Storage House, but it may be simply a cellar. The walls of the house were of roughly split planks, 4 to 22 inches wide and about 1 inch thick, which with two exceptions were set verti- cally in the sand at the bottom of the pit to an average depth of 10 de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 49 as Cc PLAN OF STORAGE HOUSE ——— — : ExcavaTeD [8 ----- RECONSTRUCTION seesnevere OBSCURED FROM VIEW a be gd Floor Plank |} Floor Plank PLANK LASHING TECHNIQUE lp Bottom’ View —— eee) —>— CROSS SECTION Su bsureace Pit 38 PLAN VIEW Dark brown Sandy midden Srttl S149 Red- brown ashy sand fem coerce y.vie #) ae ae Ficure 2.—Storage House and Subsurface Pit 38. Drawn by Donald F. McGeein. inches. The entrance was at the south end where a horizontal plank (A), 40 inches long and 13 inches wide, formed a sill some 7 inches high. Two posts and the small vertical planks (B,B), 534 inches wide, 20 inches long, and % inch thick, formed the sides of the doorway. One of the wall planks (Q), 48 inches long and 15 inches wide, was set on edge. The three floor planks, 81 inches long and 11 to 25 inches wide, rested directly on the sandy bottom of the pit. They had probably been inserted after the walls were erected, since the westernmost plank overlapped the central one and had not been trimmed to fit the 50 BUREAU OF AMERICAN ETHNOLOGY [BuLn. 192 floor space. The eastern and central planks were fastened together at the ends near the doorway. Here a rectangular area had been chiseled into their lower surfaces, into which was fitted a flat wooden crossbar. The latter was secured by a lashing of split spruce root that passed between a pair of holes in each plank. This arrangement may have been intended to facilitate lifting the floor planks in order to reach the pit below. The structure was probably roofed, but of this we have no direct evidence because fire had destroyed all of the upper part of the house. The quantities of carbonized moss and grass, found between the floor planks, between these and the walls, and under fragments of charred wood fallen on the floor, suggest that the Storage House had been chinked at floor, walls, and roof (?), presumably to prevent the con- tents from freezing. Identification of this structure as a cache was based partly on the following artifacts found in it and in the pit below. In Subsurface Pit 38: Point of a large double-edged slate blade, another fragment of which was found in the lower levels of Mound B above the Storage House, fig. 14, b 2 barbed bone points for arrows (or leisters), fig. 17, b, / Bone barb for gaff or fish spear, fig. 18, g Below floor of Storage House: Bird bone tube (422) Bone shaft fragment (420) Wooden spatula, fig. 24, c Wooden blade, fig. 24, 6 On floor of Storage House: Ulo with wooden handle and copper blade, fig. 10, a Whetstone, pl. 10, ¢ Bear canine, pl. 16, z Wooden comb, fig. 20, c Wooden figurine, fig. 22 Fragments of 2 wooden boxes or dishes, fig. 23, a, b, b’ Wooden rod scarfed at both ends, fig. 16, d 2 spatulate wooden objects, fig. 24, a (and 416) Fragment of bidarka rib (?), fig. 24, e Fragments of twined grass or bark matting (418) Fragments of carbonized two-ply cord (195, 196, 254, 365) Band of strung ryegrass stems, fig. 23, d, d’ Calcined bone fragments (428) 2 fragments of cut wood (199, 284) Just above floor of Storage House: Stone ax, pl. 5, 7 Cobblestone anvil (392) Small stone lamp (412) Slate blade for arrow or knife, fig. 14, a (associated with basketry fragments, pl. 18, a) Broken barbed point for arrow (?) (288) de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 51 Bone shaft fragment (205) Fused bone (411) Slag (200) Copper fragment (287) 2 drift iron blade fragments, pl. 4, e (and 201) 5 seraps of drift iron (258) Barbed wooden spear point, fig. 16, a Wooden box fragment (413) Wooden pin, fig. 24, d 12 fragments of worked wood (181, 182, 224, 257, 285, 286, 364, 366, 378, 386, 415, 425) Fragment of two-ply cord (208) Knotted spruce root (?) (226) Fragments of twined basketry: some with false embroidery, pl. 18, a (and 232/233); some plain, pl. 18, a (and associated with slate blade, fig. 14, a, and with salmonberry seeds) Fill of Storage House (i.e., lower levels of Mound B) 2 sea otter harpoon arrowheads, pl. 13, c, e Broken barbed bone arrowhead, fig. 17, k 2 beaver tooth chisels (881, 409) Broken bone knife or scraper (302) Bone awl (?) (406) Fragment of bone shaft (407) Notched cobblestone (295) Cut bone (197) Bone figurine worn as pendant, fig. 20, a 11 pieces of worked wood (87, 120, 171, 173, 175, 176, 178, 180, 183, 186, 187) Fragment of two-ply cord (431) 2 teeth of wooden comb (?) (179, 184) The perishable materials—wood, cordage, and basketry—were preserved because they were charred, probably in the fire that de- stroyed the Storage House. The baskets may have been used to store food and other objects, or to gather berries, as is suggested by the association of salmonberry seeds with one group of undecorated basketry fragments. HOUSE 8 The burned remains of a small house (figs. 3 and 4) were discovered at the bottom of the midden near the southern edge of Mound B (see map 7 and fig. 5). The house (first called ‘Subsurface Pit 24’’) was almost 18 feet square, and had been built inside a pit about 20 feet square, dug into the sterile sand for a depth of 30 inches. The floor of this pit was level. Vertical wall planks were driven into the sand to a depth of 1 foot, leaving a space about 1 foot wide between the walls and the edges of the pit. Later this space was filled with sandy midden to brace the walls, and the floor planks were laid. Eventually the house was destroyed by a fire that evidently started inside it and consumed all but the floor and the walls to about 15 52 BUREAU OF AMERICAN ETHNOLOGY [Buny. 192 —==—WALL PLANKS Ficure 3.—Plan of House 8. Drawn by Donald F. McGeein. inches above the floor. These charred remains were eventually buried under midden deposits 2 to 4 feet deep. The stratigraphy of the mound (see X—Y on the profile of the south face of Trench 51, fig. 5) shows that the house had been built during the early history of this part of the site; the unbroken layers above the fill attest its relatively great age. WALLS The walls of House 8 were of planks split from straight-grained wood that was free of knots, probably spruce. They were not smoothed on the surfaces. They varied in width from 1 to 2 feet and in thickness from about % to over 1% inches. When the house was burned, the upper parts of the walls fell into the house; no wall fragments were found outside. The plan (fig. 3) shows that a number of wall planks were missing. Possibly these had been salvaged after the fire. The gap in the de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 693-818—64——5 Ficure 4.—Reconstruction of House 8. Drawn by Donald F. McGeein. 53 54. BUREAU OF AMERICAN ETHNOLOGY [Burn. 192 otherwise complete line of planks along the southwest wall suggests an entrance, although a very narrow one. However, since the floor was 2 to 2% feet below the surface, this gap would have allowed sand to spill into the house in the absence of some kind of sill or ramp, no evidence of which was found. Probably the wall planks in this area had simply been removed, and the doorway was a hole cut through the wall at or above ground level, with a step inside the entrance. ROOF In the center of the northwest wall was a post that had been burned down to about 19 inches above the floor. The bottom was 40 inches below floor level. The post was originally 10 inches in diameter, and had been placed in an oval hole, 20 by 16 inches and 42 inches deep, partly filled with rocks to anchor the post. Three split planks were set around it, making a small alcovelike bulge in the wall. About 5 inches of the post below floor level showed charring, indicating that it had been exposed to this depth. The side of the post facing outside the house was relatively unburned as compared to that on the inside. When found, the post leaned several degrees toward the center of the house, its lower end well outside the floor area. Although no other posts of the same type or size were found, this post probably supported one end of a ridgepole, the other end of which presumably rested on top of the large plank in the opposite wall (beyond the container of fire-cracked rocks). The roof was partly or completely covered with sheets of bark, carbonized remains of which were found on the floor. It is reasonable to assume that the roof was gabled. If the eaves were too low at the sides to permit entry, the doorway would have been at one of the gabled ends. FLOOR The floor of the house was entirely covered with planks except at the sunken box in the center and the portions filled with fire- cracked rocks, as indicated on the plan (fig. 3). There were from 29 to 33 such planks, ranging in thickness from \% to slightly over linch. Their sizes varied, the maximum lengths being 8 feet and the maximum widths 3 feet. Their upper surfaces and edges had all been smoothed. All were charred in the fire. SWEAT-BATH ROCKS A container of fire-cracked rocks in the middle of the southeast wall was not a hearth for cooking but seems to have been a receptacle for sweat-bath rocks. It was 5 feet long and 3 feet wide, made of several large stone slabs, about 2 inches thick and set on edge —aousi : A adh A nda = besa yale Stowe Lav ob) boos bamintenn woot. pen trod : heawb novel b | 51-51 | HOUSE PIT 1 | 48-5) | 47-5| l Soutn Face Ey Gravel, small pebbles, sand. Bea Clean, unstained sand Rock. NAN Unburned wood, GB Burned wood, charcoal, 5 Indistinct gone. Thin bark, black midden. ° ' 2 FT. tb 1 1 — HOUSE PIT 1 | ora f 56-5I | 5551 | 54-51 53-5 = _ (es — De BSM osm Gsm osm ~ ~— _ - - CO rng Se e -_ csm mn UD) Ts nt QD Hs =- RED-BROWN Rocky MIDDEN) eo ree = TTD asm ia 15m conn ‘cannowseeD con woop Post Hole? $ T E R \ L E S A N D Ficure 5.—Cross section of Trench 51, through House 8 and House Pit 1. Drawn by Donald F. McGeein. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 55 (indicated on the plan by dark shading). The southwest side of this bin was formed by a plank (A) that had fallen out. The rock deposit inside this container was about 10 inches thick, the lowest part consisting of small rocks, above which was a gravel- like layer of fine rock fragments, sand, and occasional bits of charcoal and ash; on top were larger rocks, averaging 6 inches in diameter. These were mostly rounded sandstone cobbles, burned a dull red brown. The reddish gravel below had probably resulted from fracturing the heated stones when water was poured on them to produce steam. At a depth of 9 inches below the bottom of the rocks in the container was another layer, about 2 feet long and 3 feet wide and 4 inches thick, made up of the same kind of fire-cracked sandstones. This layer was at the bottom of a pit almost 2 feet below the floor of the house, which apparently represents a period when the house was first occupied. Later, it was evidently decided that the pit was too deep, so it was filled with a homogeneous layer of brown sandy midden, on top of which the rock container was built (see the cross section in fig. 5). Fire-cracked rocks were found scattered over the floor of the house outside the bin. The latter had been built up some 8 to 10 inches above the floor level, but when the house burned, the slabs and boards forming the sides were displaced, allowing the rocks to spill out. Except for the unplanked area between the container ‘and the box in the center of the floor, this spillover is not indicated on the plan. BOX A small plank-lined pit, below the level of the floor, is in the center of the house. The upper edges of the plank lining were either flush with the floor planks or projected only an inch above them. The sides of the box were of small planks, set on edge with the grain running horizontally. These enclosed a space 3 feet long, 2 feet wide, and about 8 inches deep. The planks themselves were about 3 feet long, from 12 to 18 inches wide, and \% to % of an inch thick. They were not fastened together at the corners, but were simply held in place by four stakes, 1% inch in diameter and 40 inches long, set inside the box. These projected 2 inches above the sides, but may originally have been longer. The box was covered with short boards, not part of the regular flooring, and the bottom was lined, at least in part, with pieces of bark. Quantities of carbonized moss were found on or just below the bark. Above this was a layer of burned sand, ash, and minute shell fragments, all of a very uniform texture, and a number of fire- cracked rocks. On top of the sandy fill was a large angular boulder 56 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 which, surprisingly enough, showed no evidence of having been sub- jected to fire. While such stones are said to have been put on roofs to hold down planks or sheets of bark, this stone does not seem to have fallen through the roof or the cover of the box. Its significance is unknown. The upper 6 or so inches of fill in the box consisted of a mass of carbonized wood fragments, burned sand, bark, and some pieces of shell. Near the center of the southwest wall, in front of the post, was set a vertical board (B). It was 20 inches long, 20 inches wide, and 3 inches thick, and extended 18 inches below the floor. The 2 inches that projected above the floor was battered as if the plank had served as an anvil; the wood fibers were considerably ‘‘broomed.”’ The lower end was not sharpened, so it must have been inserted into a small pit, not driven into the ground. Although this plank evidently provided a working surface for some task involving pounding, the nature of this cannot be determined. House 8 would appear to have served primarily as a sweat-bath house, and this function was also suggested by the natives with whom the field data were discussed. Such bathhouses were sometimes used also as sleeping places. One might assume from the battered board (B) and from the artifacts found on and below the floor planks that the building had been used as a workshop. These objects include the hammerstones, whetstone, grinding slab, paint, lamps, scraps of iron, small woodworking tools, adzes, and fragments of worked wood and bone. The broken war club head and splitting adz, and to a lesser degree the harpoon head and barbed arrowheads, suggest that the workers were men, not women. However, the box in the center of the floor which contained so much moss resembles very closely that described by our informants as made in the women’s birth house. This birth house was a permanent structure, used by all the women of a large lineage house at childbirth and during menstruation. It was described as containing facilities for sweat bathing. The birth pit was said to have been as deep as the distance from the fingertips to the elbow, and was filled with soft moss almost to the top to receive the baby. The woman in labor squatted over the pit, grasping a vertical pole in front of her. No remains of such a pole, it should be noted, could be identified in House 8. Further- more, this house is considerably larger than the birth hut, which was supposed to be just big enough to accommodate the parturient, the midwife, and two assistants. It is possible that House 8 was originally a bathhouse and workshop, later converted by the women into a birth house. It could hardly have been entered by men after contamination by women in childbed. Why the box should have contained the boulder we cannot explain. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 57 In any case, after the house was burned, the pit was used as a trash dump. The stratigraphy indicates that House 8 was contemporary with the Storage House. Both belong to the early part of Mound B, and the fill in both consists of its lower levels. The following objects were found in House 8: Below the floor: 2 hammerstones, one with red paint (1005, 1009) Whetstone (665) Drift iron blade fragment, pl. 4, 7 Drift iron nail, found in a small bark-lined pit (651) Broken harpoon head, pl. 13, 1 Bird bone point (687) Wooden stake (680) Piece of skin (646) In container of bath rocks: Hammerstone (1008) In central box: 2 broken barbed bone arrowheads (?), fig. 17, c, d On the floor: Splitting adz, pl. 5, a Broken planing adz, pl. 6, a 2 small woodworking tools (982, 994) Hammerstone (989) Whetstone (1003) Grinding slab (628) Rock with red paint (619) Broken head for war club, pl. 5, ¢ Wooden plank (744) Just above the floor: Small woodworking tool (995) 2 hammerstones (983, 991) Piece of red ocher (903) Stone lamp, pl. 11, a, and fig. 12, a Broken lamp (1007) Curiously shaped limestone pebble (amulet?) (547) Section of cut whale rib (1014) Fragment of cut wood (1000) Fill of House 8 (i.e., lower levels of Mound B): 2 small woodworking tools (598, 694) Hammerstone (981) 2 abrading stones or whetstones (985, 986) Rubbed stone (956) Lump of red ocher (571) Cobblestone with red paint (952) Chipped slate knife(?) (666) Sea otter harpoon arrowhead, fig. 15, b Bird bone awl, pl. 16, 0 Bone peg (621) Large quartz crystal (146) Mica scrap (907) Copper pin, fig. 18, d 58 BUREAU OF AMERICAN ETHNOLOGY [Bunn 192 Copper dangler, fig. 19, z Lump of coal (637) 2 fragments of worked wood (177, 597) Carbonized cordage (629) 2 pieces of bark (614, 615) HOUSE SPER House Pit 1 in Mound B is roughly rectangular, measuring 50 feet in length. ‘The width is about 45 feet at the front(?) end toward the southeast, but widens to about 50 feet near the rear, and then abruptly narrows to form an alcove 40 feet wide and 7 feet long. The pit is surrounded by a low pile of earth that is broken only at the south- eastern end, where it is thrown up on each side to leave a sunken approach about 20 to 30 feet long. The maximum depth of the pit before excavation was about 3% feet below the top of these earthen walls, the deepest part being just inside the sunken entranceway, where the later pit for House 9 was dug (fig. 6). As already mentioned, House Pit 1 was excavated through the lower layers of Mound B, and some of the tan sandy midden which was dug out was thrown on top of the fill overlying House 8 (fig. 5). House Pit 1 and all its contents belong, therefore, to the later levels of Mound B. In 1949 a narrow test trench was made in the northern quarter of the house pit, from the side wall just forward of the rear alcove to the center (i.e., to Surface Pit 10 on map 7). Charred timbers revealed by this trench suggested that there had been a bench about 6 or 7 feet wide along the northeast sidewall, and that a pair of central beams supported a gable roof, as on the large Tlingit houses of historic times. The floor was about 14 inches below the present surface of the pit. It was assumed that the alcove represented a line of sleeping cubicles across the back of the house, although this was never ex- cavated. | ee ee Se ee ee 4 VWiviitile ne eR CRO Se ee ae, Serene See 2 Sta [Eek Fa = eee 1 1 ee ee ees =| Seed ee ee ee (2?) Canis familiaris (domestic dog) ------------ | Meee a WR | REN (1) 2 Numbers indicate the bones removed from the test areas. Numbers in parentheses indicate the probable number of individuals represented. X indicates that remains were found in the mound, but not in the test areas, and that no numerical count was recorded. [BuLL. 192 BUREAU OF AMERICAN ETHNOLOGY 80 1e}0L, oz | st | 6t | 62 | OF | oF | 8e| 9 | zt | ez] ce | es] Fe] oe) Tt | pe | WoL UA etna | =r anal lee bane | mc | (Eanes iaaere pa aka bea NG ae So eet a Seal Wel Op Peele pee ieSen|eo | Wr eee Ge Wee: (mers. uiceealeg Teale) BoaleSeelacts| Ola Ela Wee 18) (Le Wee le, | Ol ser ai OF BEG |e) [20 Or ra | Fo 1 | 8 [OL | 6 fat et | Bajcc shar Dealceee|Ceclgtavleve tes WO |so38) Seale Got. We leer Tape ees| Se [oe CUMS aT We alse Lee a i fa CCUM Lr om ieegiiterlaed| c eleaeiee iy lee ler | Tiel ah6 2s aoa fe al ee a (ee (We Cee eas T WA NA Zee NES Te Geel ere AV) Pie siete Go tw We he hts iy | ie Wie Wp WS 2a Ga ee 6 7 ee oa el ia | aia es Come | (011 nea eee |e | | =| (nea a | Iie Q|>| 2 8/8| s/s Bl ails tow | aa) Se ElElElElelElElEIFIFIE/ 2/2] é B|e |S Sota: |e fester fe ke BI@(E/P If /ElE|E| Es gl/ele|s < — ~ “4 - * = a} — aA dan | e/e/./ 2/8 Sula e ee gen (eee | Solleea |e BIBIB|/S® Il Sl2elelel/eleslelFiale/4/2 ae B)/aeis|s|s| > @l Sle] a st Ce ie? |ftes |} ter + a | 9, 1 fo} ° (=) oO .. =} i=} B — — &|e ls g Ere Se lees 2) 8/5 s ct 5 ([B9s) DI0YT 4jJoy ‘Snowman yy | SI) GZ | 42 | 9T (4 g T Il | 6 F OL | ZI | 6 (6 g z € 17 i! T iE S\eea (4 Calera ated | cae I g|3i\s ko] re B Ee | & EI |E tex — Co &|S] 5 o| | o 439] ‘olqIpuB yy ay te NG ceee PS it T Lie ie ae ae een | noua ae aay || [i pel 4 ats (i a € 1 |----|----|---- SL pe PS a a S|/S/8| 2% 2 2 2 = BIB] EIS PS | Pie a - — ei |) eee | ee | YS | | Pale 3 5 5 Q o 5 qqsit ‘Aloypne ‘eying {ol “Aroztpne ‘eying Ti ee ee || eae ees Ce [210.L 98-08 777} 08-42 77"! $E-8T SA Seat ---"! 21-9 Tes X0-0 See lineman eae q punoyy aaa ates: cl | RCSL; Se ee SE (210. 08-¥Z --"-| ¥Z-81 ---"| 81-21 ----! 21-9 arc O-Ol | ace eae talent be Y punoyy m z a B ie} B s+ | soqour ul 8018 4SAT, qydeq spain 489) ay? ur ‘yzdap pun adfij fig ‘sauog Jowwnwm payfrpowun fo iauanba1gq—Z AAV J, de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA ™ sO tb tet Ft] WL MimONwMoe | © a ie oie cl ' ’ - 1 ' ' Py CaN er eer ae Pe erat ae) aT LS Se a a a CR, ae eae Te ee eee ae le, | Pe se ae Is pins a ap 4 = (G3) Hie hate | = a a i onl fae piel ie a) WeRe ree, po | | | | C— ho We ie a Te iT N me | N is 4) eee ey eel a | ‘ add a Ae Va eT) ' Borate, 1s ee Sete te nee reals Prr1rt nia hot 6 x ey s Bia 0 Wee nei nae se, ss Pano ALS) | | Ade i wes oo sk 8 ae eT ee Oe Lee I ee ae on NOxnmS rt tN OD OD Ht mn IMOUTOI@ Sa5e eso eee AND CUES occ SSE 8 pubes Deter | Be aad eens. VU OUNO AD serena ta oe AMUN SS $53 58 ode SP ae S| eS BItBstO tales ns ee ee eee 82 TABLE 2.—Frequency of unmodified mammal bones, by type and depth, in the test areas—Continued Oreamnos (mountain goat) Enhydra (sea otter) Phocoena (porpoise) BUREAU OF AMERICAN ETHNOLOGY [BULL. 192 Ts i ee on) IHR ON 1 Pt te oe) PUR ea ee mora | ilile| eee ee Tedre9 | ABLE) RETTIG RERT ANS) =| = somo | Titi ii ibiled ities iiy ei iiil ils sshd UL re rewmodmen | iP EE Tei die [sys a ‘era | SEETIEL Egat GA ETETAEIEIEEEEERIEG waar | ELLE iim ie biel E EE eal ale suogayourmmommr | i Sift ile PEELE EET PEPE] is woven | PEEL PEE EESEEEEEEEEET EEE] ile seppur‘smpor | LET EEL EP EEEEEET PEELE EET le Puede vorvempen | PEPE ey Ee wee ee vetoes | LU woroment | PETE PEE TEEPE EET EEE il oo [epee PPP i re ted a ae a a a a fe ree | Te wammor | PELE] Pit EET PEEL EEE EET EEG] i suogeyeurmounr | YT HEI EEE EE EET de Files vere CE Te vervema | eT EEE a i es wer‘ensommat | TPE EEE TEE wrvsrpen | PP YP EEE EEE ae veepmrvomen | HEE EEE aT le sa ‘omeye || ist | aL REEEEL HEEL EEE ile Naa alla quomae | EE HEE Es Tey pad ills 81Q9}19 A, line = [= ]=ease= [|| j ida | &| | conaeaiaes [siz ws “arqrpaey, | FHI Maat CET EG szoypne ‘sing | | : Sr ube bee: ‘{s| als ae quomes) Tras | eve ec eH EEEE AT be ded Bad ana nae era | ae | Q a= Sin MmrN oO ' mrtN OD OD SH mein ' z i 3 43 7 A e: 3 | 5 & 3 & ae & a 3 3 3 3 = = = = 2 Includes a porpoise skull, in Square 46-58, 12-24 inches. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 83 TABLE 2.—Frequency of unmodified mammal bones, by type and depth, in the test areas—Continued Lutra Castor Marmota| Rodentia Canis Whale (land otter) (beaver) (marmot) (dog) Depth s rey Test Bee. : 2¢| ® 4s area inches | & = z oo rs = "op = Q _ 2, Ss , a S 3 - 2) ro) J.| 8 |3a | 3a | 8a 3 F ra ae =| as a Abe | eb | oo =) 6 S| 3 # ct n ~ 8 wo Ge 8 fo} (3) Sone 9 = ° = ion) n a om & Fy Mound A---__ DSH lecte S| GseSse| bases] ESSs3c| betel ae Se Se Rell Se Se eet |e nl | ee GELY. | eece ce) eeose| bees) beac | bee el inst ice Se) Seer eae el See] eo! Pesce ne TEATS Boal fee | A a Se eS eee ee ee Ce ee |e Pe eee eee SS Te ees eee eee eel Repeal eases) pene On Seer r= 1 oS ees ere SED) linen eee Bees eee a Se ee at Le oe Sl ee Se Eee 8 [pases (See ae ART) rE lepers eco etaheee een ~2asce) bee eal Basecel lst ae Rese Si Me ta em eee 1 De eee Mound B---.- 0-6 10) eee ee ae | ee eee ee ee om eee een. ease koe (1 ot cere [ea Sept | aad =8) Pt nme POS Scone Oa | Set ee ee ee eens iPS eee 1 1 1 1 TDyal | 2,2 Mi a = el UO |S CSRS) | a a 7 Na ee LE eek fs a fee el |b a (th ne ee ee (SE eS gee (eter (een eee TESTA ba Se nat |p Sa fas SS | i ed S| cll ee Ee ee hl Ee 2 eee ee aS S0-86) [es a a8 ee aE ee ek | es eee Se est acces eeens- cans Motaltene eases oto 1 1 1 1 1 11S es hE Eel (een ee On ee Mound C-.._- St om be ee] | pepe [eet | eee eS Pgs ee js eee eee ee Meme (Rae | Ae Ci | Jace | [i hal eat ee | Scene A ae (needs nes ot | Sa army |S ee ee ee IA-1S eeen eee eet Oak Se oo | SE See Se eS ese as be eeee |e SSeS ee. AeA eee ee eeee ee eeseoe |aeceas |peaaws | loco e ore [beeweneaeeeess | sescee|onoce (San cecercs VAST ate tees al Label EERE SO |p Fk SN) aa (Pe aes Ses ee ee ee ee eee Se Li | kee |= cereal [apa ateal a baa bet Ge al PRES OS 2) SS Se ae Se ee ee (ae ANGRGS a es | se id Rae tf Me ea pane aoa be Bie Dba ei Sa jase) Lh Ot Ye (ay (ae | ae Mound D.-._- VR |e ll [atest |e ee al | ies] bp tier ol se et al he eee fe Bl eh el Hea oe | SPAN aie aaa pepe pl ES fe BaD ee ee a) ee oS ae ee ee ee See UP pee ee fee ee as | ea al | ec | ane Ra a YE RS a | Fe ee Be ee ee ee | eee JAS |e eee ee eee ieee es ea | ey iD) eee Se ee 1 SATE ANT oh Alf ge EBS Ne oe 2) | pe MD fea AN | Le aE Pal | Delite | Nee TCS 19 [eee eo Se ES | ee S| 29 ee | ee eee ee ee eee | eee 3 te Oi fi le Fe | Lal mel la at he” bel | atta ae IR A Rh Sil cr | ee RS le Te tS NS | eee, | ee | Se See | ee (ere ae oe) a eee Ne eee ee ANTM La 2 | [ER S| | ae (Ce ee le ee (ee ee oe ee ee ae i OY ee Ie 1 Bitentotal.|-—.---=— 1 1 1 1 1 ul 1 | 1 I Total number of animal bones from test areas Test area Depth in | Number of Test area Depthin | Number of inches bones inches bones INIGINION AS eo 0-6 Oar vlound) ©. a eee ene 0-6 3 6-12 57 6-12 7 12-18 57 12-18 3 18-24 8 he ones 24-30 3 24-30 i - 30-36) |Socsecs5-2>2 NTF a ne ee a [oe Se ee 148 36-42 1 42-48 1 NUTT O Eo 3 ia 0-6 43 6-12 95 TROL AI ee ee eee oes eo meee 16 12-18 344 18-24 DOS Noung = ee meee eee a 0-6 8 24-30 55 6-12 10 30-36 8 12-18 8 18-24 17 SU eee ie BRE hee ee 748 24-30 5 TOGA Eee eee eee Semcon 48 [BuLL. 192 BUREAU OF AMERICAN ETHNOLOGY TABLE 3.—Occurrence of molluscan, crustacean, and echinoid remains in the test areas 84 (aytiesieu yoddng) it > it srypdnd sapivbsvyy HOA Wo? a H Qedunt perms) lee | Dyad vanwuoy teal isso { (aqyom vag) snpoindind snjzosjuaoojibuons XxX xXx { 1 ' 1 ' 1 1 1 ' ' 1 1 t ! ' 1 ' ' 1 1 ' ' ' 1 1 ‘ ' ' ' 1 ' ' 1 1 1 1 1 1 1 1 1 1 1 (ejoeuseg) bala tt ‘ds snumpog biyt i x (wavy Jade3 oyoeg) 1 27)D)7NU SnasnY}0Z LY IS 1 1 || XxX (MIRIO WOR SULYSeAA YOOuTS) snaquvb1b snuopizng xX | XXXXKXKX | (ure[d Yoous 41] Ope) DIUIULDJS DIDYIOIOL | | XXXXX (apPq{ooo yayse) YPORNU WNipsD IUD => s ~— ! 1 ' 1 ' ! Synpa snqiyy (oP{UTA 20d) “ds swy 7, | KXXKXKKXK | XXKKKXK | xXXX (404113 WOZ3I0) | asuauobaso wnwwonqobipy xX 1X a CNA tate oD © BAS Toads Sbadus J lal q Seen oe ior) mei No Test area | | | | | | | (jassnur ont SSooese a / d Ficure 23 (For legend, see opposite page.) de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA Lea technically advanced type, requiring straight grained wood and skill. Cedar is ideal for this purpose and was always used for such boxes, except that Chilkat (Drucker, 1950, p. 257) and Yakutat informants reported the use of carefully selected spruce. ‘Tlingit, Haida, Tsim- shian, and Kwakiutl boxes, especially those for holding liquids, had flanged, morticed bottoms (Drucker, 1950, Trait 449). Similar boxes were made at Yakutat in recent times, and one of the specimens from Old Town has this type of bottom, even though it is oval. Very large boxes on the Northwest Coast were sometimes made with two or four separate pieces for the sides (Niblack, 1890, p. 319; Birket-Smith and de Laguna, 1938, pp. 413 f.). This type was easier to make than that with a single bent piece for the side, since it did not require skill in bending wood, nor such fine planks. Presumably, such boxes were also made at Yakutat, since they were by the Copper River Eyak and the Chugach, and are quite old among the Eskimo (Birket-Smith, 1953, pp. 57 ff., 202, pl. 26, 6; Birket-Smith and de Laguna, 1938, pp. 78 f., 413 f.). Pegging or nailing was employed to fasten the sides of these boxes together, but the method typical of the northern and central Northwest Coast is to sew the parts together. The Coast Salish used both methods (Barnett, 1939, Traits 359-361, 365-367). Both are represented by the Old Town specimens and were reported by our Yakutat informants, although we gathered that pegging was more common. The cylindrical vessel with a flat round or oval bottom, like that from Old Town III (fig. 23, c), may have been a still older type than any form of square-cornered box, for it is less difficult to make. The sides of such vessels are usually of pliable bark or baleen, or of thin wood which is easily bent after soaking in warm water. Cylindrical pails of this kind were made by the northern Alaskan Eskimo at least as far back as Birnirk and Old Bering Sea times, and have been re- ported ethnologically from the Pacific Eskimo, Aleut, Eyak, Tanaina, Tena, and many interior Athabaskan groups (Birket-Smith and de Laguna, 1938, pp. 413 ff.; Birket-Smith, 1953, p. 202). The ends of the side piece on such vessels are commonly joined by sewing, which may explain why the northern and central Northwest Coast Indians have adopted this essentially interior bark-working technique and adapted it to the manufacture of wooden pails and chests. Ficure 23.—Box fragments and band of grass. Drawn by Donald F. McGeein. a, Side of small wooden box or dish (restored), from floor of Storage House, Old Town II (No. 367); b, b’, fragment of a similar box or dish, same provenience (No. 367); c, bottom of wooden box or vessel with copper nails, from just above floor of House 9, Old Town III (No. 974); d, band of ryegrass stems, strung together, from floor of Storage House, Old Town II (No. 283); d’, diagram to show method of stringing d. 178 BUREAU OF AMERICAN ETHNOLOGY [Buiu. 192 BARK Two small fragments of birch (?) bark, one of which had been folded twice, were found in House 8, Old Town II. No birch grows in the Yakutat area, but birchbark baskets full of soapberries were sometimes traded from the Southern Tutchone at the head of the Alsek River. Both berries and baskets were considered a novelty at Yakutat. Two pieces of spruce (?) bark were found on or near the floor of House 9 in Old Town III. They may have been part of the roof, or of some big container such as the large sheets of spruce bark which were set on posts above the fire as pans in which to cook berries for storage. CORDAGE, BASKETS, AND TEXTILES CORDAGE There are seven carbonized pieces of two-ply Z-twist cords, from 2 to 8 mm. in diameter, probably made of spruce root. Six were on the floor of the Storage House, as was a knotted length of spruce root (?), and the seventh was from the fill of House 8. Examples of two-ply S-twist sinew (?) thread or string, about 0.5 mm. in diameter, were preserved on the stem of an arrowhead from Old Town II (pl. 14, 6) and on two copper hooks from Old Town IIT (fig. 18, 7). Informants mentioned heavier ropes or cords of spruce roots, ropes of untanned seal and sea lion hide, fishing lines of kelp, and braided square sennit cords of porpoise sinew for bowstrings and harpoon lines. Thread of porpoise sinew was used for sewing garments. TWINED BASKETS Carbonized fragments of several fine, twined spruce root baskets (pl. 18, a) were found on or just above the floor of the Storage House in Old Town II. The direction of twining is downward from left to right; the fragments vary from about 6 warps and 8 wefts per square centimeter to 9 warps and 10 or 11 wefts. Salmonberry seeds were found with one of the coarser baskets, suggesting that it had been used to gather or store berries. The finer baskets were decorated with false embroidery, a technique in which the northern Tlingit excel (Mason, 1904, pp. 308 ff.). ‘The Yakutat have always held first place in basketry,”’ and legend credits them with the origin of this art (Emmons, 1903, pp. 229-231). Yakutat women claim that their baskets were superior because they held the weft strands tight with their teeth while weaving, whereas other Tlingit women used only their fingers (sic). Baskets of a variety of shapes and weaves were formerly made. Those designed to hold liquids were soaked and then rubbed on the de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 179 inside while still damp with a brown bear canine, which polished and flattened the strands, making the weave watertight. A bear canine (pl. 16, 2), found on the floor of the Storage House, was probably used for this purpose, for there is a worn facet in the enamel and both edges of the root have been cut flat. There is nothing about the archeological basket fragments to distinguish them from modern baskets of the Yakutat and other northern Tlingit (cf. Niblack, 1890, pl. xxxv1). The Haida make similar baskets, except that among them overlaid designs are said to be recent. Decorated twined baskets are not found elsewhere on the Northwest Coast until we reach the Makah (and the Nootka who have recently copied them), but their baskets are rather different from those of the Tlingit. Twined, decorated baskets are, of course, made by many southern Northwest Coast tribes, including the Coast Salish and northwestern California groups. In southwestern Alaska, the Copper River Eyak, Chugach, and Tanaina also made baskets of Tlingit type, similarly decorated, and the Koniag made some twined baskets. The finest work was done by the Aleut, although their baskets were of grass and different in design. ‘Twined baskets were also made sometimes north of Bering Strait (Birket-Smith and de Laguna, 1938, pl. 14; Birket-Smith, 1953, fig. 28; Osgood, 1937, pl. 10, A-C). This type of basketry has a circum-Pacific distribution, while coiled baskets occur in areas beyond its limits. Unfortunately, archeological specimens are too seldom preserved to give clear evidence of the antiquity or sequence of types in any given area. However, twined basketry with false embroidery is found from southern Oregon to the Columbia, with radiocarbon dates indicating an age of 9,000 years. Except for the materials, it resembles Tlingit work very closely, and Cressman (1960, p. 73) reports that he saw the same kind of basketry in Heizer’s collections from Kodiak. Fragments of rather coarse, open twined baskets, as well as of coiled baskets, were found in the Platinum Village site in Bristol Bay (Larsen, 1950, fig. 57, 1-4). ‘This site seems to be older than others in the area with pottery, and the material from it shows similarities to both the Near Ipiutak of Point Hope and the lower levels of Kachemak Bay. All available evidence, therefore, suggests great antiquity for twined baskets on the Northwest Coast. MATTING A fragment of twined grass (or shredded bark?) matting was found on the floor of the Storage House. Our informants had heard that shamans used such mats in their seances, but could not describe them. At a still earlier period, mats were undoubtedly used for ordinary domestic purposes. The weft elements of this mat are about 180 BUREAU OF AMERICAN ETHNOLOGY [Bunn. 192 1 cm. apart and are twined upward from left to right, so that it is identical with specimens from late prehistoric burials of the Chugach, Aleut, and Tena, as well as with the sleeping and kayak mats of the modern Kuskokwim-Kotzebue Eskimo (de Laguna, 1947, pls. x1x, xx; 1956, pls. 52 and 53; Oswalt, 1952, pl. 18, A, B). The undecorated grass mats on the walls and floor of the Eyak sleeping room may well have been the same (Birket-Smith and de Laguna, 1938, p. 34). Coarse twined grass mats were found on the floor and sleeping plat- form of a burned house at the Platinum Village site in Bristol Bay (Larsen, 1950, p. 184). Twined cedar-bark mats are common on the Northwest Coast (Drucker, 1950, Traits 718, 733, etc.). In general, twined mats, used for bedding, seats, and for shrouds, have a very wide distribution in both the Old and New Worlds, and the oldest direction of twining seems to have been up from left to right (de Laguna, 1947, pp. 217 ff., 272). BLANKETS The geometric patterned woolen blanket, known at Yakutat in early historic (and late prehistoric?) times (see pp. 171 and 196), is an obvious link with the Tlingit of southeastern Alaska. According to our informants, a number of Yakutat women knew how to make Chilkat blankets of conventional, modern type with representative crest designs, since they had at one time been “married into Chilkat.”’ There was no specific reference to pattern boards which they may have used; certainly there is none at Yakutat now. The knowledge of weaving blankets may have extended even farther northwest than Yakutat. Thus, Captain Cook (1785, vol. 2, p. 368) who visited Prince William Sound in 1778 reported ‘one or two woolen garments like those of Nootka.” Strange (1928, pp. 42 f.) also said that the Chugach in 1786 had thick warm woolen blankets, but valued them too highly to sell any. He bought the skin of the animal from which it was obtained, and described it as very similar to a sheepskin. In a report by Potap Zaikov, who explored Prince William Sound in 1783 (Tikhmenev, 1863, App., p. 6 7), the Russians observed ‘‘. . . a blanket made of white wool, similar to sheep’s wool, plaited and fringed. The blanket was ornamented with yellow and coffee color.’”? Our Chugach informants, however, believed that it was not until after the arrival of the Russians that they themselves learned how to weave goat wool blankets (Birket-Smith, 1953, p. 64). In 1884, Abercrombie noted that the Eyak slept under woven goat wool blankets about a yard wide and 5 feet long, but our informants 2 Translation by Ivan Petroff. Permission to quote this passage has been given by the Director of the Bancroft Library, University of California at Berkeley. de Laguna} ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 181 denied that the Eyak knew how to make them (Birket-Smith and de Laguna, 1938, pp. 43, 70). The lack of specific mention of designs on Chugach and Eyak blankets and the description in Tikhmenev suggest that they were either plain or had only simple geometric patterns, like that of the Yakutat blanket, for surely anything resembling the elaborate crest patterns of modern Chilkat blankets would have been remarked. On the basis of her detailed study (see pp. 187 ff.), Carolyn Osborne recognizes a northern center for woolen blankets with geometric designs. MISCELLANEOUS WORKED BONE, STONE, AND WOOD WORKED BONE A flat piece of bone from Old Town ITI, (7.5) cm. long and 1.9 by 1 cm., may have been the handle of a tool. There is a shallow cut across one face near the unbroken rounded end. Another possible handle is a flat piece of seal (?) bone, (9.2) cm. long, 2.3 by 1.2 cm., from Old Town II. It is ground on several sides; one end is rounded and smooth, the other broken. A somewhat similar worn bone frag- “ment from Old Town IT, (4.7) cm. long, has a hole through the rounded end. Also from Old Town III is a cut section of animal rib, 7.8 em. long, with one rounded end; the other is damaged. This also may have been a handle. A flat piece of whalebone from Old Town III has been whittled into a disk, 1.8 cm. in diameter and 1.2 cm. thick (for a top?). In addition, 28 nondescript pieces of bone, most of which appear to be workshop debris, show that bones were split by cutting grooves in one or both surfaces and that bone was also shaped by adzing, whittling, and grinding. Two pieces from Old Town II are the articular ends cut from animal long bones. The distribution of these pieces is: 1 from Canoe Pass, 15 from Old Town III, 15 from Old Town II, 2 from Old Town I, and 3 from Old Town, level unknown. It should be noted that these worked bones include fragments of whale or porpoise bone, mostly from Old Town III, of large mammal (bear? and mountain goat), of small mammals, and of birds. CHERT CORES, NODULES, AND CHIPS Three cores of green chert, two about the size of the fist, from which flakes were struck by direct percussion, are from Old Town III, and another core is from an unknown level. Three nodules of green chert, about 3 by 4 cm., struck from larger pieces, are from Old Town ITI. 693-818—64——_13 [BuLyu. 192 BUREAU OF AMERICAN ETHNOLOGY 182 Ficure 24 (For legend, see opposite page.) de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 183 Some 45 chips of chert, mostly green in color, range in length from 1 to 6.5 em. A few might have been used for scrapers or knives, but most seem to be the debris of manufacture. We do not know what was being made, however, since the few finished artifacts of flaked stone include no specimens of chert. The proveniences of these flakes are: 4 from Old Town III, 2 from Old Town II, 1 from an unknown level, and all the rest from Old Town I, including the upper and lower levels of Mounds C and D and of the fill in House Pit 7. Thus, of 52 pieces of worked chert only 9 came from Old Town III, 2 from Old Town II, and 2 from an unknown level in Mound B, showing that the flaking of chert, whatever the purpose, had become virtually obsolete by the period represented by Mounds A and B. WORKED QUARTZ, GREENSTONE, AND SLATE From Old Town III is a flake of quartz, 4.8 em. long, which could have been used as a knife or scraper. In addition to the fragments of adz blades, already described (p. 95), there are 6 flakes of greenstone. One is from Diyaguna’st, one from Old Town III, three from Old Town I, and 1 piece from Old Town, level unknown. The last could have been used as a scraper. Of chipped and ground slate, a piece from Little Fort Island may have been intended for a knife blade. The proveniences of other fragments are: one from Old Town III, one from Old Town II, and three from Old Town I. MISCELLANEOUS WOODEN OBJECTS In addition to the artifacts described elsewhere, many pieces of carbonized wood were found, chiefly in the burned houses. These are mostly fragments of firewood, workshop chips or splinters, a few of which show the marks of adz and knife but only 10 appear to have been purposely shaped. Among the latter, there are two slender wooden pins, well made and polished. One (fig. 24, d), pointed at both ends and 17.5 cm. long, is from the Storage House in Old Town II; the other (fig. 15, c), from Old Town III, is now broken, but was evidently pointed at one end. A more roughly made pin, originally over 30 cm. in length, came from House 8 in Old Town II. Figure 24.—Wooden objects. Drawn by Donald F. McGeein. a, Spatulate object, from floor of Storage House, Old Town II (No. 388); 6, wooden blade, from below floor of Storage House, Old Town II (No. 423); ¢c, spatulate object, from below floor of Storage House, Old Town II (No. 429); d, wooden pin, from just above floor of Storage House, Old Town II (No. 256); ¢, fragment of bidarka rib (?), from floor of Storage House, Old Town II (No. 399); f, cut branch, from floor of House 9, Old Town III (No. 658). 184 BUREAU OF AMERICAN ETHNOLOGY [Buu. 192 A stick (fig. 16, d), 42.5 cm. long, has been trimmed smooth, and both rounded ends are beveled from one side. It may have been used to spread fish for drying or cooking. There are four spatulate wooden fragments, also from the Storage House. One (fig. 24, c), now (14.4) cm. long, has a series of finely incised chevrons on the flat blade near what appears to be the handle (on the side not illustrated). This may have been a paddle for beating up soapberries, an imported delicacy. Another (fig. 24, 6), looking like an asymmetric wooden knife, 17.8 cm. long, suggests the implement used for eating sea urchin ovaries. A third may have been intended for a wedge, although it shows no signs of wear. The functions of the other two specimens (fig. 24, a) cannot be guessed. There is also a section of a slightly curved branch (fig. 24, f), now (18) em. long and 2.5 cm. in diameter, which has been cut with a sharp metal (?) knife. Both (?) ends were originally bluntly pointed. It was found in a litter of charred shavings, twigs, and moss on the floor of House 9 in Old Town III. The proveniences of the other pieces of worked wood are: 15 from Old Town III, and 37 from Old Town II. Some of the larger pieces were sent to Dr. J. Louis Giddings, at the Haffenreffer Museum of the American Indian, Brown University, with the hope that these, to- gether with borings from livings trees in the area, might furnish materials for dedrochronological dating. ‘The samples were, however, insufficient. Other pieces of wood were given to Miss Elizabeth Ralph, in the Department of Physics, University of Pennsylvania, for radiocarbon dating. We are extremely grateful to her for the results reported below (p. 206). CANOES There are two bluntly pointed, slightly curved wooden sticks (fig. 24, e), flat on one side and rounded on the other. One from the Storage House in Old Town ITI is (13.4) cm. long; the other, smaller fragment is from House Pit I. It is possible that these were pieces of ribs for bidarkas or kayaks, since they resemble some Chugach specimens (de Laguna, 1956, p. 247), and our Yakutat informants reported that their ancestors long ago used sealskin canoes. These included large boats like umiaks, one-hole kayaks, and two-hole bidarkas. Obviously, there could never have been a complete boat in the Storage House. While the ethnographic evidence (cf. de Laguna, 1963) is sufficient to establish that the prehistoric Yakutat once made skin boats like those of the Chugach, this cannot be said of the Tlingit, with the possible exception of the Chilkat. However, the skin canoe used by the latter for crossing lakes when on trading trips to the interior, had a de Laguna} ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 185 covering of moose or caribou hide, not of sealskin (Drucker, 1950, Trait 390, p. 254), and was probably, therefore, of Athabaskan type; it was paddled, not rowed. The Chilkat tradition that they once had only skin boats, “before they knew there were other people living to the south and west on the coast,” is also suggestive of an interior origin. While the Yakutat themselves made or purchased from their southern neighbors several well-known Northwest Coast types of dugout, they made two distinctive types of their own. The forked- prow canoe for the open sea or swift currents was made only by the Eyak-speaking Gulf of Alaska Indians, from Yakutat to the mouth of the Copper River; the canoe with spoon-shaped bow and ram for sealing in the ice floes was made only at Yakutat and Icy Bay. Weare inclined to consider dugouts as peculiar to the Northwest Coast, so it is important to remember that they were made by the Chugach, Tanaina (the latter possibly in imitation of the Kenai Peninsula Eskimo), and even by some of the northern Koniag, although among these peoples the dugout was never as common as boats of other kinds (de Laguna, 1956, pp. 241 ff.). While the dugout is the modern type of craft on all the Northwest Coast, Borden (1951, pp. 46 ff.) had argued that the ancient inhabi- tants of Whalen Farm I and Locarno Beach I and II must have hunted seal and porpoise from skin canoes because they lacked antler wedges, pestle-shaped hand mauls, and large adz blades—tools which he believed were essential to making dugouts. Although the presence of such implements does indicate a well-developed woodworking industry, their absence cannot prove that boats were made of skin, not of wood, since the modern Coast Salish fell trees by burning or chiseling with a relatively small adz (Barnett, 1939, Traits 571, 572), and the Tlingit, at any rate, shape their dugouts with a small planing adz and crooked knife. This question has been further discussed by Osborne, Caldwell, and Crabtree (1956, p. 121). However, we should also note that Borden (personal communication) believes that the dugout was made in the Marpole Phase, which was in large part contemporaneous with that of Locarno Beach. Probably the question of boat types on the southern Northwest Coast in the most ancient days is not yet ready for solution. nad ‘ es. mere sits ‘ a) reas rh jung 1) nicer! ne ghee eis gotiua 9 yan: Satta ue AY A Al ahi Al paces. ak at Ot ft mayiser Phi? Soe lt atl RGR RRS, AWOL IEE pees h ‘J ar; ? . , , gl ae 4 : > M4 : POET Nds a Me eR 7 UF Po Mb iW cA atte Hy, AB eR Ub Se Ny i Ly : i | yur Pa bisa WEAR ROAN grakas : Pas a a ba a cea 7 pl hey iy salts Bs peer: ; age ; ce 1S wry ner be ; i Mies ry Rese a Lee LS LP Ke TNT “ry fogataa iy re ata’ ne a ' fsson - ~: $5 pa 7 EAR, Pe) eT a SON hee bs rofo8 -Apa ee yy eats fa ROA we eT wey ORT Oe m = : al J y 7 ' i Te erty ‘ i ND, Sf OSD ike Te a Abii he f » B OTRNUOL, Say Ces bib retin. able ve ee eee pone Toa isha el. one Tint Soil Janb pombe acy epee ore pit aa od) cull hen Hd ote Vay Rae ; (iontpdta wet if et e “3 id ; . q 2 ® } : ie % wl wt : LYCa LNT ATR RY ke OLN TEA ROE : ~ [ Hi 4 i : Pu ; we £ +i fe , P », bY hPa errr e ‘oe ane Wichune ys, Mk A LL A: 4 Tea ey ih Gene? Logon 2 Tah ee AT Nene ee ae +4 oer le he panaihiins | Nie laure a aa Rt PERL “ing eae solo te 0a) Sloe the naan ane ir «thy ae tate sin tires Mag Bar iksthietal THE YAKUTAT BLANKET By Carotyn OsBoRNE? The blanket remains which were recovered from a shaman’s grave on Knight Island, Yakutat Bay, were cleaned between two frames of plastic screening with sprayed detergent suds and clear water. Even though unfolding was done carefully the largest fragment meas- ures a very ragged 10% (warp) inches by 12% (weft) inches. This fragment, which is apparently from the main body of the robe (or ceremonial blanket), contains neither warp (top and bottom) nor weft (side) selvages. Fragments of both warp (top) and weft (side) selvages do exist, but I was unable to piece these to the other frag- ments to give continuity of design or weave or to give an indication of size of the blanket. There can be no doubt that all of the fragments belong to a single blanket. The many fragments have been placed with a high degree of accuracy, in the following order: i.e., the heavy geometric-patterned twined fragments with the fur binding as top selvage; the twilled-twined gold or yellow with the concentric rec- tangles of dark brown and with multiple tassels as the main body of the textile; the heavy warpwise-twined rows and the attached fringe and the wrapped bundles as side border. One fragment of this side border had a section of three-strand braid attached (?) to it; it may possibly have been part of the lower border. DESCRIPTION Materials—Samples of the various yarns (warp; light-colored weft; decorative yarn used both weftwise and warpwise; tassels; fur edging) were sent to the laboratory of the Federal Bureau of Inves- tigation, which very kindly consented to identify the fibers used in this and other local aboriginal textiles. Their findings are as follows: Warp yarn: Goat (includes mountain goat) dyed light brown. [There is a possibility that this might be burial staining; it seems probable that the original color was either natural white or light yellow.] Weft yarn: Goat (includes mountain goat) dyed light brown. [This isthesame color as the warp yarn.] 3 I wish to express my gratitude to the following: Mr. J. Edgar Hoover, who directed the laboratory of the Federal Bureau of Investigation to identify the fibers; Dr. Kaj Birket-Smith, who supplied photographs of the Copenhagen blanket; and the Portland Art Museum for lending the Tsimshian blanket from the Rasmussen Collection. This paper was read, in part, at the 1957 meetings of the Society for American Archaeology, in Madison, Wis. After this reading Dr. Arnold Pilling kindly sent me his notes on the original catalogs of the Cook and Vancouver Collections in the British Museum. 187 188 BUREAU OF AMERICAN ETHNOLOGY [BuL. 192 Tassel of blanket and decorative yarn: Goat (includes mountain goat) dyed dark brown. Side border decorative yarn: Goat (includes mountain goat) dyed dark brown. Fur edging: Otter (Lutra‘) or sea otter (Enhydra) natural light brown to dark brown. The warp yarns are two-ply, Z-twist (singles S-) approximately one-sixteenth of an inch in diameter; loose to medium degree in twist. They are invariably light colored. Warp yarns were set up 14 to the inch and were used in pairs or fours; never singly. The light weft yarns, which form the background color of the blanket, are also two- ply Z-twist as is the warp, but are smaller in diameter and somewhat harder twisted. The dark-brown decorative yarn and fringe yarn is considerably smaller in diameter (a minimum of one-thirty-second of an inch); it is also two-ply Z-twist and medium to hard in degree. The dark-brown side border decorative twining yarn is approximately one-eighth of an inch in diameter; two-ply Z-twist and usually hard twist. The attached fringe at the side selvage is light colored, two-ply Z-twist about one-eighth of an inch in diameter. Technique.—As previously mentioned, there are fragments of three or possibly four sections of weaving, of distinctive design and technique. Five fragments of the top border of the blanket are present, of which two show the complete complex of techniques (pl. 19, a). The larger of these measures 5% inches long (warp) by 6% inches wide (weft); this fragment and the next larger have portions of the top selvage andfur. The warp, as is common for the suspended- warp weaving of all Northwest Coast blankets (with the exception of the Salish which were ring-woven on a tension bar loom) was doubled at the top over a heavy loom cord, and secured with an inital row of plain twining. This row of twining was covered with the otter or sea otter fur band (hide and fur) about 1 inch wide, which was folded over the top edge so that it appears equally on both sides, and sewn. All of the twining in the blanket is carried over paired or quadrupled warps. The pitch of the background weave of the light-colored wefts, of the weftwise decorative twining (three-strand), and of the vertical decorative wefts (three-strand twining) is invariably up-to-the- left. This is in contrast to the elaborate stylized naturalistic Chilkat blankets in which change of direction of pitch of twining may be used to emphasize design breaks and changes. The sequence of the twining rows from the first weft working downward is: two rows of plain twining over paired warps; five rows of twilled-twining over paired warps; one row of three-strand dark- brown twining; 10 rows of twilled-twining over quadrupled warps. 4In view of the horror which the modern Tlingit and Eyak have of the land otter and the taboo against wearing its fur, we may assume that this was sea otter fur.—F. de L. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 189 The pattern then proceeds into a geometric zigzag design of dark- brown and light-brown wefts. There is no “eccentric” wefting in this section as is common in the Chilkat blankets; all of the wefts are at right angles to the warps. In all of this zigzag section, the weaving is accomplished with one light twining element and one dark. The pattern is therefore reversible; when the light appears on one surface, the dark forms the identical pattern on the reverse. Of necessity, in this zigzag design, therefore, the twining proceeds on a twill-line over paired warps (see fig. 25, a). Between each rectangle of the bold zigzag design are four vertical bands, one-half of an inch wide, alternating light and dark wefts. In the fragments preserved, there is not one of the complete rectangles of the zigzag design; their width is uncertain. The maximum of width of the zigzag block of any fragment, which is therefore a minimum for the design, is 5% inches. The warp length of the design band, which is complete in several fragments, is 2% inches. Unlike the Chilkat blanket weaving in which short pieces of weft yarns were inserted in the design pattern and locked with the adjoining wefts, all of the wefts in this textile proceed from selvage to selvage and were woven directly across the entire width, forming the zigzag designs and the vertical bars as these design elements appear in turn. Warp count throughout the blanket is 14 to the inch, used primarily in pairs. The Chilkat blankets often had warps added to aid in shaping the blanket to a curved shape. There was no evidence of added warps in the fragments of the Knight Island blanket. The weft count in these top fragments is 32 single wefts used in pairs (i.e., 16 weft courses or rows) an inch. At the base of this bold pattern is a single row of dark three-strand twine, followed by four rows of the twilled-twining of the light- colored wool; another single row of dark twining, three-strand; and two-rows of light twilled-twining. At the very bottom of one of the fragments, following the above series of twined rows, is a row of twining showing two light wefts and contiguous to these two dark wefts. These may well indicate that there were originally two complete rows of the horizontal bands of zigzags and bars. No more than this hint was present. The fragments of what appear to be the main body of the blanket are more numerous than the top border pieces. Completely cleaned and in good enough condition to be teased apart for thorough analysis were four large fragments, the largest of which is 10% by 9% inches. It is exceedingly unfortunate that none of this series could be fitted to the preceding pattern section. This portion of the blanket is primarily light wefted with designs of concentric rectangles of a deep reddish brown (pl. 19, 6). The basic weave is twilled-twining 190 BUREAU OF AMERICAN ETHNOLOGY [Buuy. 192 <& . < Ss57 ee meee eee ese ese pee = SS Se ees SS —->f- - —> —— Jam x SSS SS SS Sa SSS SE ~—— ee ee ae we ee ee ee LPL — —= ery AN Cpe S ethceeetieeetianstienstienetibentiadibemetianad ee ee me emer nm ene At pL ees eee al Rb he eli a oe ope aS eae eae ae ee ros eA INS A Ey Se = == 39 WS NOWe— —- — — - — - — é-—------- EN ONES F-— Te RCA A ICR SS Se > = Se l= OO Ae ee Che Gee SS = eae ERE A SE Soe NN ee 2 Bags SSE OIA ANNO SS SSS SS SANS eS SS RAR EXGN$5=3 pe AEN AW On 4 Ss SSS S even Ste PAO MEIN Ce See ee (Ese SY SOO SsSapolS ale = --—-¢ GN RE = = = — = SS fa mcw ie ON ONG 2 Beard aie cp aE IE BE NW Sy fs lua ay NE Ae ie ct ell Baars cand Pe SAN 0955S nS : a = se eS SS eee => = Ficure 25.—Diagrams of weave of Chilkat blanket from shaman’s grave, Knight Island, Yakutat Bay. (Not to scale.) a, Zigzag top portion of the Yakutat blanket, drawn by Arlie Ostlie; b, One set of concentric rectangles of three-strand dark wool on light-colored weft, central part of the Yakutat blanket (arrows indicate directions of yarn movements), drawn by Irene Brion from sketch by Carolyn Osborne. over paired warps. There are 8 courses or rows of wefts to the inch indicating 16 weft yarns. The decorative rectangles are accomplished in three-strand twining in which the yarns follow the outlines warp- wise and weftwise (fig. 25, 6); the three yarns needed for the rectangle were measured, halved and inserted at the midway point at the upper left-hand corner of each rectangle. At this point one-half of the de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 191 length is twined across the warps to form the upper line of the rec- tangle; these yarns, then, round the corner and, as the light wefts of the basic twined background reach this vertical line of twining, are twined around them (the wefts) to form the right side of the rec- tangle. Simultaneously, the second half of the length of yarns was being twined over the light basic wefts as these were met, and after rounding the lower left corner, crossed the warps to complete the lower side of the rectangle. At the lower left corner, the three yarns of each half were braided for about 114 inches, knotted and permitted to fall free as a tassel. The longest tassel yarns in place are 7 inches long. The complexity of the meetings of the basic weft and deco- rative yarn is increased by the fact that there are five concentric rectangles and a single center line. All of the vertical lines were inserted during the weaving process. None of them is embroidered as are many of the vertical lines in the stylized Chilkat blankets (in which the overlaid three-strand yarns cannot be seen from the reverse side of the fabric). The complications of the selvage to selvage wefting of the light yarns with the addition of the design elements during weaving is indeed a tribute to the skill of the weaver. At one stage she manipulated the basic weft to the outer vertical line; here these light-colored wefts act as warp to the dark-brown decorative yarns; then they are light wefts again for two warp pairs; then again they act as warp for the next concentric rectangle. At 10 places across each group of rectangles, the light wefts were re- leased to be the passive element as the dark-brown yarns were lifted to be twined around them. Each of these wefts moved through many such rectangles to the opposite selvage, there to turn around the end warp and begin the return in the opposite direction. None of the rectangles is complete in the fragments preserved. However, two fragments contain enough of the interior rectangles so that the length (warpwise) of 2% inches and width (weftwise) of 4% inches for the outer lines can be considered accurate. As each of the concentric rectangles ended in a self tassel, the blanket must have been well fringed. Each of the three large fragments contains portions of three rec- tangles, two of these on the same horizontal line and one on the next vertical line above or below the others. The rectangles were ap- parently distributed across the width of the blanket spaced 3 to 3% inches apart (distortion of the blanket may account for some of this discrepancy). Between the rows vertically (warpwise) the space was apparently 3% inches. The rectangles were not staggered; the top decorative lines of the rows of rectangles were inserted on the same weft line, and the vertical lines inserted between the same warps. 192 BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 There is no way of knowing the total number of rectangles in the blanket. There are four fragments of side border, of which one may possibly be a lower fringe corner as well. ‘The largest of these measures 5 inches long by 2 wide. The weft yarns, on reaching the edge of the blanket, turn on the outside warp and begin the next course below. The two outer warps are only slightly larger than the main body warps. Alternate rows begin the twilled-twining by (a) twining between warp 1 and warp 2, joining warps 2 and 3, and then proceeding to double-warp twilled-twining and (6) twining between the single warps 1, 2, and 3, to double warps 3 and 4, etc. Around each of the (b) rows is attached the extra fringe of mountain goat wool; the larks- head knot, with two cut ends, which forms the fringe is tied around the first warp, embracing the (b) row of twining. This fringe is now, at its longest, about 3 inches. The yarn is the light-colored mountain goat wool, two-ply Z-twist and as fine as the weft. Working inward from the selvage, between the third and fourth warps is a single vertical row of three-strand twining with fine dark- brown yarns. The next 15 warps are twined with the light weft yarns, but inserted between warps 18 and 19 is a very heavy (three- sixteenths of an inch wide) vertical row of three-strand dark-brown twining. Both of these are woven into the wefts as the weaving pro- ceeded to their longitude, even as the vertical lines of the rectangles. Immediately joining the heavy twined row (this connection has since broken) is a group of seven wrapped bundles. This wrapping resembles a loose coiling or sewing, inserted here in a vertical posi- tion. I cannot picture it in the blanket; there is no counterpart in ethnologic or known specimens, and these fragments were too disin- tegrated to provide an answer. I can only surmise that they func- tioned as the plaited bands which reinforce the sides of the Chilkat blankets. Nothing of its sort is present on any of the geometric blankets known. COMPARATIVE DATA The analysis of the blanket fragments may be summarized as follows for comparative purposes: 1. Warp and weft and decorative yarns are entirely of mountain goat wool. 2. The construction is entirely twining: two-strand and three-strand. 3. All of the twining wefts (excluding the dark rectangles in three-strand twining) pass from selvage to selvage of the blanket. 4. Decorative three-strand wefts were woven into the blanket; there is no sur- face embroidery. 5. Tassels of integral yarns suspend from some of the decorative elements. 6. A fur border was sewn to the top of the blanket. 7. Side selvages had an attached fringe of mountain goat wool yarns. de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 193 8. The design is entirely geometric. 9. Two colors only were used in the blanket; a light yellow or gold and a dark reddish brown. It is, of course, possible that there was also the natural wool used, but I could see no shading of the light-colored yarns. 10. All wefts cross the warps at right angles; there is no eccentric wefting. There are actually only two ethnological specimens which are com- parable to this archeologic blanket; and one ethnologic fragmentary specimen. The first of these is the comparatively well-documented ‘Swift’ blanket described by Willoughby (1910, pp. 1-10). Of the 10 above- summarized techniques and qualities, the Swift blanket shares 7 with the Yakutat fragments: 1, 3, 4, 5, 6, 8, and 10. The Swift blanket exhibits the wrapped lattice twining (Willoughby, 1910, p. 5, fig. 4) which is widely used in the Chilkat weaving but is not present in the fragments of the Yakutat blanket. This type of weaving involves weft yarns which show on one surface only. Both the Swift blanket and the Yakutat blanket exhibit a closely comparable weave: the vertical bands which divide the zigzag sections of both blankets in- volve two wefts of different colors, one of which provides a temporary lattice for the other. The yarn which acts as lattice for one colored band is the twining element for the alternate vertical band. This may be resolved as a matter of tension on one of the twining elements. This reversibility of design is not a regular feature of Chilkat blankets. It is seen occasionally in very small areas of design, such as the teeth motifs. The Swift blanket has side fringes which are extensions of the weft yarns; the Yakutat blanket has added yarns for the side fringe. Neither exhibits the braided or plaited band of the Chilkat blankets. The Swift blanket uses three colors: two dyed and an undyed natural white wool. The yellow and the dark brown are comparable to those of the archeologic specimen; as I have stated, it is difficult to be certain of the light color in the Yakutat blanket, but no variation in the light-colored wefts can be seen. There can be almost no closer design elements than the alternate zigzag and bars of the Swift blanket and the ones described above; the weaving technique of the bars is certainly similar and that of the zigzag identical (Willoughby, 1910, fig. 4, 6). The design of concen- tric ‘lazy’? H’s of the Swift blanket is very like the concentric rec- tangles of the Yakutat blanket, and the tassels pendent from the lower right-hand corner of the design appear in both specimens. The Swift blanket seems to have, in addition, tassels pendent from the bold geometric diamonds. ‘These are braided and may or may not be integral weft yarns. In general, the Swift blanket is more complex in design than is the 194. BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 Yakutat blanket. This complexity is accomplished, however, with the addition of only a single technique: the double wrapped lattice twining which appears only in the borders of that blanket. The Swift blanket was collected about 1800, without provenience, although Willoughby considered it to be a northern British Columbia coast product. J would consider it, in view of the subsequent remarks, to be from farther north. Willoughby (1910, p. 8) notes that there is a dilapidated sample of a geometric-patterned blanket in the British Museum collected in about 1793. Kissell (1928, fig. 2 and p. 118: her text has the figure numbers reversed; figure 3 is the Ottawa blanket illustrated by Emmons, 1907, p. 388, and by Willoughby, 1910, pl. 2) states that this blanket was probably collected by Vancouver somewhere between Prince Frederick Sound and Lynn Canal in 1794. I have at hand a series of notes from Arnold Pilling on the following manuscripts: Hewett, George Goodman: “Vancouver Voyage, Hewett Coll*.’? British Museum Ethnological Document #1126. Manuscript catalogue written during the Vancouver Expedition, 1793-1794. Original document in the British Museum. The following entries are noted: TiS, WNeamakizat oe 2. = Ae oe ne ee ee Shoulder piece. PAO :serhee et GGJi. ULF OLE Ss. AEs Oe. eae Bark garment. 111 Mowachut or Nootka Sound___---.-----~-- Bark and wool garment. 267), | Rock tVillaceus 22 ets. Pa ee ee Garment. None of these, it seems to me, can identify the British Museum speci- men. However (also in the notes taken by Arnold Pilling), there is mention of two blankets collected by the Cook Expedition in 1778 and presented to the Museum in 1789 by Joseph Banks. These specimens are: NWC 49 Cloak of woven fibre with heavy fringe of rough fibre and twisted cord intermixed. NWC 51 Robe of brown and white twisted cord of a woolen material woven in a diamond pattern: wool of the mountain goat. There is no way of determining the age of this catalog or whether it is Banks’ original notes or notes made by later museum workers. However, I am satisfied that item NWC 51 is this fragment, and that it was probably collected by Cook in 1778. It seems likely that Miss Kissell assumed that it was collected by Vancouver because his journals describe a native of the Lynn Canal wearing a robe of this description. Neither Kissell nor Willoughby states that this blanket is of pure mountain goat wool (though the catalog notes indicate that it is). From the fragment pictured by Kissell I would state that this frag- ment shares characteristics 1, 2, 3, 7, 8, 9 and 10 with the Yakutat fragments. There are no sections of top border pictured and it is de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 195 possible that elsewhere in the blanket there are design elements of three-strand twining. Its boldness relates it to the Copenhagen blanket. It shares the zigzag patterning with all the others. The illustration shows no tassels. What is most interesting is the sug- gestion of wrapped bundles at the lower border. A detailed analysis of this blanket fragment is certainly a desideratum. A third known blanket, very inadequately described, is illustrated by Boas (1951, pl. 10) and is a part of the collections of the Ethnologi- cal Museum at Copenhagen. This blanket was formerly part of the Leningrad collections and was acquired by early Russian explorers on the coast. It seems possible that it could have been collected by Lisiansky, who described Indians wearing tasseled blankets at Sitka (Kissell, 1928, p. 117) in 1805. Dr. Birket-Smith very kindly sent photographs of the blanket (Museum No. K.c.119) including enlarge- ments of details. Its overall similarity to the Yakutat blanket and the Swift blanket is very apparent. The horizontal diamond bands and the zigzag bands are of the same temporary-lattice type twine weave as has been noted for both of the preceding blankets. It would seem to share characteristics 1, 2, 3, 7, 8, 9, and 10 with the Yakutat robe, lacking only the fur binding and decorative three- strand twining. The tassels appear to be inserted yarns. Its design elements are bolder in overall pattern than either the Yakutat or Swift blanket; it lacks the lighter design of concentric rectangles or “lazy” H’s. In all three of these blankets the zigzag designs are separated by vertical bars. The Swift and Copenhagen blankets share the hori- zontal diamond pattern. ‘The British Museum specimen has a zigzag border both horizontal and vertical, as do the Copenhagen and the Swift. This pattern does not appear to have been present in the Yakutat blanket, but, of course, the fragmentary condition of the specimen does not permit a positive statement of the absence of the pattern. Only the Yakutat blanket has wrapped bundles. None of the four has the sewn-on plaited band or bands of the Chilkat blanket. A blanket which stands as unique up to the present is the often- illustrated cedar-bark and wool blanket (Emmons, 1907, pl. xxrv, 1; Boas, 1951, pl. x1) in the British Museum. It would seem that this cape or blanket ° is probably the ‘111” of the Hewett Collection of the Vancouver Voyage (see Arnold Pilling’s notes, mentioned above, p. 194): 111 Mowachut or Nootka Sound—Bark and wool garment. 5 All-cedar-bark garments are commonly called capes, while the wool and wool and cedar-bark garments are generally referred to as blankets. The term ‘‘cape”’ is certainly more accurate in a functional way. 196 BUREAU OF AMERICAN ETHNOLOGY [BuLy. 192 It resembles the fine cedar-bark capes of the Nootka in its open space twining. It lacks the heavy rolled and plaited side binding of the specimens that I have observed in the Washington State Museum, and it has the addition of mountain goat wool in a twined geometric upper and lower border. It is also deeply fringed at side and bottom. The geometric patterning is simpler than in any of the previously discussed blankets and the lack of a photograph from the reverse side leaves us with no indication as to the type of weave. All of the blankets discussed have been assigned to the same general period. The earliest would seem to be the 1778 (?) collected British Museum fragment. The archeologic specimen (Yakutat) could date from the 1780’s. All the rest were collected before or at the turn of the 19th century. The only one of the blankets with known pro- venience is the blanket from the shaman’s grave on Knight Island in Yakutat Bay. With, however, the added information that Van- couver saw a tasseled blanket at Lynn Canal and Lisiansky saw Indians wearing tasseled blankets at Sitka, I submit that it fixes the geographical provenience of the all-wool geometric-patterned blanket as the northern coastal area of southeastern Alaska. All of these predate any specimens known as “‘T'simshian”; and all fall in Tlingit territory. The earliest known garment of mixed cedar bark and wool comes from a more southern location—Nootka—but dates (if my assumption about the Hewett collection of Vancouver Voyage is correct) from approximately the same period—i.e., the 1790’s. It also exhibits geo- metric patterning. All of this patterning is evident in the fine twined basketry of the Tlingit and other Northwest Coast tribes, and also of the Aleut. It would seem to me that, in the earlier aspects of weaving, wool blanket weaving was, like basketry, a woman’s craft, and she used designs and patterns and skills (witness the selvage-to-selvage welts; three- strand twining; wrapped lattice weave; lack of eccentrics and vertical overlay twining) with which she was familiar as a basket weaver, in both the northern and southern areas. The new art style—the highly elaborate stylized naturalistic design of the Chilkat blanket—was probably a male development. The Chilkat blanket was woven by women who copied pattern boards made by the men, and who of necessity invented and adapted new techniques to meet the requirements of design. This new style was in accord with the total artistic and social development of its time. It is the Tsimshian who have been credited by Emmons, and by Boas following the mythology, and by most later writers, with the origin of the Chilkat blanket. Because of the apparent antecedence of the northern blankets of geometric style, and largely because of the lack de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 197 of evidence of early weaving in the Tsimshian area, I would assume that the mythology refers to the elaborate ‘‘totemic” or heraldic designing of the so-called Chilkat blanket. A total lack of information on Tsimshian weaving and the very dubious provenience of the great majority of blankets that are tentatively labeled Tsimshian makes further discussion of this origin mythology fruitless. Of the development of the historic Chilkat blanket, the following items are outstanding: There are no blankets with stylized naturalistic designs (unless the fragment in the National Museum of Canada is an exception) which do not have the cedar bark and mountain goat wool wrapped warp. There are no blankets with the stylized naturalistic designs which are dated before the beginning of the 19th century, and the oldest with known date of collection appears to be about 1830 (Willoughby, 1910, p. 10). Emmons (1907, p. 390) says that his earliest blanket was supposedly the first woven by the Chilkat and was copied from an old Tsimshian blanket; it was said to be several generations old. The introduction of the elaborate designs brought new weaving tech- niques unknown to basketry and necessary to develop the patterns of a multiplicity of small bodies of color. All of the Chilkat blankets have the rounded lower edge, the so-called five-sided shape. None of the Chilkat blankets is as finely woven as are the geometric pat- terned; none is as flexible. The addition of cedar bark made for greater rigidity; it also, of course, supplemented a probably limited supply of mountain goat wool. It permitted an expansion of the quan- tity of production which agreed with the need for wealth-display items. There are only a few blankets which fill the technologic and decora- tive gap between the geometric Yakutat, Swift, Copenhagen, and British Museum specimens and the historic Chilkat (and/or Tsimshian) blankets. A cut-up blanket, incorporated into a dancing shirt, in the former Museum of the Geological Survey of Canada (now the National Museum of Canada) is illustrated in part by Kissell (1928, fig. 3: n.b. that the description of this fragment on page 117 wrongly refers to fig. 2); by Emmons (1907, fig. 58, p. 388) and by Willoughby (1910, pl. 2). None of the authors states whether the fragment is of pure mountain goat wool or mountain goat wool and cedar bark mixed. This fragment shares the zigzag and bar design with the geometric- designed blankets and has a form of concentric triangle designs. The selvage-to-selvage wefts in the geometric pattern are interrupted by a stylized naturalistic design in which short lengths of yarn are inserted to make the pattern in typical Chilkat techniques. There are a few tassels added to the naturalistic design, though the geometric- 693-818—64——_14 198 BUREAU OF AMERICAN ETHNOLOGY [Buuu. 192 patterned section seems to have none. This fragment is totally without provenience either temporal or geographical. The blanket illustrated in Davis (1949, p. 64) is part of the Rasmus- sen collection of the Portland Art Museum (cat. No. 49.3.546). The card on the blanket reads as follows: Chilkat Blanket. Tsimshian. 10-18-39. A very old blanket collected by Mrs. Kirk, wife of the soap Mfg. in San Francisco in 1880. ... Date? Un- usually fine weaving. Moth eaten places. Sea otter fur ? Laced in top. A remarkable piece. Its faded colors match the Chilkat blanketry leggings I got from Nass River. This blanket was lent for comparative study by the Portland Art Museum. It is indeed a remarkable and beautiful blanket. It has the typical five-sided shape of the Chilkat blanket with the very deep lower fringe (never characteristic of the geometric-patterned blankets). It has a wool-wrapped cedar-bark warp and fringe. The warp and weft counts indicate greater fineness of yarns than any Chilkats that I have observed. The center yoke is a geometric-patterned, two- colored section in which the wefts proceed from one side of the yoke to the opposite. The geometric pattern is obviously not as intricate as any of the previously observed; but its technique is the same. The stylized naturalistic section surrounding the yoke uses every known Chilkat technique. The colors are unique insofar as my observation and Emmon’s remarks are concerned. The green is a decidedly yellow green and rather deep in the unfaded reverse side; the usual pale yellow, dark brown or black, and the natural white complete the colors. The pattern is simpler than in many blankets; the faces are upside down to the observer as in old dance kilts. The wing design is also reversed and it lacks the common three-section division of design. Emmons (1907, p. 388, fig. 581, a) illustrates one other blanket which he calls Tsimshian. It has no geometric patterning, but the design is aberrant from the usual Chilkat designing. This is also true of his oldest Chilkat blanket which he was told was copied from a Tsimshian blanket (ibid., p. 390, fig. 580). A few old dance shirts have bands of geometric designs in selvage- to-selvage twining. Washington State Museum cat. No. 1-631, Chilkat Tlingit, is a shirt that shows the geometric band in the back of the shirt only; it is simpler in its development than any of the wholly geometric-patterned blankets. The shirt has a wool-wrapped cedar bark warp, as, I believe, do all of the dance shirts. It is on the basis of the known provenience of the Yakutat blanket that it has been possible to place more accurately the all mountain goat wool, geometric-patterned blankets. Their more northerly location seems to indicate an earlier center of blanket weaving than that de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 199 attributed to the Tsimshian. The stylized-naturalistic design blankets seem to combine traits from three earlier weaving centers: (1) the mountain goat wool blankets from southern Alaska; (2) the highly developed art style of the Tsimshian (with its probable heraldic significance) ; and (3) the cedar-bark cape weaving of the Nootka and southern groups (which probably added the more rounded shape, the plaited bands at the side selvages, and the very long lower fringes, as well as the new fiber). The intermediate blankets show a gradual blending of design; they are few in number, and though one at least is designated as Tsimshian they are really of unknown provenience. pial «otal yh oe ea 7 Besos) ana esihele wereslt ' arhhe (BY. Sabet, teat de, sine af ae i oibiweal aidachotc ped lier): aidan Tada sty Rie ial ns Hoot will ra gal BA Cogn sade abag edi | (la ante Polya sat’ out -ond babies ydadors dots) a : mene -aiol, gem ab Dales <4 oa ties 3 bikiay ont Te ‘ll a Cede goed ald a@ He “raw: “bireted lanier R wild ali7r caine Adelie 2g find, imitash at 2 val oh AIO) iano he ‘t9 dp phe: a 15 Peak. & CONCLUSION By FREDERICA DE LAGUNA AGE OF THE SITES Since Yakutat archeological materials have many striking similarities to those from Prince William Sound, the criteria used to suggest a relative chronology for Chugach sites (de Laguna, 1956, pp. 64 f.) may be partially applicable. Chugach cultural periods were: (1) Older prehistoric (perhaps contemporaneous with Kachemak Bay sub-III and III, and with most of the lower levels at Uyak Bay, Kodiak Island): Decomposed shells in middens, incised stone plaques, relative abundance of planing adzes over splitting adzes and small woodworking tools, relative abundance of tanged slate blades and slender slate points like awls over barbed slate blades, chipped ulo- shaped scrapers, and absence of native copper. For the beginning of occupation at Palugvik, Prince William Sound, which must have been within or possibly at the beginning of this cultural stage, Rainey and Ralph (1959, p. 368) have published two radiocarbon dates. These are based on one house post (P-174 and P-192, cat. No. 33-37-476) found at the bottom of the midden. This post had 105 rings and had been coated on the outside with paraffin. The date obtained from the outer part of the post was 1753 + 105 B.P., or A.D. 100 to 310. The second date from the core of the post, from which 83 years was subtracted, was 1727 + 105 B.P., or A.D. 126 to 336. A wooden shovel blade, similarly treated with paraffin (P-173, 33-37-481), gave a date of 2265 + 112 B.P., or 419 to 295 B.C., which was discarded because of suspicion of paraffin contamination. While the dates from the same house post corroborate each other, they should only be taken as suggestive of the age of the older prehistoric period in Prince William Sound, since they are not part of a series. The assumption of relative contemporaneity with similar cultures on Kachemak Bay and Kodiak is only a guess. We should note that the single date obtained from five pieces of caribou antler from the Period III level of the Yukon Island site in Kachemak Bay (P-138) is 1369 + 102 B.P., or A.D. 487 to 691, but it stands alone. It can at best only suggest that this date fell somewhere within the timespan of Kachemak Bay, but can determine neither its beginning nor its end. Nor is it possible to exclude the suspicion of contamination from sea water or from preservatives. The same hesitations should apply to the interpretation of the single 201 202 BUREAU OF AMERICAN ETHNOLOGY [Bunn. 192 date for Kachemak Bay I, obtained by combining eight pieces of antler from Yukon Island I (P-139). This gave a reading of 2706 + B.P., or 866 to 630 B.C. (Rainey and Ralph, 1959, p. 371). Again this stands alone (de Laguna, 1962). (2) Later prehistoric (possibly contemporaneous with late Kachemak Bay III and IV, and with the upper levels at Uyak Bay): Less de- composed sheils, abundance of fire-cracked rocks (probably used for the steam bath), predominance of the splitting adz and of small chisels and other small woodworking tools over the larger planing adz, development of variants of the splitting adz such as the grooved ax, the ax-pick, and the adz-ax, presence of war club heads (double- pointed stone pick and chipped stone pick), increased popularity of barbed stone blades, appearance of native copper in the latest deposits. There are no radiocarbon dates for this period. (3) Protohistoric: The same types as in the later prehistoric except for the addition of blue glass beads of the type seen by Captain Cook in 1778. Presumably iron was also present, although we found none in Chugach sites. (4) Historic (since 1783-84, with the beginning of Russian expansion into Prince William Sound and along the Gulf of Alaska): Trade goods, especially small glass beads (like those found in the grave on Knight Island), skeletal remains with lesions of syphilis and tuberculosis, and, still later, the appearance of Christian burial. Admittedly, the only distinction between late prehistoric and proto- historic sites rests on the presence or absence of Cook type glass beads. Since these were undoubtedly rare and precious, their absence from a site that yielded few personal ornaments cannot be taken as proof of prehistoric age. It is, however, probably significant that no beads or any objects proving direct contact with White men were found at Old Town, while Cook type beads and an iron arrowhead came from Shallow Water Town on Little Lost River. This suggests that such beads would have been encountered in Old Town middens, caches, and house pits, if the latter site had been inhabited at the same time as the small settlement on Little Lost River. Old Town was probably abandoned before these beads became available to the Yakutat people. Cook type beads were among the trade goods carried by the Rus- sians, although they were disseminated to the Chugach before the Russians themselves came to Prince William Sound. They were prob- ably of Chinese manufacture, and it is tempting to surmise that the first to reach Alaska may have been the 20 strings of Chinese beads, left by Bering’s expedition in 1741 in a Chugach house on Kayak Island in Controller Bay (de Laguna, 1947, pp. 242 f.; 1956, pp. 60 ff.). In any case, we may hazard that Cook type beads, without other de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 203 trade goods, might appear at sites dating from the middle to the late 18th century, and that Old Town may have been abandoned some- what earlier. It will be remembered that scraps of iron were found only in Mounds A and B (Old Town III and IT), and not in what is considered the oldest section of the site, Mounds C and D and House Pit 7. Theoretically, there are three possible sources of such iron (de Laguna, 1956, pp. 61 ff.). First, it might have been obtained from Asiatic sources by purely aboriginal trade via Bering Strait, where the Eskimo apparently had iron points for engraving tools from one to two thousand years ago. Although this source may have supplied the tiny scraps of metal used for incising the compass-drawn dot-and- circle designs on artifacts from Kachemak Bay III, early Kodiak, and early Aleutian sites, the iron found at Old Town must be more recent, for it is in larger pieces, and there is no evidence that any but the most minute fragments were known in Alaska until after the middle of the 17th century, when the Russians had established themselves on the Anadyr River in Siberia. Thus, the earliest iron knives in northern Alaska do not antedate the end of that century. A second possibility is that iron was obtained by the Aleut and Pacific Eskimo by more direct trade with Asia, perhaps as their tradi- tions suggest, from Russian or Chinese voyagers who may have preceded Bering in the early 18th century. The third, and most likely, source of iron is nails and bolts in drift- wood and wreckage, which increased maritime activity during the 17th and 18th centuries would have made available (cf. Rickard, 1939). While we do not know how early the Yakutat people may have obtained drift iron, it is likely to have been before they acquired glass beads, since the latter could only have been obtained through trade. We may, therefore, be able to distinguish an early protohis- toric period with drift iron but without beads, and a later proto- historic period with both drift iron and Cook type beads. This is, in fact, what is suggested by the Yakutat sites, if Old Town II and III represent the earlier phase and Shallow Water Town the later. Another clue to the relative age of Old Town is provided by the occurrence of native copper. This is, as far as we know, the richest Alaskan site for native copper with the possible exception of Dixthada, near Tanacross, on the upper Tanana River, close to the source of the metal (cf. Rainey, 1939, pp. 364-371). Copper first appears at Old Town in the form of three pieces in the upper layers of Old Town I, and steadily increases in quantity through the deposits of later periods. There is, of course, some uncertainty as to the age of copper working in Alaska. Because it is undoubtedly old in many parts of the New World—the Old Copper Culture of the Great Lakes area yields radio- 204. BUREAU OF AMERICAN ETHNOLOGY [BuLL. 192 carbon dates of 5,600 or 7,150 years ago +600 years (Wormington, 1957, p. 150)—it is natural that Birket-Smith (1953, pp. 225 f.) should suggest that it was accidental that copper was found only in sites of the very late prehistoric period in Prince William Sound and Kachemak Bay. We should also note that ornaments of native copper occur in both Beach Grove and Marpole (Eburne), with radiocarbon dates from the latter site ranging from 943 B.C. +170 to A.D. 179 +60 years (Borden, personal communication). Despite this range in dates, and Dr. Borden’s hestitation to accept the earliest count from the site, there seems to be no question but that copper working was old on the southern Northwest Coast. However, the Yakutat and Atna stories about the origin of copper working lack the mytho- logical character which would suggest great antiquity for this art; rather, the protagonists appear to have been ordinary Indians, though supernaturally blessed with luck. There is, of course, no necessary connection between the copper work in these two areas. We may assume, therefore, that Old Town was settled in the late prehistoric period and abandoned before the late phase of the proto- historic period. However, it is possible that the postulated “earlier protohistoric period” of drift iron may overlap in time the ‘later pre- historic period” of native copper alone. The Yakutat natives had easier access to the great ocean beaches where wreckage was found than did most of the Chugach, and may always have had more iron than the latter, and have found it earlier. Absence of iron from Chugach sites where it might have been expected, may, therefore, be due to geographical factors. We do not know how early in post-Wisconsin times the Yakutat area may have been open for human settlement. Riddell has reviewed the geological and botanical data which may indicate at what period or periods southeastern and southwestern Alaska could have been inhabited. In summarizing these findings (1954, p. 105), he concludes that: “Even the currently most heavily glaciated region of the coast, of which Yakutat Bay is the approximate center, probably was suffi- ciently deglaciated and populated with the necessary types of flora and fauna to allow human occupation about 6,000 years ago.”” The Aleutian Islands and the southern Northwest Coast were habitable several thousand years earlier. About three to five thousand years ago occurred the post-glacial optimum in southeastern Alaska, followed by a glacial advance (Riddell, 1954, pp. 75 ff., 177). It must be remembered ... that the present icefields in the Pacific Coast region are not remnants of the Wisconsin advances, but are the residue of a rela- tively small ice advance of approximately 3,000 years ago. This period of ice advance is sometimes referred to as “‘the littie ice age.”’ Before this advance, the glaciers in Alaska had retreated a greater distance than have the present glaciers. [Ibid., p. 108.] de Laguna] ARCHEOLOGY, YAKUTAT BAY AREA, ALASKA 205 In general, during the past two millennia, the glaciers in the Pacific Northwest have been retreating, except for local fluctuations, for example, an advance culminating between A.D. 600 and 920 in Icy Bay and between A.D. 970 and 1290 in Yakutat Bay (Plafker and Miller, 1958). The earlier “‘little ice age’ and this (and other local) readvance may not have prevented human occupation of all of the northern Northwest Coast and Alaska Gulf Coast regions, yet they may have been severe enough to have discouraged or even pre- vented travel and hence cultural exchanges along this part of the coastline. Furthermore, we cannot be confident that we yet know what parts of that coastline or what present foreshores were above sea level in critical periods in the past, although raised beaches in the Gulf of Alaska and in southeastern Alaska attest to uplift in both geologically and historically recent times. In any event, there are no archeological remains known at present which suggest habitation in the Yakutat area before the glacial advance between A.D. 970 and 1290 in Yakutat Bay. These dates are based on radiocarbon analyses of wood from an end moraine near Ocean Cape. Occupation may well have begun, however, some time during the 15th or 16th century, when the ice had retreated. According to recent studies made by the U.S. Forest Service, the forest growing on the outwash apron at the outer margin of the end moraine near Yakutat is a relatively even-aged spruce stand with an average age of nearly 550 years .. . The oldest of 27 spruce trees on which an accurate age count was obtained in 1953 was then 553 years old .. . This indicates that recession of the Yakutat Bay lobe began before 1400 A.D., assuming that the outwash apron did not become stabilized and suitable for forest growth until after the recession had begun .. . the spruce forest probably did not become established on the outwash apron near Yakutat for at least 50 years after the tidal ice front began to retreat into Yakutat Bay. ([Plafker and Miller, 1958.] Knight Island itself would not have been uncovered until somewhat later, and a permanent village would hardly have been established directly under the end of a huge glacier. During 1953, logging opera- tions were carried out near Redfield Cove, on the east shore of Yakutat Bay, opposite Dolgoi and Kriwoi Islands. Borings from some of the trees cut at that time were examined by Dr. Giddings, who informs us (letter of April 4, 1958) that these trees began to grow in A.D. 1530, 1630, and 1660. This information gives us some clue as to when Knight Island might have been occupied. Spruce trees growing in House Pit 7, and thus indicating the abandonment of the oldest part of the site, had circumferences ranging from 5 feet 6 inches to 8 feet; a hemlock had a circumference of 3 feet 10 inches. Borings of spruce trees growing in House Pit 7 and in other parts of the site were examined by Dr. Giddings. Unfortunately the blade of 206 BUREAU OF AMBRICAN ETHNOLOGY [BuLL. 192 the core extractor was only 12 inches long and thus much too short to reach the center of the larger trees. It was impossible to cut them down and secure core samples from logs, as was done with the speci- mens from Redfield Cove. In consequence, we do not know when the oldest trees on the site began to grow.