LATE WOODLAND CULTURAL DYNAMICS IN, THE NORTHERN. LOWER PENINSULA OF MICHIGAN . Dissertation far the Degree of Ph. D. MICHIGAN STATE UNIVERSITY WILLIAM ANTHONY LOVIS, IR. 1973 ' ‘ cw" 4 - < “ {NWf-uhiw '. z ‘ JI’lll/IIIIIIIIWNr J" ‘ 000596112 V Michigan State: ’ University T-l-li—smis to certify that the A thesis entitlea~ A -- -——. ._._.. LAFTEflOODLAND hCULTURAL DYNAMICS IN THE NORTHERN‘meI; PENINSULA OF MICHIGAN ,. AMI-.4; .r-:;‘.'.:“ 1:- . - .___.. .,._-. ,,_ presented by WI‘BEIAHfiIWONY LGVIS , JR ._ __ has been accepted towards fulfillment of the requirements for Doctoral degree in Anthrogology I: jor professor Charles E. Cleland W5 1‘ "‘ k.. ——_< _ Date 0-7639 i I I PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before dde due. ll DATEQQE DATE DUE DATE DUE “ MSU Is An Affirmdive Action/Equal Opportunity Im smarts-9.1 W - ABSTRACT LATE WOODLAND CULTURAL DYNAMICS IN THE NORTHERN LCNER PENINSULA OF MICHIGAN By William Anthony Levis, Jr. The partially stratified O'Neill Site (20CX18) Charlevoix County, Michigan, was excavated by the Michigan State University Museum during 1969 and 1971. Ebcamination of the material culture coupled with a radiocarbon chronology indicates an occupation span from ca. A.D. 1000 to A.D. 1700. Ceramic analysis demonstrated use by groups common to both the Straits of Mackinac area to the north and the Traverse Bay area to the south. A sparse chipped stone assemblage belonging to a flake industry is described and interpreted. Warm season occupation of the site, from late spring until early fall, is documented from the faunal remains. Spatial analysis of the assemblage points toward a number of localized occupation areas. Combined with large amounts of unworked raw material, chert cores, and debitage, the primary role of the O‘Neill Site is believed to have been as a chert processing station utilized by small, transient groups. Ceramics from seven Late Woodland sites are examined within a temporal and spatial framework. Key style attributes are exposed through the use of proximity analysis, and a distance-similarity hypothesis based on decorative attributes is tested and rejected. Testing of difmsion hypotheses leads to a reformlation of the extant interaction model. Implications of the latter are discussed, fixture research objectives are proposed, and a testable model of ceramic similarity in the study area is g enerated . LATE WOODLAND CULTURAL DYNAMICS IN THE NORTHERN LWER PENINSULA OF MICHIGAN By William Anthony Levis, Jr. A DISSERTATION Submitted to Michigan State University in partial Mfillment of the requirements for the degree of DOCTOR OF PHILCBOPHY Department of Anthropology 1973 U at} ACKNOWLEDGMENTS Although a single name is appended to the title of this manuscript, in actuality it is the product of input from a number of persons and organizations without whose suggestions, help, and guidance it could not have been completed. Funding for the project came from a number of Isources. Excavation of the O'Neill Site was partially financed through a National Science Foundation Undergraduate Research Grant (GI-2825), National Science Foundation Research Grants entitled "Prehistoric Cultures of the Grand Traverse Bay Area, Michigan" (GS-1669), and a grant funded by the National Historic Preservation Act through the Michigan Department of Natural Resources (Project No. 26-71-00025), awarded to Charles E. Cleland, Michigan State University. Monies for computer time were made available by the College of Social Science, Department of.Anthropology, and Computer Institute fer Social Science Research, Michigan State University'under a proposal entitled ”Cultural Dynamics of the Traverse Corridor." I am indebted to Rollin H. Baker, Director of the Michigan State University Nbeeum, for making space, facilities, and.supplies available while the analysis was in progress. Personal support, funding for manuscript preparation, and typing were financed by a Woodrow Wilson Dissertation Fellowship. 'Without the support of this organization, my finances would have been sorely strained. A special debt of gratitude is owed my Doctoral Committee: Charles E. Cleland, Chairman, Lawrence H. Robbins, and Joseph Chartkoff. Their ii guidance of the research and their advice on manuscript drafts was helpful and perceptive, while the diligence with which they'managed to meet tight personal deadlines is greatly appreciated. For his constant personal and professional interest Dr. Cleland, who originally'proposed the research problem, deserves special mention. James A. Brown, Northwestern University, both participated in many informal discussions of the analytical procedures and perused the manuscript and.made many constructive comments despite his heavy'schedule. I hope his contributions are evident. Investigation of the O'Neill Site would not have been possible without the permission of the property owner, Jehn O'Neill. His cooperation in ‘ allowing excavations to take place assured that the information from.this important site would not be lost. Laboratory'handling of the assemblage was aided by'a number of persons: Tracy'Mengerink worked on the ceramics while at The Museum as part of the Kalamazoo College Career Service Program; Steve Devine and Jean Farr did.much of the preliminary flint analysis; and Denise King, MSU Department of’Anthropology, analyzed the faunal remains. Aid on much of the statistics and computer programming was furnished by Robert MainfOrt, whose help was instrumental in the completion of the analysis. Additionally, the help and encouragement of my friends and colleagues Patrick Martin, Nancy Newak, Donnie Fitch, and Margaret Rogers must be acknowledged. I, however, take sole responsibility fer any'errors of analysis or interpretation that might occur. iii TABLE OF CONTENTS Page ACKNWIEDMNTIS .0...OOOCOOOOOOOO0.0...O...OOOOOOOQOOO...0.000.0000 ii LIST OF IILUSTRATIWS eeeeeeeeeeeeeeeeoeoeeeeeeeeeeeeeeeeeeeeeOeeee Vii LIST OF MEI—m OOOOOOOCOOOOOOOOOOOOOOO...OOOOOQOOOOOOOOOOOO0.0.0... x Chapter 1 Dmmmmmflm..n.u.u.n.n.n.u.u.u.n.u.n.n.n.n 1 The Corridor Concept .................................... 1 History of Investigations ............................... 13 Glacial Geology and Topography .......................... l7 sails and vegetation 00......0.0.0...OOOOOOOOOOOOOCOOOOO. l9 Fauna 000......CCOOC.OOO00......00.0....00.000.000.000... 21 Cljmte 00.0.0...00......O...0...OOOOOOOCOOOOOOOO0.0.0... 23 Descriptions Of Excavations 0.00.00.00.000’COOOOOOOOOO... 2y Stratigraphy and Dating ................................. 28 2 CERAMICS .................................................. #1 Body Shards ............................................. 43 Formal Ahalysis and Description.......................... 50 Discussion .............................................. 9O 3 THE CHIPPED STONE INDUSTRY ................................ 107 Raw Material ............................................ 107 Cores ................................................... 110 Lithic Debitage 125 iv TABLE OF CONTENTS (Cont.) Chapter Page EHMms.n.u.u.”.u.”.n.n.n.n.u.n.n.u.u.n. IDl Unifaces ................................................ 152 Discussion .............................................. 157 h PECKED AND GROUND STONE ................................... 171 Packed Stone ............................................ 171 Ground Stone ............................................ 178 Discussion .............................................. 187 5 NATIVE COPPER, PREHISTORIC PIPES AND WORKED BONE .......... 188 Native Capper ........................................... 188 Prehistoric Pipes ....................................... 190 Worked Bone . 191 6 EURO-AMERICAN ARTIFACTS ................................... 199 European Trade Items .................................... 199 Modern Artifacts ........................................ 207 Discussion .............................................. 212 7 FAUNAL REMAINS ............................................ 213 Mammals ................................................. 213 Fish .................................................... 214 Avian Fauna ............................................. 215 Nfiscellaneous ........................................... 215 Discussion .............................................. 216 8 FEATURES .................................................. 221 Structures .............................................. 222 9 SPATIAL.ANALYSIS AND CONCLUSIONS .......................... 2ND Spatial.Ana1ysis ........................................ 240 V TABLE OF CONTENTS (Cont.) Chapter ’ Page Conclusion and Summary .................................. 251 10 LATE WOODLAND CULTURAL DYNAMICS ........................... 255 The Hypotheses .......................................... 256 Theory .................................................. 258 Technique ............................................... 269 RBSflts Of we AnaJij-S 00....OOOOOOOOOOOOO0.00.0.0000... 277 APPENDIX A O'NEILL CURVILINEAR TYPE DESCRIPTION .................. 322 APPENDIX B O'NEILL SI'IE CORRELATION MATRICES ..................... 323 APPENDIXC EMCANRBU‘E LIST OOOOOOOOOOOOOOOOOOOOOOOOOOIOOOO. 330 REFERENCES CITED O...0....0......O...OOOOOOOOOOOOOOOO00.000.00.000. 335 Figure 1 Map of the Traverse Corridor .................................. 2 Presettlement vegetation in the Traverse Corridor ............. 3 Presettloment vegetation in the Traverse Corridor ............. 4 Presettloment Vegetation in the Traverse Corridor ............. 5 Location of the O'Neill Site, Including Offshore water Depths .. 6 Map of the O'Neill Site Showing the Location of'Excavation Units ....................................................... 7 Stratigraphic Profiles of the O'Neill Site .................... 8 Stratigraphic Profile Across Area A and Area B of the O'Neill Site ........................................................ 9 Cumulative Comparison of Body Sherd Frequency in Area A of the O'Neill Site ............................................ 10 Cugmlztive Comparison of Body'Sherd Frequency'at the OfNeill 1 ........................................................ 11$ Mackinac Ware From the O'Neill Site ........................... 12 thkin c ware From the O'Neill Site ........................... 13 Rim Profiles .... 14 Skegomog Ware From the O'Neill Site ........................... 15 Rim.Profiles .................................................. 16 Juntunen Ware From the O'Neill Site ........................... l7 Juntunen ware From the O'Neill Site ........................... l8 Juntunen ware From the O'Neill Site ........................... l9 Juntunen Ware From the O'Neill Site ........................... LIST OF ILLUSTRATIONS 10 15 30 33 37 S 56 58 62 68 7o 72 75 LIST OF ILLUSTRATIONS (cont.) Figure 20 Juntunen ware From.the O'Neill Site ........................... 21 Rim.Profiles .................................................. 22 Traverse ware From the O'Neill Site ........................... 23 Protohistoric Ceramics From the O'Neill Site .................. 24 Restored Unclassified vessel From the O'Neill Site ............ 25 Map Indicating Areas with Greater than Ten Percent Distribution of Mackinac and Skegomog wares ................. 26 Map Indicating Areas with Greater than Ten Percent Distribution of Juntunen and Traverse Wares ................. 27 Map Indicating Areas with Greater than Ten Percent Distribution of O'Neill Curvilinear, and Trailed Ceramics ;;; 28 Cores From the O'Neill Site ................................... 29 Cores From the O'Neill Site ................................... 30 Cores From the O'Neill Site ................................... 31 32 33 3h 35 .36 37 38 39 to #1 cores Front-0116 O'Neil—1.51136 ..........OIOOOOOOO0.00.00.00.00... Cores Frm the O'Neill site 0.0.0..........OOOOOOOOOOOOO0...... Percent Frequency of Flat Flakes by Length at the O'Neill Site 0.00....00.000.00.000.......OOOOOOOOO0.0.0.0.... Projectile Points and Gravers From the O'Neill Site ........... Projectile Points and Drills From the O'Neill Site ............ Lanceolates and Triangular Knives ............................. Ovates, End Scrapers, and Miscellaneous Bifaces ............... Choppers and Miscellaneous Bifaces ;........................... weight Distribution of Retouched Flakes at the O'Neill Site ... Diagram of Options in Core Manufacture ........................ Cumulative Comparison of Flake Type Frequency at the O'Neil]. Site 000...... ......OOOOOOOOOOO....OOOOOOOOOOOOOOO... viii Page 77 81 83 88 92 98 100 103 114 117 119 124 130 134 139 14? 1A9 151 155 162 167 LIST OF ILLUSTRATIONS (cont.) O Figure 42 43 44 45 46 47 5o 52 53 55 56 57 58 59 Cumulative Comparison of Flake Type Frequency in Area A of the O'Neil-l Site 0.00.0.0.......00......00.0.00...0.000.000.0000. Pecked and Ground Stone ....................................... Pecked Stone .................................................. (hmmdsmme.unn.u.n.u.u.u.n.u.u.u.“.u.u.n.u.. Ground Slate and Shale ........................................ Native C0pper and Prehistoric Pipes ........................... werked Bone ................................................... Worked Bone ............. .. European Trade Items .......................................... European Trade Items .......................................... Location of Features in Area.A, Occupation Zone I ............. Location of Features in Area A, Occupation Zone II ............ Location of Features in Area A, Occupation Zone III ........... Location of Features in Area B ................................ Map Showing Location of Sites Used in the Analysis ............ Map Illustrating Pattern A Diffusion in the Traverse corridor 0.0.0......................OOOOOOCOOOOOOOOOO00...... Map Illustrating Pattern B Diffusion in the Traverse corridor ......OOCOIOOOOOO......OOOOOOOOOOOO00.000.000.000... Map Illustrating Revised Corridor Diffusion Mbdel ............. Page 169 173 176 181 183 193 196 198 203 209 225 228 232 236 '272 304 306 311 Table kn \OmVO‘x 10 14 15 16 17 18 19 20 LIST OF TABLES Body Sherd Frequency, Weight, and Mean Weight by Area and Occupation Zone at the O'Neill Site .................... Ceramic Vessel Density at the O'Neill Site ................... Flake and Raw Material Frequency by.Area and occumtion zone 00.0.0000...0.0.0.0000.........OOOOOOOOOOOOO Length and Diameter of Block Cores from the O'Neill Site ..... Comparative Measurements of Bipolar Cores from the O'Neill Site .........OOOOOOOOOOOOOOOO........OIOOOOOOOOOOOOO0...... Attributes of Projectile Points from.the O'Neill Site ........ Flake Type and werking Edge of Retouched Flakes .............. Hammerstones from.the O'Neill Site ........................... Unmodified Shale from the O'Neill Site ....................... Pounds of Usable Meat Per Category at the O'Neill Site ....... Features in Area A, Occupation Zone I ........................ Features in Area A, Occupation Zone II ....................... Features in Area A, Occupation Zone III ...................... Features in Area B ........................................... Features in Test Units ....................................... Two by Two Contingency Table for the Calculation of Phi ...... Key Style variables -'Wycamp Creek Site ...................... Key Style variables - Ponshewaing Point Site ................. Key Style variables - Eagle Island Site ...................... Key Style Variables - weed Site .............................. x Page 42 106 108 115 121 132 153 177 185 218 223 226 229 233 237 273 279 280 281 282 LIST OF TABLES (cont.) Table 21 Koy Style variables - O'Neill Site .......................... 22 Key Style Variables - Fauver Site ........................... 23 Key Style variables — Skegomog Point Site ................... 24 Lip Area, Direction of.App1ication in Early Phase Sites ..... 25 Lip Area, Type of Tool in Early Phase Site .................. 26 Motif Variables, Early Phase Sites .......................... 27 ‘Exterior Decoration, Tool Type in Early Phase Sites ......... 28 Spatial Behavior of Style variables ......................... 29 Phi - Coefficient Matrix in Q‘- Mode fer Early Phase Sites .. 30 Lip Area, Type of Tool in Late Phase Sites .................. 31 Motif variables, Late Phase Sites ........................... 32 Exterior Decoration, Tool Type in Late Phase Sites .......... 33 Phi - Coefficient Matrix in Q - Mode For Late Phase Sites ... Appendix B 1 Machine Code variable Numbers for Artifacts in Area A, Occupation Zones I, II, and III Correlation Matrices ...... 2 Correlation Matrix, Area A, Occupation Zone I ............... 3 Correlation Matrix, Area A, Occupation Zone II .............. 4 Correlation Matrix, Area A, Occupation Zone III ............. 5 Machine Code variable Numbers For Artifacts in Area B Correlation Matrix ........................................ 6r Correlation Matrix,.Area B .................................. xi Page 283 284 285 289 290 291 292 293 294 296 297 298 300 323 324 325 326 327 328 CHAPTER 1 INTRODUCTION The Corridor Concept Since 1966, archaeological investigations undertaken by the Michigan State University'MMseum.and the Department of.Anthropology have been concentrated on the coastal plain of Lake Michigan between Traverse City and Mackinaw City, Michigan, which has been called The Traverse Corridor. The rationale behind this long-term program has been outlined by Cleland (1967) who describes the Research Universe as a north-south lakeside passage differentiated from surrounding areas by a complex combination of natural factors. To the west, the study'area is bounded by Lake Michigan; while to the east a major physiographic transition takes place when the hilly, rocky, terminal valders moraine is encountered. The morainic system initiates at varying distances from the Lake Michigan shore -- as near as 50 yards and as distant as 12 miles. The area encompassed between Lake Michigan and the morainic interior is:a flat, sandy, littoral (Fig. l). The coastal plain, as a result of its elevation, the proximity of Lake Michigan, and the modifying air masses which result from the latter, differs radically in both climate and vegetation from the interior. Recent studies of climatic differences in the study area have demonstrated that the littoral enjoys between 120 and 140 frost-free days, while the interior uplands can depend on less than 110 frost-free days (wagner n.d.). The coastal plain, therefore, is situated within a thermocline representing the effective limits of prehistoric agriculture (Yarnell 1964:128). This difference in climate is reflected in the natural biotic communities as well, since the coastal plain flora represents a narrow extension of the goofinaoo 022389 23 Mo on: H 93mg ‘ —— ’I”flll’l Canadian-Carolinian transition zone into what is predominantly a Canadian biome. That the latter phenomenon is not merely a contemporary situation has been clarified and substantiated by Isleib's (n.d.) analysis of presettlement forest communities within the study area. Utilizing primary data consisting of original county surveyor's notes, Isleib was able to recognize.l3 different vegetational associations within the looamile-long coastal plain and adjacent uplands. These various floral associations form.a mosaic (Figs. 2-4) but include species common to both the Canadian and Carolinian biomes, creating an ecotonal or edge effectd(0dum.1965:278). The numerous micro-environmental interfaces that result from a mosaic of this type seem, preliminarily, to have been optimal locations for prehistoric settlements (N. Nowak, personal communication). Contrasting .strongly'with the edge effect present on the Lake Michigan shoreline, however, is the predominantly Canadian community found on the Lake Huron coastal plain. Thus, Cleland (1967:2) concludes that "it (the littoral between Traverse City and Mackinaw City) is the only continuous region favorable for settlement linking the northern and southern cultural and natural areas.” Testing the hypothesis that prehistoric settlement patterns were, for the most part, confined to the naturally favorable corridor was accomplished early in the program (1967). Archaeological reconnaissance in the morainic eastern uplands has demonstrated the virtual absence of sites in this region. The few sites that were located are apparently small, short-term campsites dated to the Late Wbodland. This sparse occupation contrasts strongly'with the numerous (N=36) settlements located in the Corridor, demonstrating aboriginal occupation from the Early Archaic through Historic periods. Therefore, the data gathered to date Figure 2 Presettlement Vegetation in the Traverse Corridor 2 \ m ”I‘m-idhfilmihflh—n, -.—.h¢.I—‘d~vy in“-.. IZ‘ “HM“.dfidfl‘ fl“ flfifl’ mm.“ . mull-In- h‘.-fl‘~ 15-1.!- !- hip-dun"... 9 m ‘fi—‘fi D-fl-d—‘fl-t ll‘-_.‘~-~. hop-n LOCI-NIH z (Org UIU I T Puffin VEGETATION AS RECORDED ON SURVEYS OF 1839 0 I I l l nooflnaoo omno>wna mop 2H scavaowo> psosoappomonm m onswfim ......n III-III!!!- Ialllll..l.l.l£l .u .‘IIlaaI.ll-.= algal!!!- Illl .9 lilies-.1]. .. Eli-1.2.1.... .l-ialiulillfl’. .. .... —un 'IIOF. NHQVS VKN onc— mO m>m>~5m ZO nfiOxOUmx m< ZO:.<._.m0m> M m: w 1 w r ... m T .. m imam £2 . N u z — mix—2a max 2.. far ...:E...E .H «Farrier: ,, VIII 2 T3. 17 10 VEGETATION AS RECORDED ON SURVEYS OF 1839 11 seems to support the contention that prehistoric adaptations were funneled and highly confined to the coastal corridor. Although the Traverse Corridor has yielded cultural evidence demonstrating aboriginal utilization from ca. 8000 B.C., not until the Late Woodland (A.D. 700-A.D. 1650) did the coastal plain undergo intensive exploitation and occupation. The 29 sites assignable to the Late woodland period present a quantum increase when contrasted to the feur late Middle 'Woodland settlements in the study area. Because of infermation derived from these Mfiddle and Late woodland villages, evidence was generated that the Traverse Corridor acted as a passage fer cultural transfer of peeple or ideas. Beginning in the late Middle WOodland (ca. A.D. 400), we see a juxtaposition of typically northern archaeological manifestations with southern complexes of the same time period. Populations manufacturing Laurel tradition ceramics penetrate as far south as the Pine River Channel Site in Charlevoix, while two other occupation sites of these typically northern peoples are located north of this point, towards the Straits of Mackinac. Likewise, Havana tradition penetration is documented as far north as the Pine River Channel Site; the only other Havana ceramic- bearing occupation in this area is located south of the town of Charlevoix. No "intermediate” assemblages have yet been recovered. The nascent north-south bipolarity of the Middle Neodland period, as exhibited through ceramic expression, is a long-lived phenomenon which continues at least through the fifteenth century.A.D. The Skegemog Phase (A.D. 900-A.D. 1200) of the south is roughly contemporaneous with the northern Mackinac (A.D. 700-A.D. 1000) and Bois Blanc (A.D. 1000-A.D. 1200). Both Mackinac and Skegemog Phase ceramics maintain a certain 12 impression of'similarity characteristic of early Late'WOOdland Ceramics in Michigan, whereas the pottery of the Bois Blanc Phase represents a noticeable style horizon. Bois Blanc may, indeed, need reappraisal due to itsilimited distribution (Egg,”Ceramic" discussion). In later periods, we see Traverse Phase peoples (A.D. 1200—A.D. 1450) juxtaposed to populations manufacturing Juntunen ceramics. With the exception of scattered materials from the'wycamp Creek Site and the prehistoric component at Fort Michilimackinac, our knowledge of the material culture of the protohistoric inhabitants is minimal. These data demonstrate that some overlap in the spatial distribution of the northern and southern style traditions does occur - the extent of this area of mutual use had not yet been investigated. This general pattern, the penetration of the Corridor from both the northern and southern termini by peOples practicing different ceramic traditions, continues into and through the Late'weodland period. Not only does the filling of a favorable natural vacuum minimize the possibility of spatial isolation of Late Woodland populations, but, furthermore, cultural traditions practicing different adaptive strategies are brought into close proximity. These stratagies have been summarized for the Historic inhabitants, which in turn enabled generalizations pertaining to Late woodland adaptations to be formulated (Fitting and Cleland 1969). It has been argued that the northern end of the Corridor was inhabited by groups practicing a hunting/fishing/gathering subsistence base, whereas the southern, terminus populations were supplementing a similar strategy with the products of domestic agriculture. Thus, the contrasts documented for natural communities and ceramic style traditions seem to carry over to subsistence practices as well. This model, however, still needs to be rigorously'tested. 13 As mentioned previously, one of the primary goals of the Traverse Corridor Project is clarification of temporal and spatial relationships of the two Late woodland ceramic traditions recognized in the study area -- a problem that had not yet been approached. In order that these relationships be defined, it was necessary to possess prOper temporal and spatial control from.specific sites in the study'area. Few of the Late 'woodland sites located could provide the necessary temporal separation -- stratification of occupations is rare, while superposition and mixing of occupations in an undifferentiated soil zone is common. Testing Operations performed at the O'Neill Site in 1969 revealed that it was best suited to answer the questions posed. A partially stratified site which thereby provided the necessary temporal separation, the O'Neill Site contained village material from the early Late Woodland through proto- historic periods, thereby providing a complete Late'Wbodland sequence; furthermore, the site was located in the approximate center of the study area. Its potential for examination of the initial problem not only'far exceeded that of other sites considered but was augmented by both excellent faunal preservation in the stratified areas and the presence of house structures. To collect an adequate sample from the site, major excavations were planned for 1971. The results of these excavations are presented in the following sections of this report. History‘gf’Investigations The O'Neill Site (zoom) is located in the NE-t- of the SE11; of Section 7, T‘33 N, R 9 w, about 3 miles south of the Pine River in Charlevoix County, Michigan (Figure 5). The site was discovered by field crews from the Michigan State University'Nbseum.engaged in a settlement pattern study 14 Figure 5 Location of the O'Neill Site, Including Offshore water Depths 15 O '14 9* I SCALE IN MILE". ')("JTH MHNT LN \ x‘ \‘ \\1 J 16 of the area between Traverse City and Mackinaw City, Michigan. Under the direction of Dr. Charles E. Cleland, this survey was partially financed by a National Science Foundation Undergraduate Research Participation Grant (GT-2825) and National Science Foundation Research Grants (GS-1026; GS-1669) entitled "Prehistoric Cultures of the Grand Traverse Bay'Area, Michigan.” The area of the site brought to the attention of the investigators was located on a Nipissing beach with an elevation of 610 »feet-612 feet a.s.1. The elevation, coupled with the lack of surface ceramic materials, prompted a preliminary'Archaic designation for the site. I fjnguring the summer of 1969 Dr. Cleland returned to the O'Neill Site with a crew of six Michigan State University students to initiate test excavations. The first season's field crew consisted of Richard Clute, William Lovis, Jeffrey Tordoff, Donald weir, and Daniel wymer. Primary interest at this time was focused on the Archaic occupation. Indications of a Late woodland occupation were subsequently noticed in a disturbed road cut connecting Lakeview Drive with the beach, at elevations ranging between 580 feet and 595 feet. The decision was made to test excavate the Late woodland component as a secondary phase of the Operations on the Archaic occupation. Due to the success of these initial testing procedures, 1750 square feet of the woodland component was excavated that season. These investigations revealed a partially stratified Late WOOdland village site, containing ceramic materials of the Mackinac and Juntunen Phases (MbPherron 1967a) and the protohistoric period. Following the 1969 operations, the O'Neill Site was nominated to, and eventually placed on, the National Register of Historic Places. During ” the winter of 1969/70 a Detroit, Michigan, development company revealed plans fer a resort subdivision on the site area. The U.S. Park Service 17 responded to this threat to both the natural and cultural components of the shoreline by reconnending that these and adjoining lands be brought under Federal ownership. Plans were proposed for an interpretive reconstruction of the site as a major component of the National Woodland Indian Memorial. These developments prompted a grant request under the National Historic Preservation Act to pursue excavations during the summer of 1971. This grant was awarded to the principal investigator, Dr. Charles Cleland, through the Michigan Department of Natural Resources (Project No. 26-71- 00025). From 21 June until 12 August, a crew of seven students under the direction of the author, excavated 2100 square feet of the site, with a concentration in the stratified areas of the site; an effort to determine the limits of the occupation was also pursued. The second season's crew included P. Fisher, J. Kimball, J. Burdick, C. Nern, N. Nowak, B. Reed, and J. Tordoff. These field workers were occasionally supplemented by visitors; Dr. Cleland was present for most of the field season. A preliminary report on the second season's work was prepared for the Michigan Department of Natural Resources (Iovis and Cleland 1971). Glacial Geolog 39g Topography Discussion of the geologic glacial sequence in the Grand Traverse and Mackinac Straits region relating to aboriginal occupation of the area commences with the glacial substage known as the Valders Maximum. At this time the water levels of both Lake Michigan and Lake Huron rose to an elevation of 605 feet a.s.1. forming Early lake Algonquian (Kelley and Farrand 1967: 14-15). Retreat of the Valders ice mass ca. 11,000 B.P. freed the area, at which time resultant rises in lake level back up to the ———~_.’-—-——«—— 18 605 foot level submerged the coastal regions with waters of glacial Lake Algonquian. The fossil beaches of glacial Lake Algonquian are visible in Charlevoix County both inland from the present Lake Michigan shoreline and on Lake Charlevoix; they generally attain an elevation of about 690 feet a.s.1. as a result of isostatic rebound. Following this Algonquian high- water stage of ca. 9500 B.P., continued retreat of the Valders ice margin freed the Trent Lowland of Georgian Bay. The Michigan basin drained via the Straits of Mackinac into the Huron basin, and then out the sea level drainage at North Bay, eventually reaching the lowawater Lake Chippewa stage. During the drOp frothake Algonquian to Lake Chippewa-Stanley, a number of brief, dropping lake stages are recorded by fossil beaches. These “upper group” beaches are not visible in the immediate vicinity of the O'Neill Site, although several beaches of this group occurred at elevations ranging between 540 and 390 feet a.s.1. Finally, Lake Chippewa-Stanley'reached its lowest level of 230 feet a.s.1. ca. 9500 B.P. Throughout the post-Valders sequence the land was rebounding from the removal of the weight of the ice, with rebound taking place along a generally north to south axis. Crustal rebound terminated the drainage of Lake Chippewa-Stanley by raising the level of the North Bay outlet above water level which initiated a renewed rise in water level. The impounded waters rose until the levels of the old outlets (St. Clair River and Chicago) were reached, at which point they stabilized at an elevation of 605 feet a.s.1. Drainage also continued through the North Bay outlet until, once again, rebound lifted it above lake elevation. Lake Nipissing, as this 605 foot stage is called, reached its maximum at about 4200 B.P. In the Charlevoix area this stage is represented by fossil beaches at elevations of ca. 610 to 612 feet a.s.1. Continued drainage through the l9 southern outlets caused down-cutting of the glacial till in the St. Clair River, dropping the lake level to 595 feet a.s.1., the Algoma stage. The current topography of the O'Neill Site area suggests that from as early as post-Algonquian upper group times, the Inwood Creek drainage area was a series of rolling and sometimes steep dunes. A similar situation was apparent at the Mt. MoSauba Site north of Charlevoix, where steep dunes were the only indication of the Nipissing lake stage (Charles E. Cleland, personal communication). Furthermore, the Archaic Inwood Creek Site (southeast of the O'Neill Site) seems to have been located on a sandy Nipissing stage beach. During modern times and as early as the Nipissing stage, these dunes ran parallel to the beach, occasionally cut by either wind or water eroded gullies formed at right angles to the shoreline. The stratified portions of the O'Neill Site attest to the presence of a similar situation as late as A.D. 1500, while the occupied areas to the east of Inwood Creek corroborate these observations. Soils and Vegetation The O'Neill Site is geographically positioned within what Cleland (1966: 5, Fig. 1, p. 6) characterizes as an ecological transition zone between the Carolinian and Canadian biotic provinces (Dice 1943). Despite its position at a microenvironmental interface, however, the transition zone itself retains a distinct identity'due to diversity of the over- lapping communities it contains. This identity has been described as an ecotonal or edge effect (Odum 1965: 278) and is quite visible in the natural communities within which the O'Neill Site is located. Soil surveys of Charlevoix County were undertaken during the years 1940 and 1949 under the cooperative auspices of the Agricultural 20 Experiment Station, Michigan State College of Agriculture and Applied Science, Conservation Institute; the Soil Science Section, United States Department of.Agriculture, Division of Soil Survey; and the Bureau of Agricultural Economics. Unfortunately, a report of their survey'was never published. A County soil map and key are available for inspection, however, and the information presented here is drawn from these sources (Schneider 1949). [within 1 mile of the Lake Michigan beach, moving inland, one rapidly crosses three soil zones which run almost parallel to the shoreline. The stabilized dunes on which the O'Neill Site is located are dominated by Eastport and Bridgman sands. These well-drained soils, which generally occur on hilly slopes with a gradient of 15 per cent, presently support a second-growth vegetation of aspen, white birch, oak, jack pine, and some hardwoods (Schneider 1949: 10). As one moves further inland, highawater, tabletsoil associations are rapidly encountered, in particular the Carbondale muck—Rifle Peat association and the Detour loam-Munuscong loam association. Both of these soil associations are generally found on slopes of less than 3 per cent. The former presently supports second- growth black spruce, white cedar, ash, elm, balsam fir with tamarack (Schneider 1949: 9), while the latter is predominantly second-growth aspen, black spruce, white cedar, and balsam, although some elm, ash, and hemlock also occur. Moving to higher elevations these soils finally grade into an Emmet Lloam-Onaway loam-Detour ~loam association which occurs on hilly areas with less than 7 per cent gradient but which contain occasional, short escarpments. Second-growth forest cover in these areas includes aspen, sugar maple, elm, ash, with some white pine, hemlock, black spruce, cedar, and balsam (Schneider 1949: 3). 21 The above soil and vegetational associations indicate that the O'Neill Site is located at a modern microenvironmental interface and that the inhabitants had access to a variety of faunal and floral resources. Isleib's analysis of the composition of the presettlement forests in Emmet, Charlevoix, and.Antrim Counties (n.d., Map 1) also supports this contention for the primary vegetation. Her data, drawn from the original county surveyor's notes, demonstrate that the O'Neill Site was located at the juncture of three microenvironmental zones. North of Inwood Creek and bordering the shoreline, the plant community was composed of cedar, tamarack, and hemlock with some spruce, fir, black ash, alder, and white pine. To the south of the creek, the coastal flora consisted of hemlock with white pine and elm, while the interior supported a community'of'sugar maple, beech, and hemlock with minor amounts of white pine, elm, basswood, birch, ironwood, and cherry. This diversity of floral communities illustrates the variability present in an ecotonal area. The area of the O'Neill Site may be generally described as a mixed forest due to the mosaic effect evident in both the primary and secondary communities. That human habitation occurred in these extant microenvironments cannot be merely fortuitous but must be viewed as a maximization of potential subsistence with respect to edge area diversity. Fauna As a result of the transitional nature of the environment on which the O'Neill Site is located, an interesting admixture of mammalian fauna may be observed. These have been described by Burt (1954: 1957), Cleland (1966), and Dice (1925). For the transition zone between the Canadian and 22 Carolinian biotic provinces Burt (1957: 3) indicates l6 northern species with their southern limits in the ecotone, while 17 southern species have northern ranges in the area. Today, only 28 species are found in the transition zone (c.f. Fitting 1966: 143-144). Conspicuously absent from the modern lists are some larger Canadian mammals such as the elk and moose. The relative abundance of avian fauna in Charlevoix County'may in part be due to its location along a minor tributary of the Mississippi Flyway (Hawkins 1964: 185, 187). Van Tyne (1923) and Wood (1951) list 118 species that occur in modern times in the vicinity. Of these, the number of available game birds has varied through time both in respect to the species utilized by various inhabitants of the site and those which might be included in modern game lists. This will be explored more fully in the faunal analysis. In recent years some migratory aquatic species have wintered in the area. Local informants claim.that this phenomenon is due to the increased amounts of open water made available by Coast Guard vessels during winter months. The availability of aquatic resources is dependent on the interplay of a number of variables. water temperature, depth, gravel spawning-bed availability, direction and speed of currents, and the availability of both mollusks and plankton are some of the major factors to be considered. In terms of these variables the O'Neill Site and adjacent Lake Michigan shoreline bear a remarkable resemblance to the lacustrine environment described by Brose (1970a: 13-15) for Summer Island. The shallow Antrim.1ow Shoretype (Humphrys, et. a1. 1958: 1), with its sand and gravel.bottom, has a water surface temperature ranging between 15 and 16 degrees centigrade (Ayers, et. a1. 1958: 21). Surface 23 currents move slowly in a north northeast direction out of Grand Traverse Bay (Ayers, et. a1. 1958: 30). Even these slow currents are probably moderated by the shoal areas surrounding Fisherman's Island (Fig. 5), thus sheltering the bay into which Inwood Creek drains. These factors, coupled with an abundance of small, distinctly northern mollusks that play an important part in the diet of bottom-feeding fish such as sturgeon (Walker, 1896: 96-97), encourage an abundance of those species preferring shallow, protected coastlines with sand and gravel bottoms. It should be noted, however, that rocky, unprotected, but shallow, shores occur both north and south of Inwood Creek. In summary, the O'Neill Site lies in a rather rich natural area. The ecotonal vegetation provides a variety of microhabitats for terrestrial mammals, while both the aquatic and avian fauna supplement these resources starting in the early spring. Peak resource availability would therefore occur in the late spring and early summer with the return of migratory waterle and the beginning of the spring spawn. Elisa—ts. The popularity of the Charlevoix area as a summer resort is the product of many natural factors. Aside from obvious variables such as its proximity to Lake Michigan and its location on the inland lakes, its moderate climate is also an appealing and desirable asset. Data recorded at the Charlevoix weather station (U.S.D.A. Yearbook of Agriculture 1941: 727-728, 914-924) for the years 1899-1938 indicate that radical extremes in temperature occur but generally the climate is rather moderate. For example, the maximum temperature recorded for these years was 99 degrees F., while a minimum of minus 33 degrees F. was recorded. The average January 24 temperature was a not too extreme 21.7 degrees F., rising to a moderate 68.2 degrees F. average during July, Snow cover, however, generally remained fer a minimum of 120 days, a product of the 70 inches of snow that falls annually, Average annual precipitation is 28.47 inches. In striking contrast to the morainic interior, the coastal plain on which Charlevoix is situated enjoys a long growing season of 157 days. 'wagner (n.d.) indicates a probability of 100 per cent for a 120 day season, the sufficient minimum.for corn agriculture (Yarnell 1964: 136). A primary natural variable responsible for this situation is the wind direction during the growing season, which is predominantly from the north and west, thus dealing a major moderating lake influence on the coastal areas. Descriptions; 9; Excavations Survey ‘With the initiation of excavations at the Archaic component of the O'Neill Site, a point of origin for the superimposed Cartesian grid system was chosen. This center point, designated NO/EO, was located on the northern periphery of the site area, atop a raised knoll free of short vegetation which allowed sightings to take place with a minimum.of brush clearance. From the origin a north-south and a eastawest transect were laid out utilizing a surveyors' transit, stadia rod, and 100 foot engineer's tape. Ten-by-ten foot excavation units encompassed by the grid were shot in from these transects and designated by the location of their southwest stake in relation to the center of the grid. Thus, a unit whose southwest corner was located 30 feet south of the origin and 20 feet west was labeled unit S30/w20. When investigation of the Late Woodland 25 component was begun, the north-south transect was extended north to the beach, with surveyor's.stakes placed at 50-foot intervals, again using the methods stated above. Since the N350 stake fell in the center of the known site area, a second east-west transect was shot from this point with surveyor's stakes located at lO-foot intervals along the line. Once again excavation units were labeled according to their distance and location from the origin. All Late woodland component excavations, however, were located north of the center reference stake. The O'Neill Site was mapped with a transit and stadia rod while excavations were in progress, utilizing several stations whose elevations above lake level were recorded during both the 1969 and 1971 field seasons. Their true elevations were then calculated from tables provided by the water Level Section, United States Lake Survey, for Lake Michigan. Excavation StrategyiapdpArtifact Recovegy Testing procedures carried out during the 1969 field season at the O'Neill Site were concentrated in two areas of the site, one deeply stratified and the other bearing a number of cultural features as well as portions of a large structure. Preliminary analysis of the recovered materials posed a number of questions which it was hoped could be answered by further excavation. Due to the lack of surface evidence in the form of artifacts or soil features, many of these questions were still of an exploratory nature. 1. What are the limits of the occupation area? 2. What are the intrasite relationships of the cultural strata? 3. Are the cultural strata products of single or multiple occupations? 26 4. Was more than one structure present at the site; if so, how many more? 5. How did activity areas relate to the structures? Generally, these problems were investigated through a tripartite strategy. Problem 1 was handled by a systematic testing Operation parallel to the shoreline between Lakeshore Drive and the beach and was carried out in a manner that tested both the ridges and gully areas of the site. Supplementary core samples were also performed to extend these test investigations. Clarifications Of problems 2 and 3 were accomplished through the placement of a 5-by—6O foot linear block connecting the stratified and unstratified parts of the site as well as through a continuation Of'block excavation in the deeply stratified zones. The last two questions were approached by continuing investigations in the structure and feature bearing portions Of the occupation. All Of these operations were hindered by the heavy vegetation present over some parts of the site which at times dictated the placement Of excavation units. Sampling Of the site was accomplished with excavation of 5-by-5, 5-bye 10, and lO-byhlo foot excavation units, while vertical controls within the units were dependent on both the stratigraphic sequence and the thickness of each stratum, Humus zones were removed in one unit, unless it exceeded 0.3 feet in depth, in which case another level was begun. Arbitrary 0.3- feet levels were utilized in the sandy subsoil below and between occupation zones. As each vertical provenience unit was closed within a given unit, depths were recorded for all corners. In the case of sloping or uneven occupation floors, depths were recorded for all discontinuities. The base of each level and the tOp Of all occupation zones, were troweled clean, 27 and color positive 35 mm slide photographs:were taken. The cleaned and photographed floor was then mapped; all color differences, hearths, pits, and postmolds were plotted on square sheets. In the case of discrete occupation strata, all artifacts on the floor were plotted and photographed in situ. Cultural features such as hearths and refuse areas were collected as separate provenience units. Half bushel samples of soil, or smaller but proportional-sized samples, were collected from these features and subjected to flotation. Pottery and flint concentrations were plotted and excavated as features, although no float sample was collected. Soil stains suspected of being post molds were cross sectioned to sustain their validity'and, when verified, were recorded on the square sheet fer that level. Upon reaching sterile layers, the walls of all units were cleaned, troweled, and photographed, after which they were drawn in detail with a line level, plumb hob, and engineer's tape. All excavated soils except fer flotation samples were screened through tainch hardware cloth, and each unit was terminated when no more cultural materials were gathered. A11 excavation was performed according to the above criteria until sterile sands or gravels were encountered. various units were subsequently'excavated to much greater depths as a check on premature termination of the square. Additional data was collected in the form of soil samples and pH tests from widely varied parts of the site. The soil samples are still in the process of being analyzed. During the 1969 field season, 1750 square feet Of the site was excavated by six people in a four week period. In 1971 a crew of eight excavated fer eight weeks to remove 2100 square feet of earth. A total of 3850 square feet of site area, therefore, was excavated to an average 28 depth of three feet. Thus, only slightly less than one man hour was expended fOr the excavation of 3 cubic feet of site area. Stratigraphy‘gpg.Dating The stratigraphy of the O'Neill Site did not allow surface collection of cultural materials in areas other than disturbed secondary'road cuts that penetrated cultural soil strata. Additionally, the shape, depth, and thickness of the various soil zones changed consistently across most of the occupied parts of the site. A single stratified occupation sequence was located.within the block excavation labeled Area A (Fig. 6), while the part of the site called Area B had predominantly undifferentiated occupational strata. Test Pit investigations revealed both stratified but unoccupied areas near the lakeshore and undifferentiated occupations fer excavations further inland. Radiocarbon dates from Area A and Area B, coupled.with the artifactual material, allow us to reconstruct the depositional sequence on the parts of the site most intensively utilized by the Late WOOdland inhabitants. Alas—A Due to the stratification of cultural layers revealed in Area A, this part of the O'Neill Site presents us with a more complicated depositional record than excavations on other parts of the site. Overlying parts of Area A are loose, wind-blown sands which have not yet become stabilized and which cover the forest floor with up to 0.5 feet Of still active sands. This sand layer did not bear artifactual material. TherefOre, this layer was shoveled off until the buried humus zone was exposed. The deep gray humic zone, which was composed of‘loamy sands ranging between 0.3 feet and 1.0 feet in thickness, graded gently upward from an elevation Of 588.5 29 hoes: seasoneooe .8 eofieooa can monsoon Ben 382.0 one. no on: m ondmflm 3O - utn 4...”..0 31 feet a.s.1. to 590.5 feet a.s.1. Aboriginal materials and European trade items recovered from this layer have been comparatively dated to the late seventeenth century, although some modern cultural remains also occur. This soil layer was designated Occupation Zone I. The east part of Area A differed slightly in the deposition of this occupation stratum, A light gray sand zone branches downward from the humus zone to a depth of 1.0 feet, and there is an intervening sand stratum.that separates the two. This soil layer was labeled Occupation Zone Ib. Immediately beneath the humus zone on the west part of.Area A is a black occupation zone which lenses out and.disappears as one moves south toward Area B. The same cultural stratum on the east part of Area A has an intervening windpblown sand stratum separating it from the humus zone. As this soil layer was followed across the site, it was designated Occupation Zone 11. This surface was essentially level across Area A. Underlying Occupation Zone II is a layer of water-layed white sands of varying depth that contain thin black laminae. This intermediate, culturally sterile stratum.is a plano convex in profile; the upper plane is formed by the base Of Occupation Zone II, and the bowl-shaped bottom section is directly above another occupation stratum. The interpretation Of water-deposited sands is supported by the eroded appearance of ceramics recovered from the basal occupation stratum. The above sequence is illustrated in Figure 7A. McPherron (1967a: 34, citing wright) and Mason (1966: 50-52) in reference to similar water-deposited sand phenomena at the Juntunen and Mero Sites attribute this to storm beaches rather than a major rise in lake levels. This is one explanation. A second explanation, which may well be unique to the O'Neill Site, would pertain to current occurrences 32 mpfim whoa oaozmoxmq HO hommawflpmnpm HmOHohe I m e cone to acoeaweoesom . a seam Haeoz.o one no hoaaeoao oeeoeeweeeaom a cheese 33 4.21.5 024m 2.; g 02¢ 02%. >45 EOEEEEM 2.: 233a air—U $sz 22:: 550-55 233' 024m 55 E03! ozqm 23 93 «343mm. mi... 58g . .. 2 833.2 m_xoON mtm u_.:mZ.O 34 that might possibly have had a prehistoric analog. Throughout both season's excavations at the O'Neill Site, the outlet of Inwood Creek into Lake Michigan was observed to shift almost daily, an occurence dependent on wind direction and wave action. Quite often the outlet is small with a sluggish water flow. In this event, a backwater pond forms, always in the same area, a hollow behind the beach crest. Even when waters are not impounded, the beach is wet and mucky. Thus, little modern utilization of this part Of the beach occurs; a situation that is identical to that found in Occupation Zone II. McPherron reports similar occupation-density decreases for those occupational layers overlying the flooded parts of the Juntunen Site. The basal occupation layer, which ranges between 0.05 feet and 0.30 feet in thickness, reflects intensive use by Late woodland populations. This zone is actually a series Of thin, gray-black sand layers that meet and diverge across the floor. Separation of discrete layers was not possible; the entire layer was designated Occupation Zone III. Two radio- carbon dates were Obtained on charred wood from a hearth on this occupation floor; A.D. 1210:100 (Ma2406, unpublished) and A.D. 12901100 (Ma2405, unpublished). These are compatible with artifactual material from this layer. A single date of A.D. 1115 (N-1268, unpublished) was received on charred wood from 0.3 feet below Occupation Zone II in the water-deposited sand stratum. ‘Considering its position above Occupation Zone III, it may have been deposited by human or natural agencies or may even have been contaminated by the inundation of this strata. In either case, it is clearly not compatible with the stratigraphic position of Occupation Zone II. 35 A_rea .12 The stratigraphic sequence of Area B is essentially the same as that reported for the upper layers of Area A. An overlying layer Of tan wind- blown sand, a dark gray loamy humus zone, and an underlying occupation strata of black sands rest on a yellowetan sand subsoil. However, variations do occur. For instance, parts of the north portion of Area B do not have an occupation zone. When viewed in profile, it assumes the shape of a truncated dune crest (Fig. 8). A thin sand stratum occasionally intervenes between the base of the humus and the occupation strata on the western part Of Area B. Further variation is evident in the linear profile block. Here, a second, thin gray, undulating stratum was located beneath the occupation zone, and separated by a layer of yellowhtan wind- blown sands. This lower stratum probably represents a stage Of light grass cover stabilizing the back beach dunes. Radiocarbon determinations were Obtained on two of the above layers, the occupation zone, and, f indirectly, the basal sand layer. The latter was Obtained on a hearth, Feature 3, which was isolated in the basal sands. This date of A.D. 10001140 (Me2401, unpublished) indicates that Area B was still in an unstable dune situation until this time. Ceramic materials in the vicinity of Feature 3 corroborate the validity of the Carbon-14 date. This age Of the undifferentiated occupation strata in Area B was based on a Carbon-l4 date of A.D. 14551100 (Ma2398, unpublished). Stabilization of this zone had probably occurred by this time. Furthermore, this date may well correspond to the intermediate Occupation Zone II in Area A. 36 even HO O O OH OH H9 0 HH . Zoo 39 mo m eon < one a eoa< mmoa emoa . nH .oaenm . we a pm m onowam 37 .o e V 38:: \ ZO_h(>U4u d): \k b! mung can! be m 93 nod :39 I .g ......... ..o I .- I \“\.0 one . .---././ L «00 . QQ\\\\\\\‘\\\\\.fi\\\\.\\t\\i\.\\.\\\ . A \\\\\\\\\\\\\\h . En: 3.1.4.6,r In-.. . . _ com 2 0mm z 032 one 2 buwu 2. mg 1 wzfi Om kmwg 020.3 m 024 < mdmmd mmomod wmiomo 1.5102 0» IHDQT w_ x0 ON mtm .jQZaQ 38 Isaac. Two types of stratigraphic variation were recorded for various test pits placed to the east of our block excavations at the O'Neill Site. Mest Of our investigations encountered a humus, occupation, yellowetan sand sequence similar to that recorded in Area B. As we moved toward the beach, however, more complex stratigraphy and fewer occupational debris tecame apparent. A typical profile from these test pits is presented in Figure 7B. Several differences may be seen between the strata in Area A and the outlying test pits. First, groups of thin gray sand strata separated by tan blow sand are present in the position of the humus zone. NO water- laid sand layers were exposed. Major buried humus zones were first recognized at a depth Of about 2.0 feet. Explanation Of the thin, gray sand strata groups are based on Observations of the site during autumn and winter. During the fall, the grass cover on the dune-forest edge becomes covered with leaves, and the grasses die and begin to decompose. ‘Winter snowoover then buries these organic materials. Then sand, blown from the exposed beach, covers the snow. As the spring thaw melts the snow, this sand cover is deposited on top Of the dead organic materials from the preceding year, thus Sealing the stratum. As time passes, some Of these layers merge, as suggested by the increasing thickness of the layers as one excavates deeper. Occasional major deposits of wind-blown sand are attested to by the upper soil layers illustrated in Figure 7B. we could not accurately correlate the deeper humus zones from these beach excavations with either inland test pits or the Area A and Area B excavations. ‘we can, however, state that a complex history Of dune migration, stabilization, and the.repetition Of these two phenomena 39 occurred until the modern forest became established. Further research which may clarify this situation might employ a back hoe to transect the site from Lake Shore Drive to the beach. Depositional History The depositional history of the O'Neill Site is a story of dune formation through both wind and water action with the ultimate stabilization of the dunes caused by these natural factors. At about A.D. 1000 those parts of the site called Area B were partially stabilized sand floors, bordered on the lakeward side by deep, grass-covered dunes running parallel to the beach. Occupaticnrof.Area B and to some extent Area A during this period of unforested succession is demonstrated by occupational debris on these floors. Sometime prior to A.D. 1200, some further stabilization of the dune gullies exposed in Area A took place. At this time a series of occupations of both Area A and Area B occured. Following the stabilization and utilization of this part of the site, Area A was inundated by either unusually high.water stages, possibly caused by storm activities or impounded waters due to the inhibited out- flow of Inwood Creek into Lake Michigan. As water-laid sand layers accumulated, this low area of the beach was not used by the prehistoric inhabitants; when reoccupation did begin, it was still very limited. This second series of occupations in Area A was covered by a layer of wind- blown dune sands. By the fifteenth century, however, Area B had stabilized and was reoccupied along with Area A. The dune crest separating these parts of the site was still subject to wind action which prohibited stabilization and the formation of a discrete occupation strata. Continued occupation of the site, combined with deposits of sand and organic materials, covered the earliest humus zone with a modern sand loam. #0 This process of sand deposition and stabilization still occurs; in Area A at least, the final stages of stabilization took place during the seventeenth century, when the site was visited by Historic Indian groups. CHAPTER 2 CERAMICS The 3850 square feet of the O'Neill Site which was excavated over the course of two field seasons yielded 9194 ceramic fragments. These materials were analyzed in the laboratory in a number of different ways. Separation of the sherds into gross categories based on the presence or absence of decoration and whether or not it was a rim fragment was the first step in the process; this group included 777 shards. These shards were then sorted and combined until, with reasonable certainty, we had combined all shards representative of single vessels; the result was a minimal vessel count of 80. The body'shards were treated in a somewhat similar fashion. Initial stages in the processing of these materials were to sort out all those shards too fragmented or too weather worn to discern surface preparation. This excluding process left a sample of body sherds which could be classified on the basis of the amount and type of surface texturing. All of the resultant subdivisions were counted and weighed, and mean shard weights by excavation area and Occupation Zone were calculated for both the identifiable and unidentifiable sherds. These data are presented in Table 1. Few body sherds could be linked with rims and decorated sherds due to the rather striking similarities in surface treatment among the different vessels in the sample. No evidence of coiling was present in the sample studied. Manufacture ‘was probably done by patching or through arpaddla and anvil technique. There were no repair holes in any of the shards, and only five minimal vessels displayed any cooking residue on the interior of the rim. Tampering particles other than crushed granite were not present in this #1 1+2 ..m @4de lam 3H 06 On 0.: m.m ma m5 ha in Emacs a mmnom i8 mm mama m3 SH wHH Nam amHH ma Emacs Sam «How a mRH mNH w: R Ham eon mH 0.58 332. H.H .2 9m «a 9m H.m ca as“ H.H ...... Eat; a “HHN NeHH .m :3 SH mm . mm. om 8 o Emacs $3 $3 .H 3H :0 S, amt. an R 0 2:8 BE puma 9m in 0.: o.m m6 o.m m5 9m 3: Nam Emacs x 8% Han om am? onH a . 3. . on Ham Hm Ewes: mHen Hmmm m mum mH H ma. 8 NNH 3 280 m 8:. hm 9H 93 .m as ...... --.. mfi ma 9m E303 x HHH 88 Ram 03 OH . mom EH 0, 0. 8H in m flat: 83.980 mmmH SN. .H man on o o on as H 9550 a. 3.3. m5 m.n --.. we 06 -..- ...... n.m m5 l... 23%. x HH 83 8% $3 0 :NHN am 0 c o . 8m «mm o 2303 53258 3.: own 0 Rm 3 o o, 8 mm o 0.550 a 3.2 N..H e5 -..- m.: ...i ...... ...i ...... Hi o.H Emacs m H 28a #3 mom 0 mum 9. o o o. oHH m Emacs 9033580 8m Rm 0 EH 0 o o 0 mm H #58 < 8.3. H309 553.5 .95 eoepooam 3.25 28 28 28 Baum 58am 88 ch33.388 5.8.5 as 0.: ad ._ as wand cowpwpmoxm b! HEM HHHMEB E 94‘ QZON ZOHBfiSooo 9?. 53¢ Hm agHmz 24m: 924 ..HmOHmz .Hozumbammm Guam Hnom 43 sample. Due to the lack of cooking residue it is quite possible that most of the vessels with which we are dealing were used for storage or water containers and not food preparation. m Shards Assignment of all the 8417 (20,522 gm) body sherds from the O'Neill Site to individual vessels was not accomplished during the ceramic analysis for a number of reasons. Primary among these was the inability to determine surface treatment for 5912 (70.2 per cent) of the body sherds in the sample; these sherds either had been broken into fragments too small to distinguish surface treatment, had had their surfaces eroded through.wind or water action, or had exfoliated along lines of cleavage longitudinal to the sherd's axis. Further, those vessels from the O'Neill Site that were restored gave indications that many combined two or three surface techniques; thus, all smooth sherds did not necessarily belong to only one or two vessels. Exceptions to the above treatment included those body sherds recovered from ceramic features; these shard groups were restored as completely as possible. Our working sample, therefore, included 2505 sherds weighing 12,248 gm which were complete enough to discern types of surface finish. The body shard sample was then subjected to individual examination and grouped into one of nine categories: (1) undetermined impressed - including those sherds with either cord or fabric impression but which could not be distinguished between the two (c.f. Quimby, 1961); (2) fabric impressed; (3) smoothed impressed -- those sherds having either cord or fabric surface treatment which was subsequently'obliterated by smoothing; (4) smooth; (5) sandy - including those sherds that were both smooth and L14 had adhering granules of sand; and (6—9) cord impressed. In an attempt to define temporal and spatial differences in the use of cordmarking at the O'Neill Site, this latter category was further subdivided into class intervals of 2.0 cord impressions per centimeter of surface area. The coarsest of these class intervals, 0.1 - 2.0 impressions per centimeter, contained no sherds. Following the formal division, the body shard sample was grouped by Area and Occupation Zone, using both the number of sherds and their weight in grams. Since both of these techniques conformed closely when percentages were calculated, it was arbitrarily decided that graphic descriptions would utilize the number of sherds rather than their weight. This not only provided uniformity but omitted unnecessary duplication. Cumulative frequency graphs were then calculated for the stratified portion of the site; Area A, including the three occupation zones; Area B, the unstratified block excavation; and the grouped Test Pit data. From the graph comparing the Area A occupations (Fig. 9), it is evident that the preferred types of surface finish for both Occupation II and Occupation III are proportionally similar; this includes minor quantities of coarse cord impressing, fabric impressing, and sandy finish, while smoothing is the dominant surface finish technique for both. The rather close alignment of the graphic curves shows little if any difference in the proportional occurrence of the above body shard surface preparation. A striking difference, however, is noted in the curve representing Occupation I. Neither cordmarking nor fabric impressing are present in the body'surface finish inventory, while 82.5 per cent of the sherds were smoothed and the remainder sandy. Restored vessels from the O'Neill Site indicate that the sandy finish sherds are in all likelihood from the 45 Figure 9 Cumulative Comparison of Body Sherd Frequency in Area A of the O'Neill Site 46 M w w m w w a w m momuzm >oom no ...zmommm 8 OCC. mu smth. sandy mpr. impressed SURFACE FINISH A v 2-4 4-6 as fabric cord ......4’ W, N - IO33 #7 bottoms of vessels; these are a product of placing still plastic, finished vessels on the sand to dry, thus obliterating any prior treatment of the basal parts of the vessels. This may explain the minimal occurrence of this shard type throughout all temporal and spatial groups on the site. On the other hand, they may also represent partially eroded smooth sherds, although the uniformity of the finish would tend to negate this inference. In an attempt to determine either patterned similarities or differences in the distribution of body shard surface treatment over the area of the O'Neill Site excavations, a second cumulative frequency graph for grouped data from all three areas of the site was calculated (Fig. 10). In this case as well a rather close alignment of two of the curves was noted; those for Area A and Area B, and once again a smoothing of prior cord or fabric impression predominated in the style pool. A divergence, however, was noted for the grouped Test Pit data. The curve indicates that on this part of the site, at least, fine cord impression and fabric treatment were more popular than on other parts of the site. Explanation of this phenomena may take the form of a temporal hypothesis. In Area A we noted that unmodified impressions on body sherds dropped out of the style pool with the most recent occupation, an occupation that was associated with European trade items, while the earlier utilizations of this area had minimal but higher frequencies of unmodified impressed surface treatment. Therefore, those occupations of the site that produce higher frequencies of unmodified impressed body sherds may be earlier in the occupational sequence of the O'Neill Site. This hypothesis, that the use of smoothing increases through time, will be tested in the following portions of the ceramic analysis. 48 Figure 10 Cumulative Comparison of Body Sherd Frequency at the O'Neill Site PERCENT OF BODY SHERDS 49 IOOT 90" 80- 70% 60’ 50-. 201’ .65; .-" '0} XREA A undet. 2'4 4'6 0'8 fcbtic Olilfli'd withsficndy Impr. cord Impressed lmpr. SURFACE FINISH N ' 2505 50 Formal Analysis and Description The 777 ceramic sherds from the O'Neill Site that were either rimsherds or bore decoration other than the body surface treatment described in the preceding section, when sorted and combined, comprised portions of a minimum of 80 vessels represented by 723 sherds; the remaining 54 decorated fragments could not be linked with either rimsherds or other decorated sherds in the sample. Within the 80 minimal vessels recovered from our excavations, 29 were represented by single sherds or sherds too small to enable us to discern the decorative sequence on the rim. The working sample, therefore, included 51 vessels complete enough that most, if not all, exterior decoration was apparent and which made formal classification feasible. Three factors controlled the handling of the O'Neill Site ceramic analysis; the rather small size of the sample, the high degree of internal heterogeneity, and our ability to apply prior ware, type, and variety designations to the sample. Initial attempts to derive typerey diagrams to explain positive and negative covariation among and between attribute groups met with little success. Resultant low frequencies for both observed and expected values at those divergent node levels which would define the formal attribute clusters gave little confidence in the utility of this form of analysis for the sample at hand. Thus, little statistical credibility would have ensued frOm this approach, and it was abandoned. In view of the above, the following formal discussion relies on extant ceramic typologies which could be applied to the ceramic sample with little hesitancy. Q:- E2 Ec—to. use of quantitative analysis of formal attributes displayed by the types and varieties resulting from the classification was not performed; this would only have demonstrated that 51 other analysts' formal systems, whether quantitative in orientation or otherwise, were indeed based on observed attribute combinations. The format of the following discussion will be ordered sequentially from.early to late ceramic ware groups within the study area. Attributes such as vessel color and hardness will not be discussed. Although ranges for both oriface diameter (in centimeters) and collar length are mentioned, we are hampered by the small sample size in determining if these attributes behave in a regular fashion. Profiles of the various wares described are found in Figures 13, 15 and 21 and may be referred to as each is encountered in the text. A single, provisional, new ceramic type has been named; hopefully, this type will be useful in other analyses of late woodland ceramic samples in the Upper Great Lakes. The category "Unclassifiable” includes those rim fragments which were large enough for nmst of the decoration to be visible but which do not fit any of the extant descriptive and formal categories, while the "Other” category is reserved for the sample of 29 small rims and fragments which could not be classified because of a lack of visible decoration. Mackinac 2332. Mackinac ware from the O'Neill Site was represented by the types Mackinac Undecorated and Mackinac Punctate (McPherron 1967a: 88). The 118 sherds recovered could be grouped into a minimum of seven vessels. Within both the Mackinac ceramic sample and Skegemog Ware sample (to be discussed in the following section), there was a great deal of internal style heterogeneity. To do justice to the ceramic sample an individual description of each vessel should be included; for the sake of brevity, however, this has been avoided. As a whole, the Mackinac wares from the O'Neill Site did not display the extreme aversion of the rim.common to 52 the type site on Bois Blanc Island, nor did the T-shaped or splayed lip profiles occur. Exterior bevels, flat, and rounded lip profiles did occur (Fig. 13). Lip decoration occurred in three cases, all oriented transverse to the.axis of the lip. In one instance, the lip decoration was the impression of a sharp-edged tool, while the two remaining were decorated with finely wound, cordwrapped stick impressions. The lip on one vessel was formed by the upper edge of an applique strip applied at right angles to the upper rim; this was subsequently decorated. Surface preparation other than unmodified cord or fabric impression was present on only two rims, both of which were slightly smoothed. Exteriorfdecoration, in the ferm.of punctatas, was present on three vessels, A single example of double-row motif was visible, decorated.with closely spaced rectangular punctates (Fig. 11A). Both vessels possessing single row motifs had the impressions placed close to the lip, one with circular punctates oriented at an angle toward the lip (Fig. 12B) and the other with deep, wedge-shaped punctates which raised interior bosses (Fig. 12A). Oriface diameter varied between 8.0 cm and 24.0 cm. Pottery assignable to the Mackinac Phase has constituted portions of the ceramic samples from a number of Late woodland sites in the Upper Great Lakes. Other sites in Charlevoix County which yielded similar materials are the Charlevoix City Park Site, Pine River Channel Site, and the Eagle Island Site. Likewise, the wood Site, the wycamp Creek Site (Lovis, n.d.a.), and the Ponshewaing Point Sites, located in Emmet County, Michigan, alSo have varying proportions of Mackinac ceramics in the pottery sample. However, in all of the above cases sharply averted rims with splayed lips occur, whereas these attributes are absent from the sample of these wares at the O'Neill Site. A single site reported by 53 Figure 11 Mackinac ware From the O'Neill Site A - Mackinac Punctate B - Mackinac Undecorated 54 55 Figure 12 Mackinac Ware From the O'Neill Site A—B -.Mackinac Punctate CHE - Mackinac Undecorated 56 r. .J'l l 4. a. \ u '..(\ sac 57 Figure 13 Rim Profiles Upper Row - Mackinac ware Lower Row - Skegemog ware 59 L. Griffin on Black Lake in Cheboygan County, the Eisen Site (1963: 73-78), seems to be solely'a Mackinac Phase occupation site. Scattered materials from Fort Michilimackinac bear close similarities to the Juntunen Site Mackinac Ceramic sample (Lovis and Mainfort, n.ds; Maxwell 1964), while a newly tested site to the west of the stockade proper seems to be early late Woodland as well (Lyle M; Stone, personal communication). Other than the type site on Bois Blanc Island, the Juntunen Site, Mackinac ceramics have been reported by Janzen (1968: 56) from the Naomikong Point Site, where three types occur. Ranging farther afield, both Wobst (1968: 259- 262) and Bigony (1970: 209-210) report Mackinac ceramic materials from the Saginaw River drainage System. Dating of the Mackinac Phase within the study area is supported by a series of radiocarbon dates from the O'Neill Site, the Juntunen Site, the Pine River Channel Site, and a possibly related data from the wood Site. A single early phase data of.A.D. 10001 140 (MH2401, unpublished) from Area B at the O'Neill Site may pertain to either the Mackinac or Skegemog occupations of this part of the site. It falls within the range of dates on Skegemog Wares from the Skegemog Point Site of A.D. 9001'120 (PL-1865; Crane and Griffin 1968), as well as dates on the Juntunen Site Mackinac Phase strata of.A.D. 835i75 (Mall42; Crane and Griffin 1961) and A.D. 910:75 (Mall4l; Crane and Griffin 1961). A preferred corrected date from the.Mackinac Phase site at Pine River Channel is slightly later, A.D. 1110180 (NH1266, unpublished). Although predominantly Juntunen Phase relationships are seen in the Wood Site ceramic sample, a low frequency of Mackinac wares and a radiocarbon date of.A.D. 10201120 (NL2057, unpublished) would argue far this data pertaining to the scattered Mackinac component. Thus,.a range of ca. A.D. 800 to A.D. 1200 may'be argued for the Mackinac ceramics within the study area. 6O Skegemog'Ware Skegemog ware is a classificatory name applied to a ceramic group from the type site of Skegemog Point, located in Grand Traverse County, Michigan. ‘Within the stated ware grouping Cleland (n.d.) defined three types; Skegemog Straight Rim, Skegemog Collared, and Skegemog Curled Rims Two of these types were present in the ceramic sample from the O'Neill Site, the Collared and Straight Rim variants. To avoid duplication, both the ware and type descriptions of the original sample will be omitted. Our discussion of the ware at O'Neill will, however, make use of the descriptions found in Cleland's (n.d.) Skegemog Point Site report and examination of the collections from this site housed at the Michigan State University Museum. The two Skegemog Collared vessels are represented by 19 rim and neck sherds; both are collared as the name implies. Lip profiles are round and square, respectively (Fig. 13), while both possess a single band of widely spaced, oblique punctates applied with a rough or notched-ended rectangular tool (Fig. 14A, D). One example has a border below the primary band composed of closely spaced impressions thought to be the corner of the tool utilized in applying the punctates. Surface preparation was obliterated by smoothing prior cord or fabric impressions in both cases. Lip decoration consisted of oblique cordwrapped stock and linear plain tool impressions. Rim aversion was moderate, although in one case it appears more extreme due to the thickness and wedge shaped profile of the collar. Diameter of the orifaces was 24 cm and 20 cm. Fourteen rim.and neck sherds resulted in a minimum.count of three Skegemog Straight Rimfvessels (Fig. 14B, C, E). All three display only slight aversion of the rim, No smoothing of the cord impressed surfaces 61 Figure 14 Skegemog Ware From the O'Neill Site A, D — Skegemog Collared B, C, E - Skegemog Straight Rim 63 was accomplished, which in two cases consisted of vertically aligned cordwrapped paddle impressions of 4 per cm, while the remaining vessel was fabric impressed. Lip profiles were either exterior beveled or flat, decorated with cordwrapped stick or paddle impressions. Interior decoration existed on one vessel in the form of vertical cordwrapped stick impressions. Exterior decoration was applied in horizontal bands of punctates, the latter either circular or rectangular with an oblique orientation. One example of a double horizontal band motif was present. The only'date on Skegemog ware was recovered from the type site at Skegemog Point, which resulted in a preferred date of.A.D. 900i120 (ML1865, Crane and Griffin 1968). Similarities between Skegemog wares and Mackinac wares, with the exception of collaring and the use of decorative motifs in the former, argue for contemporaneity with both the latter as well as wayne ware ceramics and Spring Creek wayne variants. Thus, they are another addition to a strikingly homogeneous array of early Lake woodland ceramics from the western Great Lakes. Similarly, their internal variability is also extreme. The above radiocarbon determination would support their coevality with the Mackinac ceramic series discussed in the preceding section. The sample from O'Neill might well be dated by the radiocarbon date of A.D. 10001140 (Ma2401, unpublished) which could also pertain to the Mackinac wares from the site. Juntunen Eggs, Eleven minimal vessels were assigned to the Juntunen Ware category (McPherron 1967: 111-116). Two of the three types recognized by MoPherron were present in the O'Neill Site ceramic sample; Juntunen Linear Punctate and Juntunen Drag and Jab. 64 Juntunen Linear Punctate was defined by seven vessels composed of 155 sherds. All are characterized by slightly averted rims with flat lip profiles and collaring (Fig. 15). None were castellated. Collars ranged from 2.1 cm to 3.0 cm in height. Surface preparation of the upper rim area prior to decoration was either smooth or a smoothed over paddle impression; the initial use of cord or fabric could not be determined. Primary decoration varied in two manners within the sample. The dominant decorative motif, present on five vessels, was a series of diagonals confined between the lip and the base of the collar (Fig. 16, l7A-B, 180). In each case.this primary band was bordered by a horizontal band of widely spaced punctates. All butone of these diagonal motif vessels was decorated with a round ended tool aligned in closely spaced punctates; the exception was decorated with loosely twined linear cord impression (Fig. 18B). Only one vessel shows secondary decoration in the form of obliques running in the opposite direction to the primary motif (Fig. 17B). Lip decoration varied from vessel to vessel, including linear cord applied transversely, oblique cordwrapped stick, unmodified paddle impressions, and smoothed paddle impression. Interior decoration was present only on the vessel with secondary motif. This decoration was applied in the same manner as on the exterior, but the motif was a vertical zig zag. The remaining two vessels were both decorated in horizontal bands, one with a paddle edge (Fig. 18A), the other with a plain tool impression (Fig. 18D). Lip decoration on both consisted of a transverse linear cord, while the interior decoration on both were composed of vertical impressions, one with cordwrapped stick, the other with plain tool. Four of the Juntunen Linear Punctate vessels were recovered:from datable strata in Area A, Occupation Zone III. The dates bracketing this 65 FigureL15 Rim Profiles Juntunen ware 67 Figure 16 Juntunen ware From the O'Neill Site Restored Juntunen Linear Punctate Vessel 68 69 Figure 17 Juntunen ware From the O'Neill Site A-B - Juntunen Linear Punctate (Note secondary decoration on B) Scale m Cm 71 Figure 18 Juntunen ware From the O'Neill Site A-D - Juntunen Linear Punctate Scale 0 rm 73 occupation, A.D. 12101100 (Ma2406, unpublished) and A.D. 12901 (M;2405, unpublished), are compatible with, and inclusive within, the range for these ceramics at the Juntunen Site (McPherron 1967: 112). The faur remaining vessels within the Juntunen ware classification were of the type Juntunen Drag and Jab. Two were only slightly collared and two were not collared. The square or flat lips on these ceramics were decorated in a variety of techniques: transverse plain tool impression, transverse cordwrapped stick, and paddle impressed. Both vessels decorated in a true drag and jab manner were castellated, with chevron motifs breaking the horizontal bands at the castellation (Fig. 20A). These vessels were also decorated with a plain tool on the lip, although only'one had a series of linear plain tool impressions on the interior of the rather straight rim. Also decorated in a stab and drag fashion, uncollared, and castellated, was a vessel that did not have the chevron interrupting motifs (Fig. 20B). Interior decoration was a duplicate of that on the exterior, except for the wider spacing of the punctates. The last vessel in this group was a large, elaborately decorated pot that was uncollared, castellated and decorated with numerous horizontal twisted cord bands broken by chevrons at the castellations (Fig. 19). This vessel was smooth to the shoulder and had a secondary'motif executed in stab and drag technique terminating at the shoulder. Mouth diameters fer the Linear Punctate vessels ranged between 21.5 cm.and 32 cm, while the same ‘measurement for the Stab:and Drag sample ranged between 20 cm and 28 cm. Intersite ceramic relationships are present between a number of sites both within and outside the study area. Dating of these materials, however, is dependent on the O'Neill Site and Juntunen Site radiocarbon sequences. Sites in Charlevoix County which have produced Juntunen Ware 74 Figure 19 Juntunen ware From.the O'Neill Site Restored Juntunen Drag and Jab Vessel (Note secondary decoration on shoulder) 75 76 Figure 20 Juntunen Ware From the O'Neill Site A_B - Juntunen Drag and Jab 77 78 ceramics include the Eagle Island Site on walloon Lake. Three sites in Emmet County, Michigan, also had small Juntunen Phase components, none radiocarbon dated; the'Wbod Site, near Seven Mile Point, the'Wycamp Creek Site, and Ponshewaing Point Site on Crooked Lake. These occupations are all multicomponent sites. Both Juntunen Linear Punctate sherds and Juntunen Stab and Drag sherds were recovered from.the thirteenth century occupation zone of Area A. The former type has been assigned a temporal range between A.D. 1070 and A.D. 1330 at the Juntunen Site, while the latter reaches a peak of popularity at that site ca. A.D. 1300 (MoPherron 1967a: 112-115). Outside the;study area, two undated Juntunen Phase components with Stab and drag decoration were present at the Naomikong Point Site, Chippewa County, Michigan (Janzen 1968: 56), and the Pic River Site, Ontario (wright 1966: 62-63; MoPherron 1967a: 116). Traverse £9.32 As in the case of the Skegemog wares, the sample of Traverse Wares recovered from the O'Neill Site excavations are described in reference to a group of ceramics from the Skegemog Point Site, Grand Traverse County, Michigan. This ware category, as defined by Cleland (n.d.), consists of four types: Traverse Scalloped, Traverse Pinched, Traverse Punctate, and Traverse Plain or Undecorated. All of the named types were present in the O'Neill Site ceramic sample. One of the primary discriminating attributes of Traverse wares is the presence of horizontal bands, or zones, of varying width that are characterized by different surface preparation on the upper rim.prior to the application of decoration. Thus, moving sequentially from rim to shoulder, a number of variations are possible: (1) smooth- unpatterned paddle impressions; (2) patterned paddle impressions; (3) smooth-unpatterned paddle impressions; (4) patterned paddle impressions- 79 unpatterned paddle impressions; (5) smooth-smoothed over paddle impressions, and so on. Decoration, whether it be some form of tool or finger impressions, may then be placed (1) at the juncture of different bands of surface preparation, (2) below a patterned band, or (3) on an intermediate band. All but one of the Traverse'Ware vessels from the O'Neill Site display the above attribute criteria. Seventyasix rim.and neck sherds were grouped into a minimum.of 18 vessels; nine Traverse Punctate, six Traverse Undecorated, and one each of Scalloped, Pinched, and Aberrant. The sample of Traverse Punctate vessels (Fig. 22A) comprised of 47 rim.and neck sherds exhibits all of these patterned body surface preparation. All nine vessels were slightly averted, almost straight, with flat or squared lip profiles (Fig. 21). Lip decoration in three cases is transverse linear cord impression, three plain tool impressed, one cordwrapped stick, one cord paddle, and one is obliterated. Smoothing of the lip was perfOrmed subsequent to application of this decoration. Exterior decoration was in the form of one or two horizontal bands of punctates, applied to either a smooth area or to the juncture between different zones of surface preparation. Punctate shape varied considerably'and included square, rectangular, round, and wedge- shaped tool impressions, as well as deep-100ped cord impressions. The latter is not a common form of decoration for this type. Only one vessel had interior decoration: a series of vertical plain tool impressions in a single horizontal band. A slight thickening of the upper rim.area was visible on three of the vessels in this type group. Rim.diameter ranged from 19 cm to 24 cm. The six Traverse Undecorated vessels (Fig. 22B—D) from the O'Neill Site were representedlby 17 rim and neck sherds. Two of these are too 80 Hmoflfiflflbhdo HHHoZ-O l AIM oaw3.omnm>dhe l hl¢ moHflHOHm Sflm Hm 2de a .2 .. 1:2: 82 Figure 22 Traverse ware From the O'Neill Site A — Traverse Punctate B-D — Traverse Undecorated E - Traverse Scalloped F — Traverse Pinched 83 84 fragmented to discern the presence of differential surface preparation. The remaining five, however, display zonation as either patternedpsmooth or patterned-unpatterned. Again, rim profiles are only slightly averted. All but one vessel have flat lip profiles, the exception being round. Three have smooth lips, two are decorated with plain tool impressions, and one has cordwrapped paddle impressions. No interior decoration is present, nor are there punctates, etc. on the exterior. Thickening of the upper rim, or slight collaring, is displayed by three vessels. Rim diameters ranged from.24 cm.and 36 cm. The single Traverse Scalloped rim shard from the O'Neill Site (Fig. 22E) is too fragmentary to determine a surface preparation sequence; the upper rim is, however, patterned. Deep cordwrapped dowel (1.4 cm in diameter) impressions are the diagnostic criteria for this type. Neither interior decoration nor exterior decoration were visible, nor was any thickening of the upper rim.noted. This shard was too small for oriface diameter calculations. The three sherds which, when combined, formed a single Traverse Pinched vessel exhibited a smooth-patternedpsmooth preparation sequence (Fig. 22F). The square lip, decorated with two bands of longitudinal linear cord, curved upward to a peak or small castellation, although this modification was not completely present. Exterior decoration consisted solely of a single band of vertically oriented finger pinches applied to the smooth surface band of the upper rim; the latter is the primary diagnostic attribute of the type. One vessel, smoothed throughout, with a square cordwrapped paddle impressed lip, was classified in the Traverse Ware group. This vessel, represented by 11 sherds, does not display surface preparation zonation. 85 Two bands of shallow toothed tool impressions on the slightly averted rim would perhaps indicate its affinities with Traverse Punctate. Rim diameter of this vessel is 14 cm. The coevality of Traverse wares with Bois Blanc Phase and Juntunen Phase ceramics is suggested by radiocarbon dates from three sites in the study area. A date of A.D. 1420:100 (ML2063, unpublished) from the Traverse occupation of the Henderson-lamb Site may'well data the most recent limits of the type. Two dates of A.D. 12101120 (Ma1863; Crane and Griffin 1968) and A.D. 13101110 (NL1864; Crane and Griffin 1968) on zTravarsa Wares from the type site at Skegemog Point may well bracket their peak cf popularity. These dates coincide well with a single radiocarbon determination from the Wycamp Creek Site of A.D. 1220:110 (Ma2059, unpublished), which could relate to either a Bois Blanc component or a Traverse component. Traverse wares, however, are twice as frequent as the Bois Blanc Wares from this site. Two vessals of this ware category'were recovered from the thirteenth-century Occupation Zone III of.Area A at the O'Neill Site. This coincides well with the earlier dates on these wares cited above. A data of A.D. 14551100 (PL-2389, unpublished) from O'Neill Site Area B is compatible with the A.D. 1420 date from Henderson-Lamb and is also within the range of Juntunen wares. Thus a temporal span contemporaneous with the Juntunen ceramic range is not unlikely for these materials; ca. A.D. 1200 to A.D. 1500. O'Neill Curvilinear Two vessels represented by 60 sherds are a ceramic style not yet described fer the Upper GreatLakes. The use of this type name should be considered provisional due to the small sample size with which we are dealing in this discussion. A conventional type description is presented in Appendix A. 86 The vessels were not castellated; neither were they collared nor did they possess upper rim thickening elements of any kind. Decoration consisted of three and four bands of horizontal twisted cord impressions which, intermittently and probably based on the diameter of the rim, curved upward and mat the rim vertically (Fig. 23A-B). Interior decoration was not present on either vessel, and both lip profiles were flat (Fig. ZlK-L). Lips were decorated with either linear punctates or a cordwrapped paddle. Beneath the primary decoration was a single horizontal band of wedge-shaped impressions which acted as a border. Surface preparation on both vessels involved the smoothing of prior paddle impressions. Eversion of the rim is moderate, while diameter of the oriface was 20 cm and 24 cm. Neither vessel could be completely restored; vessel shape is still unknown. Stylistically these two vessels could easily be an outgrowth of the Juntunen ware series. The use of a row of punctates as a bordering device was recognized for the Juntunen ware series (McPherron 1967a: 112), although in terms of motif the latter is linear and not curvilinear. Curvilinear motifs may possibly have arisen from the use of wavy linear motifs beneath castellations, a phenomenon that occurs both in the southeastern part of the state (Fitting 1965: 156, Plates XV, XVI), as well as near the Straits of Mackinac. ”In any case, its use on this type constitutes a break with the geometric motif tradition, while the continued use of horizontal linear motifs demonstrates some continuity as well. These materials may be provisionally dated on the basis of their stratigraphic position within Area A; one vessel was recovered from Occupation Zone I and the other from Occupation Zone II. The latter was in close association with historical trade items dated to the late 87 Figure 23 Protohistoric Ceramics From the O'Neill Site A-B - O'Neill Curvilinear C - Incised, CheVron Motif, Unclassified Scale In Cm 89 seventeenth and early eighteenth centuries, as well as in association with a trailed incised chevron motif vessel to be described later in this section. These vessels, therefore, post date the thirteenth century occupation of Occupation Zone III and continue until European contact. Occupation Zone II of Area A would fall between the early fourteenth century and middle seventeenth century. The A.D. 14551100 (Ma2398, unpublished) data from Area B, cited previously in reference to both Traverse and Juntunen ceramics, would fall within this time period. O'Neill Curvilinear, than, arises near the end of the peak of popularity for Juntunen Stab and Drag and Juntunen Linear Punctate ceramics and continues until contact. Unclassified A single vessel from the O'Neill Site, represented by 165 sherds, was decorated with a chevron motif executed by a flat tool in a trailed technique (Fig. 23C). 'Width of the trailing was 0.3 cm. This design is than bordered by a series of oblique punctates which occasionally become a double row, the upper a group of three, beneath the invertedIV.part of the motif. The 1ip:area on this vessel is missing; we cannot therefore determine whether or not it was castellated. No collaring is present, and the surface preparation is uniformly smooth. This particular vessel was recovered from Occupation Zone I of Area A in close proximity to a single O'Neill Curvilinear vessel. No ceramics with similar styles of decoration have been recovered from other sites within the study area. Given the wide range of decorative style present during the seventeenth century among aboriginal ceramics, this vessel might well be either the product of trade, movement of people, or an experiment with something new. It is not, as seems to be the case with O'Neill Curvilinear ceramics, an'out- growth of recognized precontact ceramics. 90 One thin, straight rimmed, undecoratad vessel was partially restored from 44 rim and body shards (Fig. 24). Recovered from Occupation Zone II of Area A, this vessel had a thin wedge-shaped lip profile with an interior bevel. The exterior surface treatment consisted of smoothed over paddle impressions, while exterior, interior, and lip decoration were completely absent. Rim diameter is 10 cm and the height is 10.5 cm. This vessel's stratigraphic position would place it either with the Juntunen or Traverse ware ceramic series, and it has been interpreted as an undecoratad utilitarian item, perhaps a bowl. No food residue was present on either the interior or exterior. Seventeen shards representative of two rather fragmentary vessels bore similar kinds of decoration when it could be observed in a sequential rim to shoulder manner. One possessed a slight collar and was straight rimmed, while the other was averted and uncollared. Interior decoration in both cases consisted of two rows of shallow lunar impressions probably (executed with the edge of a round stick. Lip profiles were flat, and lip decoration was either linear tool impressed or a series of round punctates. Exterior decoration occurred in two bands; an upper motif of one or two rows of rectangular tool impressions. These items might broadly fit a Juntunen Linear Punctate category. The differences observed between these vessels and Juntunen Linear Punctate ceramics from the site argued against this classification. Rim diameters were about 24 cm in both cases; this is an estimate due to the shortness of the restorable rim fragments. One vessel, partially restored from nine sherds, bore extremely unusual decoration for the Late woodland ceramics of this area. The rim was peaked and collared. Exterior decoration consisted of four rows of dragged, oblique, toothed tool impressions reminiscent of Middle Woodland 91 Figure 24 Restored Unclassified Vessel From the O'Neill Site 92 93 Laurel decoration. A similar technique was used to superimpose zig zag diagonal lines across the horizontal bands. A single row of similarly executed decoration was placed on the upper part of the rim interior. The flat lip was smooth, and the rim fragments were too small to calculate diameter. The rehaining 12 sherds in the sample were assigned to two vessels. One was straight rimmed, and the other was extremely averted in profile. Both had flat lips. The averted rim vessel had transversely applied twisted cord lip decoration, and interior decoration consisting of one oblique row of theLsame impressions. Exterior decoration was smoothed after it was applied; it might possibly have been horizontal bands of the same cord impressions, although we have speculated on the possibility of a patterned fabric impression as well. The straight rimmed vessel has a transversely notched rim, and a single band of round punctates around the upper interior rim. Exterior decoration occurs in three bands; (1) two rows of oblique toothed tool impressions applied in opposite directions to achieve a herringbone effect; (2) a band of three rows of stab drag decoration executed with the same tool; and (3) a band of diagonals below this, again applied in a stab drag fashion with the same tool. Rim diameter of the latter could not be calculated; for the former it is 20 cm. 221.933 This category includes rim and decorated sherds that, when combined into a minimal vessel count, were still too small for classificatory purposes. Twenty nine of the 80 vessels recovered from our excavations were placed here. 94 Discussion The O'Neill Site ceramic sample is rather typical of Late Woodland pottery industries in the northwestern lower Peninsula of Michigan. Its interest lies in the spatial distribution of related materials within the study area. The prior descriptive discussion of the ceramics was based on two typological systems, each applicable to specific ceramic samples located approximately 120 miles apart and at Opposite extremes of the study area. The data support the contention that the O'Neill Site was utilized by peoples practicing both ceramic decorative systems for the duration of its occupational history. For example, those sites with comparable Skegemog'ware and Traverse ware ceramic samples occur, for the most part, south of the O'Neill Site; these include the Fauvar Site, the Henderson-Lamb Site, the Skegemog Point Site, and the Schular Site. One exception to this general rule lies in the presence of Traverse wares from the wyoamp Creek Site in Emmet County. Likewise, ceramic relationships of the Mackinac wares and the Juntunen wares lie to the north, including Pine River Channel, Charlevoix City Park, Eagle Island, woods, Eisen, Ponshewaing Point, Wycamp Creek, and Juntunen. This framework, as well, has exceptions in the occurrence of Mackinac ceramics in the Saginaw drainage area, although if the discussion is confined to the littoral between Traverse City and Mackinac City the statement holds true. No other sites within the study area show such a marked mixing of ceramics from.both northern and southern ceramic traditions. Conspicuously'absent from the northern tradition was the ceramic taxa, Bois Blanc ware, defined at the Juntunen Site. This ceramic style is a Blackduck related middle Late Woodland ware dated to between A.D. 1000 and A.D. 1200 (MoPherron 1967a: 275-278). Only two assemblages within the 95 study'area included these ceramics: a small quantity from the Wycamp Creek Site, Ennnat County (Lovis, n.d.a.), and what appears to be a major component at the Ponshewaing Point Site, Emmet County. Spatially, this ware is restricted to the northern third of the study area, and its absence or paucity in ceramic samples from other sites in this area may be cause far explanations other than its derivation from the Mackinac Phase. If, indeed, it is an intermediate phase, than little if any occupation of the northern part of the study area can be demonstrated for the period A.D. 1000 to A.D. 1200. On the other hand, we may be dealing with a ceramic style that leaves evidence of a shOrt-lived penetration of the area by a non-indigenous group. Ceramic material evidence illustrates that some kind of contact was made with peoples from the Skegemog Point Site where a single Bois Blanc Bracad Rim vessel (McPharron 1967: 105-107) was recovered from the excavations (Cleland, n.d.). Intrasite distributions of'both early and late ceramic ware classes were calculated by grouping shard counts from excavation units and features from both Area A and Area B. Contour distribution maps of the various wares were then drawn, using contour intervals of 10 per cent, based on the total number of sherds assignable to that ware category. When completed, composite maps of Area A and Area B were drawn for temporally coeval ceramics, such as Skegemog and Mackinac, based on those contour lines delimiting areas which contained a greater than l_0. m ggnt distribution of each ceramic ware sample. Thus, using this arbitrary contour interval, we were able to define the areas of'most intense occupation for both northern and southern series ceramics, as well as the early and late subdivisions within each. 96 Results of the mapping procedures are illustrated in Figures 25-27. When dealing with contemporaneous groups of ceramics, a mutually exclusive pattern of space utilization is apparent. Mackinac and Skegemog 'Wares, are, far the most part, confined to the unstratified parts of the block excavations (Fig. 25). The encouraging mutually exclusive tendencies of these ceramic distributions were duplicated even when dealing with overall ceramic occurrence; very little spatial mixing of the components is discernible. Unfortunately, these data do little to explain the occupancy of either of the structures or to clarify which component the A.D. 10001140 (Ma2401, unpublished) radiocarbon date pertains. The feature from which the latter was Obtained falls between areas of Skegemog and Mackinac ceramic occurrence and could, with equal validity, be placed with either. A low frequency, (less than 10 per cent) of Mackinac wares was recovered from the vicinity of Structure 2, whereas no early period ceramics came from Structure 1. Structure 2 may, therefore, be part of the Mackinac Phase occupation. It is, at the least, early. The fact that a number of localized ceramic distributions are present might well be due to specialized task activity areas or because the site was used by contemporaneous peoples at slightly different times, both of whom possessed different ceramic assemblages consisting of the northern and southern groups. In contrast, both Juntunen and Traverse ceramics display distributions which take advantage of the partially stable dune bottoms uncovered in Area A, whilefin Area B we maintain discrete high frequency areas with a great deal of overlap in the overall distributions (Fig. 26). The overlap occurs around Structure 1, the longhouse. This structure, therefOre, is probably late, assignable to either of the above occupations. 97 8.33 mosomoxm one omsfixomz mo cowpdnfiapmflm psoonom com. swap 93.3.5 33...,» means. msflmeHpsH mm: mm cadmium 98 3.302% III: QUE 00—2 ............ 09:6 3...... -..-.. 33 5 23m 8800. ON“. 0 m_xoON 35 262.0 99 mesa: 6298.5 one zozdpcfiw mo 835926.39 peooaom com. meow nopmonw Aims. moons. wsflpmoflozH me: am enema 100 8.268% I 5:322. ............. 3.26..» I u u .. 88 5 23m eiflmmj m_xoON 35 262.0 101 Although the 0.14 date of A.D. 14551100 (M.2398, unpublished) was recovered from the Juntunen occupied part of Area B, the proximity of a localized Traverse component, coupled with similar overall distributions, still leaves doubt to which component this sample dates. It could apply to either, but a case for relationships to the Juntunen Phase is stronger. Included in the Late Late Woodland and Early Historic distributions were the O'Neill Curvilinear and trailed Chevron motif ceramics. Both (Fig. 27) are localized to Area A and, furthermore, are stratigraphically as well as spatially distinct from the other ceramic groups. These late occupations occur closer to the lake than the earlier ones and also display similar restricted spatial tendencies. The overall pattern of space utilization through time as reflected in the identifiable ceramic distributions displays a lakeward shift. This phenomenon may well correspond to the patterns of site stabilization described earlier (seg_"Depositional History”). In general, structures at the site seam to have been constructed in stabilized zones along the beach, while other activities were performed on the more active lakeshore dunes. As the latter became covered with vegetation, more evidence of use arises, until during the final occupation stages the now stabilized lakeward dunes were the only occupied areas of the site. Techniques similar to those used above would not have been appropriate for viewing the distribution of identifiable ceramics from the Test Pits. Two factors account for this: the large distances between test excavation units and the paucity of ceramic materials recovered from these units. Contour mapping would have given a false impression of spatial patterning; high frequency contours occur in areas where ceramics were present not in areas of high frequency'for a large group of ceramics. Thus, the data 102 moflsaaoo poHHwae one enmosflaw>a50 HHHoz.o Mo soapsnfiapmfim pgooaom sea gasp aopaono ape: momma mnfipaoflsz mm: am enemaa 103 an wmuafioahm ll. “um—5:. .03 5 zoom ace ___oz.0 ---- a on 9 8 ow. o. o QXQON r_ _ . 25 282.0 104 derived would illustrate where ceramics occur, and would not illustrate patterning of occupations. For instance, two Mackinac ware pots and one vessel each of Skegemog and Traverse Wares were recovered from the Test Pits, each from a different excavation unit. Each unit would have been contoured as an area of intense occupation for the respective phase. Although a statement concerning the presence or absence of occupation can be made, one could not be made concerning either intensity or the overall spatial patterning. In our discussion of the body sherds from.the O'Neill Site, the hypothesis that the frequency of smoothing of prior surface preparation increases through time was proposed. The description of classifiable ceramics and their surface preparation tendencies would support this; unsmoothed surfaces do not occur in late period ceramics. Rhoh higher frequencies of unmodified cord and fabric impression from the Test Pits would lead one to suspect, if this is true, more traces of Mackinac and Skegemog occupation than Juntunen and Traverse. Vessel frequencies, however, generate only minimal support in this direction. ‘We are, in this case, probably dealing with fragments of’more vessels than we were able to retrieve and identify. Statements pertaining to the relative density of different components in the Test Pit area, based on the body shard data, would at the least be premature. Ceramic density for grouped pottery types can be calculated for all the excavated areas of the O'Neill Site, and verifies the preceding discussion about the lack of ceramic debris from.the Test Pits. Each of the 80 vessels in the sample, whether classifiable or not, was grouped on the basis of shard weight into one of the three excavated areas of the site, Area A, Area B, or the Test Pits, depending upon where maximum shard 105 weight occurred. When divided by the number of square feet excavated in each area, a mean vessel figure was arrived at (Table 2). The density figures for Area A and Area B were quite close, .024 and .025 vessels per square foot respectively. The Test Pits illustrated a density of less than half that of the block excavations, .010 vessels per square foot. Occupation of this part of the site seems to have been much less intense than the other areas, a fact supported by the flint chip and artifact distributions as well. 106 TABLE 2 CERAMIC VESSEL DENSITY.AT THE O'NEILL SITE Ceramic Category Area A Area B Test Pits Total Mackinac 1 4 2 7 Skegemog l 3 1 5 Juntunen 6 5 0 ll Traverse 8 9 1 18 Unclassified 4 2 1 L7 Other 4 l9 6 29 Protohistoric 3 0 O 3 TOTAL 27 42 ll 80 Square feet of excavated area 1100 1650 1100 3850 Vessels / ft.2 .024 .025 .010 .023 CHAPTER.3 TEE'CHIPPED STONE INDUSTRY The two season's excavations performed at the O'Neill Site produced 33,716 pieces of raw material or non-utilized debitage with a total weight of 142,446 gm (313.4 pounds) and an additional 370 chipped stone artifacts. 'Within the latter category were 181 cores or core fragments, 59 bifacial artifacts, and 125 unifacial artifacts. Description of this chipped stone industry relied primarily on the shape of the working edge (pointed, transverse, or lateral) within the bifaca and uniface groups. Functional explanations were then proposed for the formal categories when possible. During the transition from formal to functional analysis, microscopic observation was occasionally performed to study wear patterns; for some formal categories use attribution has not been possible. Raw'Material ‘Within the large sample of flint debris from the O'Neill Site was a large series of 795 rather large, unmodified chart blocks weighing a total of 64,556 gm (Table 3). No subdivisions were discernible within this category, although a series of small pebble cobbles were grouped in other chipped stone categories (cores). The pebbles were probably gathered from the glacial till, while the cores were all gathered from the Norwood (Piawan-go-ning) Quarry Site. This chart has been previously designated as Eastport chart (Binford and Papworth 1963: 80) but has been renamed Norwood chart in Cleland's (1973) discussion of the Piawan-go-ning Prehistoric District. Norwood Chart, as it will be called throughout the remainder of this discussion, is tabular in its most common form; the chart is sandwiched 107 108 meomm o mHm mmbH meem mamm mmmH mm om: momH mmeHH meme 0: eH mmH mm: wmmH moan son eH db N3“ assoc eat: 0.58 essmsH bHann sea Han 0mm: mHm oemm de sobmH mmbm sede mama mmmm emsoH be domm nmmHH mam emmmH HmMH bmmsb new esmHez memos mbep thee 88m nanm men .mn. moan NMH mmdm com mHmm memH momm new: new: ammoH HNH QmH mmbm mmH HmomH awn can: mil HemHez esseo m sets Gamma emm mmm 0mm emeH oomH mam mH deem dem mmem noon om Hm mm mom 0mm HmoH mmH emm mom 0: meet: page HHHHV a sees ppHm com mod mom sow mm m Hde mnHH momm HomH sow mH mH em mmH mm: mm was bum mm memae3_ esseo AHHV a sees Hme o mHm new mmd mum wmm m mmOH Ham mmam mHmH gages H tease m m.n-m Hm m.eus mnH d.m-m Hem d.m-m 0mm m.HuH mm m.ouo em memeHa meesHeu ma hemeHm mxeeHm mm HeHeepez sum esmaez. esseo AHV a mess . mzom 2031388 92 same. am Ezmmemmm games: 3d a: E m mummy 109 between two layers of thin limestone cortex occurring on opposite faces of the block. This tabular form distinguishes it from both Bayport chart, common to the Saginaw drainage (Dustin 1935) and Lambrix chart from Oceana County (Rabert Green, personal communication), both of which are nodular forms. Its grey/white/blue, banded coloration is a characteristic unlike that of other chart types in Michigan. The Norwood Querry Site exposure, the most common source of this material, is located approximately 5.5 miles south of the O'Neill Site along the Lake Michigan shoreline. Other sources of similar, if not the same, chart are also available from the glacial drift around Hayden Point in Lake Charlevoix (Charles Cleland, personal communication). Both sources would have been available for exploitation by the inhabitants of the O'Neill Site, and both could be the ultimate source of the unmodified blocks recovered from the site. The distinction between these unmodified pieces and modified block cores was based on the observation that all unworked material exhibited smoothed, waterworn cortical edges and flake scar ridges. The latter are a result of the chart eroding from the outcrOp and falling on the beach, where they often fracture due to natural causes. Likewise, the drift around Hayden Point extends into the water, and similar conditions could have occurred at this location as well. No platform preparation was present, although battering was frequently visible on the edges of the blocks. These characteristics imply that these pieces of raw material were gathered from the beach below the outcrop at Norwood or from a shoreline situation at Hayden Point, both of which are littered with eroded chart cobbles (Charles Cleland, personal communication) rather than quarried from the outcrOp itself. llO 9.2m This analysis uses the definition of cores, as a formal category, prOposed by Binford and Papworth (1963: 83). That is, they are pieces of raw material that have been prepared so that systematic flake removal may be performed. Using these criteria, 181 cores and core fragments from the O'Neill Site were assigned to five major categories based on differences in the technique and pattern of flake removal. Block cores and block core fragments constituted the largest core category recovered from the O'Neill Site. All 68 block cores were manufactured on Norwood chart. This sample includes lb block core nucleii, lfi intact but unexpanded cores, and 40 broken or fragmentary examples. Both the intact examples and the expanded nucleii exhibited cortical faces at two ends; these Opposing faces also functioned as the striking platforms for flake detachment (Fig. 28). This phenomenon was noted in the Eastport Site lithic industry as wall, in which the block cores' ”striking platform frequently was the original cortical surface of a rough tabular nodule” (Binford and Papworth 1963: 97). All but three of the 41 block cores from the O'Neill Site possessed at least one flat cortical surface that had been utilized in this manner. The two cores which showed evidence that both cortical faces had functioned as platforms illustrated that flakes would have been driven from the opposite side of the tabular block. Although 13 of the cores examined in the sample had flakes driven from non-cortical areas of the block, this was infrequent and probably resulted from a process of core trim. Likewise, four of the block cores assumed a rather regular, polyhedral shape on the worked faces, specifically when lamellar flakes were struck in series (Fig. 29). 111 Figure 28 Cores From the O'Neill Site Upper Row - Block Core Nucleii Lower Row - Block Cores 112 I 2 3 4 5 Scalo'ncm 113 Figure 29 Cores From The O'Neill Site Polyhedral Block Cores 114 L- ‘ ‘. ‘ ,\! . ‘ .‘ _ -t"’ VJ ' a Scale m cm 115 Metric attributes Of the block cores from the O'Neill Site were gathered from the unbroken block cores and the expanded nucleii (Table 4). Length measurements were made from cortical face tO cortical face. Thickness and width cannot be applied to the block core, since they are essentially cylindrical; therefore, maximum.diameter will be used. Both the mean and the range Of the length Of the two groups corresponds closely, while the measurements Of diameter are rather disparate. The lower limits Of block core length and diameter were both made on an extremely small example Of this type. TABLE 4 LENGTH AND DIAMETER OF BLOCK CORES FROM THE O'NEILL SITE N X Length Range X Diameter Range Nucleii 15 5.4 cm 4.0 - 6.1 2.5 cm 1.6 - 3.3 Block Cores 14 5.6 cm 3.5 - 6.7 4.4 cm 6.3 - 3.0 The second most frequent group Of cores from the O'Neill Site were those displaying Opposed zones Of percussion as a result Of a bipolar, hammerstone and anvil technique. Binford and Quimby‘s (1963: 289-296) six-fold classificatory scheme was utilized to group the 48 bipolar cores in the sample. All six categories were recognized in the O'Neill Site ‘specimens; (1) basal area-point, N=4 (Fig. 31); (2) Opposing point, N=2 (Fig. 30); basal area-ridge, N=l4 (Fig. 30); Opposing ridge, N=9 (Fig. 30); (5) ridge-point, N=5 (Fig. 31); and (6) right angle ridge,;N=5. Additionally, nine fragments were also present in the sample. All of the latter exhibited areas of resolved flaking and battering that were similar to those present on zones Of percussion on the intact specimens. Indeed, five Of these cores exhibit a sharply concavo convex edge which is reminiscent Of MCPherron's "gouged-and" artifacts, whereas only five Of 116 Figure 30 Cores From The O'Neill Site Upper ROW'- Opposing Ridge Bipolar Cores Middle ROW'- Area/Ridge Bipolar Cores Bottom Row - Point/Point Bipolar Cores 117 m. m 118 Figure 31 Cores From The O'Neill Site Upper Row - Ridge/Point Bipolar Cores Lower ROW'- Area/Point Bipolar Cores 119 Scolo'llom 120 the intact bipolar cores in the sample displayed this attribute. There is a possibility that bipolar cores may'have been utilized as tools (BinfOrd and Quimby 1963: 288-289), and MePherron speculates that the ”gouged and” artifacts may have served as gouges (1967a: 137-142). Brose's (1970a: 104-106) microscopic analysis Of wear patterns has not clarified this problem other than to demonstrate that wear striae are present. It is tempting to speculate, however, that five Of the nine broken core specimens from the O'Neill Site had a "gouged" and and that the breakage may possibly be the result Of heavy use. Metric attributes Of the bipolar cores from the O'Neill Site are presented in Table 5. As a group the area-ridge cores are larger than the other bipolar cores in the sample. This is true Of the Summer Island bipolar core sample as well (Brose 1970a: 104), although not true of the materials analyzed by Binford and Quimby (1963). The area-ridge cores from O'Neill may also be favorably compared with some Of the rejected block cores from the site, specifically those that were seemingly discarded by a hinge through fracture. This phenomenon produces a ridge that may have functioned as an anvil support for continued knapping Of cores which otherwise might have been rejected. All but two bipolar cores in the O'Neill Site sample were manufactured on Norwood Chart. The exceptions were also the smallest cores in the sample and were a very waxy, high quality material; this characteristic may be the primary selection criteria (Binford and Quimby 1963: 353.354). On the other hand, this type Of industry is generally located in areas where primary flint sources are scarce or nonexistent and the occurrence Of higher quality charts is a function Of the proportions Of these materials in the glacial till. None Of the O'Neill Site bipolar 121 :Hflz H.d I o.H ac m.H m.: I w.H ao m.m n.o I m.N BO 3.: Omoflmlmon¢ mnz o.m I 5.0 so 3 m.m I m.H SO 0.N ¢.m I o.~ SO w.m Omoflm oomommo m.~ I m.o so m.H m.m I o.H BO Jom o.m I w.H 5 ma paaomIomeem muz o.H I 3.0 So w.o m.n I m.a BO w.N 0.: I m.H ac m.m omeam edema .am mama onoo euz a.m I H.H So m.H 0.: I m.a SO w.m e.e I H.m , SO m.m psflomlmoa< n.H I m.o SO o.H m.m I :.H a 3 m.m I N.m SO N.d pgaoaIesaoa mHHm qHHmz.o Mme 20mm mmmoo m¢qomHm mo mezmzmmbm¢mz m>HB¢msnamI||III onoo xOOHm deIomcfimlllll mooaosz ooocoQNm ozozlllll. No>QOOflm asseesfimsoqll No>eooIoemam . Mo>sooIosmHm . enclomnfim no IlIIII amaodflm nmdoaanomgflm mooaodz economxm onelelll 28.8 assaeeflmseoj moaaoez ooocamxm oeolelllI ZOHHAO Hmmmzoomm mw¢mmmmm mo Mmozmm amHomflm ZOHBmo HmmZHmm :Oflmmfionom pooaflmlllll. oooshoz amafiflmelowamq moabnom ooosaozlnoz no noosaoz HHoEm Ademfiaz 3am 163 The 172 other chipped stone artifacts from the O'Neill Site fall into a wide range Of formal classes that functionally overlap. Readily recognizable secondary tOOls utilized in the manufacture Of‘WOOd or bone implements compose a small part Of the total industry; gravers, drills, and choppers account for only 3 per cent Of the tool inventory. Cutting and scraping tools, which may well have served secondary functions, dominate with 75 per cent Of the sample; projectile points and fragments account for 15 per cent Of the total. The presence Of projectile point tips and the large number Of cutting and scraping tools should reflect a large ameunt Of both hunting and processing of game animals. These subsistence activities, however, probably were a necessary, but secondary, adjunct to the flint processing activities. Examination Of the O'Neill Site industry to deduce the primary subsistence activity perfOrmed at the site was accomplished by comparing the ratio Of bifaces to unifacas. A resultant ratiO' Of 1.75 unifacial tools per bifaca would support the conclusion that the occupation relied for the most part on fishing (Fitting 1969: 367). Two considerations may negate this interpretation. First, Fitting notes that the derived ratios are not applicable to WOOdland sites because Of the omission Of certain tool categories from Taggart's (1967) study from which the ratio was calculated. Secondly, 17 Of the bifacial implements, a total Of 27 per cent Of the bifaces, are manufactured on small flakes that have been shaped by bifacial.marginal retouch. This includes projectile points, knives, and scrapers. MePherron (1967a: 149) Observed that unifacial Juntunen points occurred in the Juntunen sample. Upon examination many of the Late WOOdland chipped stone industries in the Upper Great Lakes seem to be flake industries, and minimal bifacial stone work is necessary. The 164 O'Neill Site fits this description; laterally'modified unifaces sufficed in places where the Saginaw Valley Archaic would have manufactured bifaces. The above statement is supported by the frequency distribution of flat flakes in class intervals Of length. It was Observed (Fig. 33) that 54.6 per cent Of these flakes were between 1.0 cm.and 1.9 cm, and 80.6 per cent fell between 1.0 cm.and 2.9 cm. Only 13 artifacts in the sample would not fit on flake blanks in this size range: the lanceolate bifaces, ovate bifaces, large triangular bifaces, preforms, and lunate chopper. Considering the size, quantity, and quality Of raw material available, there seems to be a conscious selection for small flakes and small-flake- derived tOOls. Spatial and vertical distributions Of both debitage and flint artifacts at the O'Neill Site revealed major differences in the density Of artifacts and the types Of activity carried out on different parts Of the site. The lack Of temporally diagnostic lithic materials, coupled with a paucity'of artifacts from.stratified portions Of the occupation (Area A), hindered attempts at eXplaining fermal variability in terms Of temporal trends. The only flint artifact class that could be analyzed in this manner was cores. Vertical distribution in Area A illustrates a sharp reduction in the frequency Of cores during the most recent occupation (I), while the frequencies remained virtually equal throughout Occupations II and III. Despite this sharp reduction, we can still define a selection for plano—convex cores for this latest occupation. Likewise, bipolar cores are mere popular during Occupation II; Occupation III reveals less clearly defined preferences with some emphasis on block cores. There was considerable differential distribution and density Of chart artifacts between excavation areas at the O'Neill Site. Artifact density 165 in Area A was .0210 per cubic foot; for the Test Units it almost doubled to .0417, while Area B jumped to .1117 artifacts per cubic foot. The relationship then can be illustrated as B;>Test Units:>A and can be said to hold true for tools, cores, and chippage frequency as well as overall density; A similar relationship also pertains to the stratified occupations in Area A, where all Of these groups except for cores decrease through time: III:>II:>I. For cores this would be illustrated as II3>7 III:>I. Area B, if the above relationships are true representations of the amount and kinds Of activities carried out on different parts of the site, represents a focus Of activities dependent on both primary and secondary chipped stone tools. The stratified parts Of the site, on the other hand, seem.to have functioned in a different manner for all activities involving chipped stone tOOls with the exception Of core processing. Chipped stone materials from the Test Pits, however, indicate a lack Of secondary tools, implying that the manufacture Of wood and bone implements was localized to Area B. In addition, the three preform fragments from.the site were all recovered from one unit in the Test Pits, N370/E90. This activity may have been localized to certain areas on the fringes Of the occupation. The consistency'in the prOportion of chippage classes and raw material found on different areas Of the site made it impossible to prOpose statements concerning the types Of flint knapping activities perfOrmed across the site (Fig. 41). Differential proportions Of chippage types recovered from each Of the three occupations Of Area A, however, led to assumptions pertaining to the stage Of manufacture that was emphasized during each (Fig. 42). Throughout Occupation II more care preparation, in the farm Of block and primary flakes, took place than processes involving 166 Figure 41 Cumulative Comparison Of Flake Type Frequency at the O'Neill Site 167 0.00.0”... ’ ’ ........ ..I. .4... ea. {0’ w . n . m . . ... #1053 wao aaHsOaHo naHdOnHo onnvaom an: Iwemfiae Ha>o naHsmeap Iooaosm Hmao aastaap IOoaQSm amHsOnfio carom ZHMZ m.am.mm¢m domdosm Nutmeg humanism memeosm méeéosn Néeeoiem m.aa.mesm m.mm.wesm cement nooaoz mOHapao 178 Ground Stone Grindstones The O'Neill Site yielded no groundstone material which could be classified as mortars or pestles. On the other hand, certain Of the plano-convex cobbles from the site, although waterworn, possess surface sheen on their flat faces. The three cobbles in this group have maximum diameters between 10.3 cm and 6.4 cm, while they are between 3.0 cm and 3.6 cm thick. A fourth example, measuring 7.8 cm.x 5.6 cm.x 1.3 cm.in maximum.dimensions, had visible wear striations running parallel to the longitudinal axis. This artifact was manufactured on a waterworn plano- concave beach pebble Of irregular shape. The worked area was approximately 1.8 cm x 5.1 cm. 'Wear striations continued Off one end Of this pebble, and terminated in deeper gauges 2.7 cm from the Opposite end. Whetstone A single piece Of micaceous schist was also recovered from the O'Neill Site and has been classified as a whetstone (Fig. 43C). The artifact is fragmentary, measuring 4.7 cm.wide and 1.3 cm thick. Both lateral margins are heavily polished from use. Throughout the excavations at the O'Neill Site, any material that was not Of local origin and.which may have been transported to the site, whether worked or unworked, was saved and retained for laboratory analysis. As a result of this recovery, 192 fragments Of slate and shale were analyzed. Due tO the high frequency Of exfoliation, transverse fracture, and surface erosion, it was necessary to restore much Of this material as completely as possible; they were then subjected to both visual and 179 microscOpic analysis. Twenty-seven Of these fragments, four slate and the remainder shale in various states Of decomposition, were restored into 12 either whole or fragmentary artifacts that were intentionally modified tO some degree. The 12 artifacts were subsequently subdivided on the basis of presence or absence Of ground margins or faces. Seven artifacts within this sample had some fOrmal alteration prior to utilization, while the remaining five had none. 'Within the category Of modified materials were a gorget blank and a fragmentary pendant. The gorget (Fig. 45A), manufactured on a large piece Of partially eroded shale, was quadrilateral, with one and wider than the other. NO holes had been drilled, and the corners of the larger and were broken. This ornament measured 13.6 cm in length, 6.4 cm.in width at the smaller and, and was extremely uniform in thickness ranging between 1.1 cm and 1.2 cm thick. Both the longitudinal and transverse cross sections were biplano. The pendant, also manufactured on eroding shale, was broken at both ends (Fig. 45B). Its ovoid transverse cross section had a maximum dimension Of 2.1 cm and a minimum Of 1.1 cm. .A single, scratched, decorative line encircled three fourths of the circumference Of the artifact, located in the center Of the fragment. A single ground slate ”knife" (Fig. 46D) was recovered from Occupation Zone III Of Area A. One end Of the artifact is broken. The upper blade edge is straight and ground for a distance Of 3.6 cm, while the lower blade is curved, ground along 6.2 cm Of the edge and terminates in a shallow notch assumed to be a haft. This haft arrangement may have been duplicated at the Opposite and, since the thinned area presents a structural weakness; the haft was broken and restored on the portion Of 180 Figure 45 Ground Stone A Gorget Blank B Pendant C Abrader 181 182 Figure 46 Ground Slate and Shale A-C Unmodified Shale "Fish Sealers" D Ground Slate Knife 183 - - ‘ .- A B C D o l 2 3 4 5 Scale In cm 184 this artifact that was analyzed. Wear striations were noted along the central portion of the blade edge, oriented in two directions; perpendicular to the working edge, or on a mean Of 30 degrees toward the hafted end. . Abrading tools at the O'Neill Site were represented by one slate fragment (Fig. 45C) and one shale fragment. All intact edges on these artifacts are ground smooth. Both have haphazard scratches oriented parallel to the longitudinal axis, while the larger slate piece has deep, intermittent gauges perpendicular to the edges as well. The two remaining artifacts in the sample Of modified materials cannot be classified functionally. Despite their size disparity, the larger measuring 15.9 cm.by 11.5 cm by 1.0 cm and the smaller 7.8 cm by 1.5 cm by 1.2 cm, both have ground edges. The larger piece Of shale may possibly represent an initial stage Of blank preparation fer a gorget. Those five pieces Of shale unmodified by grinding (Fig. 46A-C) had wear striations present on at least one edge Of the artifact. These striations did not demonstrate any directional uniformity in relation to the working edge. TWO had wear on the ends Of the artifact, while the remainder had the striae confined tO the long edges. All of these artifacts were elongated and waterworn with tapering, thin edges. Measurements Of these implements are summarized in Table 9. Unmodified but utilized artifacts similar to those described above are not reported from any WOOdland sites in the Upper Great Lakes. This may be because they are Often broken into small fragments and look much like water-rolled beach cobbles. The shale from which many Of these implements were fashioned does not occur locally. It does, however, occur at the Norwood Quarry Site, interbedded with the Traverse limestone and 185 moonmoo mm moonwoo N mooamoo on mooamao ow moonmoo m macawoo m: amom mnflxnoz.aoam mafiapm mo OHms< a. a.e so m.o ac m.H as m.o ac m.o ac m.o mmochHze as a.m as a.s so H.m so a.a a. e.s a. a.m opon 20 m.mH as m.oa SO w.MH So m.ma gamete mHHm AHHmz.O mme 20mm mHmoxm 0: mm NH aoasnz endpmom 224 H OQON soapmmoooo .< moaa ad moaspmom mo mafipeooq mm areas .. . r...w... a .o .. . n .. I»: ... u.,L. t. u .. 2.5.: a) p.753. .... .mlma must. ca. ..l 1.-...- . 3.... Isjluz 5 2 2 1.x... L: _. 1.1.4:. “5.3.5.1.. l . ...» ... r. \\ . Ax _. A.” _ r l a. ...; _ ... 3:5 I , E: 3; .W Estes; r|.|lL I I4 rlllL “I- . s r. . r. l. 1 fl ‘L rill L ‘x :1 . a ...... E: 5. l 1‘ Ty Ll. qQVV“ 1) _ l . ‘io‘ . u V a v . v 1 .9. . v . W . A V a ‘ ~ V . .‘ . r . k b . a As. :Eutoaa 8 madman 32.38 S 5.3 uric. 2. massed... Z - mud Z. undue“ m. . H M719” Zipdlj .J A. d aid. in? jazz L--... 226 AOQOHQSoosfi soapepoamv onoz Aopoamsoocfi soapmpoawv osoz menopzoo semen esteem refines £95 9.0 x .m.m x .o.m semen sodasmm meadow .awdflondo .N.o K .H.m N .m.H soapepoamaopsH ememm mQOHonSflQ ngOHpossa HH mzom ZOHHdeooo .« eoxm ma 0H aonaoz snowmen 227 HH OQON QOflewmsOoo a< moa< me moedpmom Ho soapmooq mm enemas 228 ...... .....a....m.1-._cu tabled a... brew 6.3.1 .....,._..........xw uo 3:2,: mm .. .. 2.....1...c um again no mom ......» bum... mmm» wxnow ...muahdUc Q24 Eldmt » 1m» H011 Imppdom +£1de h; mmfiidmuj; _ wxfiLQMu/L.‘ m U..- - V (Us. . O 00 00 .I‘ ......u. 31. . be + .s mzox Z Faun 7: me u 7 ,2. 14 ~( ._ r flurd A; (v 229 aeHo nnfim m I eesan Asa eel HeHnopma are H I anHm been one cameo pom AspOHmaoosH :OHpepOHMV osoz Anm :va hnqeeam mm .Aam a.ev am Haaeaeab Ase omv unseen as I beads mOHSdnoo Aam mwv m haeSHnm .Anw m.mwv asaaeeab Hm .Anm we: a base .ca 93 HeHaopms_3ma H I pcfiHm .Asm mama .eseeass mma .Anm OHV commonaSH .pooss H .Aam mwdv em oompooem .Aam MHV hence a I mOHsmnoo mesopcoo OHHQ bemoan msHmmooonm boom HHonWQESQ msOH He>o noHdmoanH semen seldom eSHHsAm .Heso nausea condom semen .neHoOAHo ease :OprpoamaopsH omwgm HmQOHuosdm .H.H K .N.N .m.o N .N.H K .H.H .m.o N .H.m x .w.m A-m.m K -:.0Hv .m.o an .H.+~ K .m.m .m.o x .m.e x .e.m mQOHmmoaHQ HHH mzoN ZOHBameooo .a amma.zH memeeams NH mamaa os3\ommz omzxomnz oe3\ommz erommz om3\oemz $522 om3\ommz oezxoamz pHQD OOHpmbaoxm Hm mH HH 0H m nonesz anagram 230 heHo owsenoneon oopoemaoo AOQOHmaoosH sowpepOHmv osoz AOQOHQEOOGH SOHpepOHmv osoz AO¢OHQEOOQH SOHpeQOHHV noszm pen H oommonmaH once :mH pepenoooo w mSHa m I mOHawnoo mesons .88 we mSHn : I mOHesaoo mesopsoo onspoehssea ceasaoo .sowvdnsmonm ease a3 banshee gpaeom apnoea meadow nosHmpsoo OHeuaoo oopawomHm nosHepsoo Ostaoo oooawomHQ soepmvoamnopmH HeQOHpossm sHmep 3OHHenm aessom semen.zOHHenm .Hebo neHsmonnH samen.30HHemm .Hmbo answoanH semen aoHHenm .He>o ommmm .N.o N .w.H N .m.H TwH N .N..HV .H.o N .w.o N -N.H A-M.N K .w.Nv .H.o N -N.H N -B.H A.w.m x .o.:v .N.o K -m.m N -0.N .N.H K .m.H .o.H N .m.o msoemsoeam HHH mzom ZOHBdonm mm mm mm wN nogfisz caspaom 235 Figure 55 Location of Features in Area B —— _‘ ‘\_¥ '2 236 K) ,i" I SCA'IiEkIN Eur . I 2 I, ‘s‘.’« . FELT'JRE '5 ‘ ‘V,,A‘-«{ ',.-".P, - ' W, ’ \\ I T‘FfArJHL g,’ ' FEAT m5 :4 : MODERN) . 9F W o' ‘ .. kg:‘\ 0 "r I . . V. .'\.. ‘ ICFL'NT 4.". . . o . ° _ I x " I. .1. r *FEAT 8 o :9 .0 0 c lirfl o b . . o x — ' A v- _ --—_J I ~ +FEATJRE 3;, i i I fl . - - O f ‘ l ' 1% . L”, . , .,..v' c 0 ’ r.” \ o . o . ‘ ° . o . 0 O . . I ' __.J NEIZHYH A.“ 1).,‘2'Yf :1 ”E. L -4 m ,m , r 0'. 9.0—C lNY E‘a YREt pEL.‘, _- 5“ [3'3 Hf“? W4 '. . FECIY'JQE 5') [;d _g_I_ I'q C“ t‘ “a...” . V AJ [-13 F—'~'.3“1|£ l“ 3“ Y'QH' .\ 'i £‘."1‘.‘.'.". ‘III'II l ' | 237 W» H3 8.83535 3H 2m 3 vaauougad wonwauapomna H .Anm o: noguooau :H .Anw mm cannon afiunnu H: .mn adv cannongafivnoo n .An NV hflcum H I mafiauuoo 2m «.8 3893 H: I anE Aopoflnabona cowpuuoamv onoz Anm mmv nopunmm nonH Hum 93 H H3330. .2» as H E83 4% and m Hawnopua.3um I pnfidm mpcopnoo ponwupnoo oafiuaoo ucunuomwn nvnuom HE megawfifl nowpwponmnounH dazedpoqsm mBHZD 9mmH 2H mambadmm ma mnde 23% :OHHBHSuHu H¢>o sands mood .naHSonHo .¢.H N .o.H .m.o N .m.o N .w.o .H.H K .w.o mnoamqofia 363mg Smflommz omHEoomz was: :owpwbdoxm on mm mm nonauz andpmoh 238 statements concerning their function. An entrance is located at the southwest and of the house. It should be noted that the areas of’missing posts along the northwest wall of this structure correspond to areas where the occupation zone was truncated by natural action; further, little artifactual material was recovered from this area. The second structure includes a semi-circular line of post molds in the southeast portion of.Area B (Fig. 55). A possible entrance is located on the west side of this partially'excavated house. Post meld diameter ranged between 0.25 feet and 0.35 feet. No hearth was feund'within the exposed area, although a large concentration of flint was included within the post perimeter. If this structure is circular, only half was excavated, and it may indeed be the end of a house similar to Structure 1. Comparative interpretations cannot be made until.mcre is known. House structures similar to the longhouse at the O'Neill Site have been excavated at the Ponshewaing Point Site, Emmet County, and at the Juntunen Site, Hackinac County. The structure from the Juntunen Site is constructed of a double row of wall posts (McPherron 1967a: 235-239) and is wider than the O'Neill Site example. Although differences may'also be found between those from the Ponshewaing Point Site, the similarities are much more striking. Rounded ends, parallel single post walls with compatible post mold diameters, and entranceways located at the ends of the house all bear close similarities to the O'Neill Site Structure 1. The major difference that could be observed is the use of heavy, pit- placed support posts near the ends of the side walls. This variation is relatively insignificant and.may'bear more directly on the type of saplings or covering used in construction or even the type of ground surface on which it was constructed. The Ponshewaing Point Site is a \‘I' 11 ‘ll tl .239 Late Woodland site with artifacts dated to between A.D. 800 and A.D. 1300. Since final analysis of this site has not yet been completed, it is not possible to refine our dating of the structure from this village. I.ll inl- I11 I 11" ll IIIII'llI‘ GHAPTER.9 SPATIAL ANALYSIS AND CONCLUSIONS Spatial Analysis Although the O'Neill Site does not offer ideal opportunities fer analysis of intra-site spatial utilization, particularly in view of the multi-component, only partially'stratified nature of the site, it does afford greater possibilities in this direction than many other sites in the study area. Descriptions have been forwarded of what seem, after a partial analysis of distribution of artifact classes, a series of localized habitation areas confined to areas away’from the location of dwellings. Thus, there is a strong chance that these distributions represent task-specific activity areas. The rest of the analysis will attempt to test the following question: is there a tendency fer certain chipped stone artifacts, ceramics, and faunal debris to be distributed in patterns in space? An attempt, therefbre, will be made to define relationships between and within groups that may well be functional. To accomplish this, certain classes of artifacts were coded according to their frequency'by.Area and Occupation Zone. It was necessary'to delete one of these classes from each of the analyses because of’uncontrollable factors. For instance, using bone frequency as a variablefin Area B could have given a skewed picture of the relationships of faunal debris with artifact classes, due to differential preservation. Faunal remains were, therefore, deleted from.Area B, but were retained in all three occupation zones from.Area A, where the conditions for bone preservation were present. 240 241 Measurement The measure chosen to test associations was Pearson's product-moment correlation coefficient, a measure that defines a linear relationship between two variables. The correlation coefficient, 3;, ranges between values of +1.0 and -l.0; the former indicates that as one variable frequency increases, the other does-as well. Likewise, the negative value would illustrate an inverse relationship; as the frequency of one variable increases, the other decreases accordingly. A value of g approaching, or at, 0.00 indicates no patterned relationship; one that could occur by chance. It should be emphasized that non-linear correlation between two variables can be computed; that is, the relationship is defined by a curve and not a line. This analysis, however, will be limited to the linear case. The intercorrelation matrices were obtained as output from ICMATRIX, a computer program designed for use on the CDC65OO computer at Michigan State University (Paulson, Donaldson, Price, and Carroll 1972) and programmed by the Computer Institute for Social Science Research Stat Systems Group. These matrices are presented in Appendix B. To maintain some security in the following statements, levels of significance were computed for each of the four intercorrelation matrices. Area B, with 18 d.f. and r=.5000, has a p<.05 of being due to chance alone. Likewise, the Occupation Zones I, II, and III, with 9 d.f. and r=.6000, are significant to .05; they are not due to chancedsampling error) 95 times out of 100. On occasion, correlation coefficients with less significant values will be used to help strengthen specific statements. Occup_ation II; Analysis of the spatial relationships within Occupation Zone III of Area A, the most intensive occupation, verifies some of the concluding 242 statements contained in the ceramic analysis. Correlation coefficient values between ceramic types were either close to .0000, an association that could occur by chance alone, or negative, in which case they do not occur simultaneously. This substantiates" the concept of the occupation of the O'Neill Site based on the contour distributions obtained from the different ware categories. Strong positive 3 values resulted between a variety of artifact groups and in part help refine the interpretation of the subsistence patterns practiced by some of the inhabitants. Distribution of Juntlmen ceramics, for instance, correlates highly with both fish bone (fJI6529) Ed mammal bone (_r_=.5282.l+). Subsistence emphasis does not seem to have been an either/or proposition; both terrestrial and aquatic fauna were being prepared or consumed. No strong positive or negative faunal associations were obtained between other ceramic wares; their occurrence with these wares could be due to chance. The strong associations between Traverse wares and bifacial cores (f.96l89) and Skegemog wares with the few plano-convex cores from this stratum (3567802) may well indicate that both of the latter inhabitants utilized the partially stable dune area as a special function locus for flintknapping, whereas the Juntunen pepulations were employing it primarily for either food preparation or consumption. Occugtion E Explanation of O'Neill Curvilinear ceramics as a stylistic continuation of Juntunen wares was strengthened by an a value of .6046 between the two in Occupation Zone II. Their similar spatial relationship is perhaps indicative of other ties as well. The possibility exists, of course, that the ceramic remains were deposited at different times, in which case little stylistic seriation should be attempted. Both stylistic 243 and spatial indications, however, point towards major similarities existing between the two. An added factor is high correlation (256708 and _r_=.99816) between Juntunen ware and O'Neill Curvilinear respectively and bifacial cores. On the other hand, only Juntunen ceramics are distributed in a similar manner to the fish bone in this stratum (£56038), a situation present in Occupation Zone III as well. Occupation Zone II is prObably a summer occupation, while Occupation Zone III is primarily a spring utilization of the site. In either case, the Juntunen occupants seem to ' be exploiting aquatic resources and were either preparing and/ or consuming them in the northeast part of Area A. Occumtion _I_ As noted previously, the most recent occupationr of the O'Neill Site is most likely the product of a number of short-term camps spanning the late prehistoric and both early and middle historic periods. The historic inhabitants left a thin scatter of trade materials in both Occupation Zone I, Area B and the Test Pits, while the late Late Woodland occupants seem to have had a rather confined occupanp'y, primarily Area A. O'Neill Curvilinear ceramics and the chevron-incised vessel were both recovered from a single unit. The resultant intercorrelation matrix between artifacts from Area A illustrated high coefficients of correlation between these ceramics and fish debris (35 5892) and two core types: block cores (356517), and plano-convex (35 5079) cores. Block cores and mamalian fauna also display this relationship (_r_=.1+083), as does the former with fish (357015). On the other hand, plano-convex cores and mamalian remains rarely occur together (£53018) . The interpretation of this phenomenon is that block cores have a distribution overlapping that of the area defined by O'Neill Curvilinear ceramics. The resultant coefficients 2141-!- also have some bearing on the possibility that the historic period's occupants are responsible for the mammalian fauna in this occupation zone. The fact that late aboriginal ceramics and plano-convax cores occur together, but plano-convex cores and manual remains do not, may perhaps indicate dissimilar temporal relationships. This conjecture is partially' upheld by the presence of butchering marks, some of which may'have been done with a metal blade, on the mammalian remains from.0ccupation I (D. King, personal communication). If this is true, then it follows that the ceramic bearing occupation may not be a summer encampment but that the historic occupation probably'did take place during the summer months. mas Unlike the three occupational strata in Area A of the O'Neill Site, Area B contained few faunal remains and an abundance of chipped stone artifacts. Beyond this, the evidence of dwellings in the unstratified excavations would presuppose some sort of functional differentiation between the two areas. Of the above contrasts, the first cannot be controlled. The differential preservation present in Area B has been noted. Thus, two points of departure are left in the analysis of intra- site functional differentiation. ‘Within the group of'3# artifact variableston which correlation coefficients were obtained, two did not occur in Area B: CI, Tl projectile points, and bifacial preforms. Given the rather poor control on temporal relationships of chipped stone artifacts from this part of Michigan, the first part of the analysis attempted to define chipped stone artifact associations with datable ceramics.. Although some of these associations are probably-the result of prehistoric disturbance and scatter, the derived associations may'be tested for similar relationships 245 at other sites; hopefully, then, these results may be viewed as future testable hypotheses. Due to a high correlation between Juntunen ware and Skegemog ware (35 8210), temporal dichotomies will be difficult to define for artifacts associated with these types. Although a similar situation obtained between Juntunen ware and O'Neill Curvilinear in Occupation Zone II, the contexts differed enough that the problem of superposition warranted alternative treatment. First, Occupation Zone II was virtually devoid of chipped stone artifacts, making it difficult to derive temporal associations, whereas Area B provided more Opportunity in this direction. Second, the temporal disparity between Skegemog ware and Juntunen wares is great enough that ignoring their superposition would have given a distorted view of the associated chipped stone materials, while O'Neill Curvilinear is probably coeval with Juntunen Wares at its inception. Furthermore, both wares were highly associated with bifacial cores, the only differences in association was the Juntunen Ware and fish bone correlation. However, both Juntunen wares and Skegemog wares covaried highly with the same chipped stone materials in Area B: a total of four artifact group. Mackinac wares illustrate similar distributions with small triangular unhafted knives (£56296), small triangular hafted knives (35 7293), awls (3:.9677), and with both retouched and utilized flakes (£=.1P107, 3:.6393 respectively.) Traverse wares do not correlate highly with any of the artifact types tested; all associations both positive and negative are within the realm of chance. One artifact class which covaried highly with either Skegemog wares or Juntunen wares was end scrapers, which are distributed in a similar manner to Skegemog wares (f.l5887). Additionally, a high a (A378) was obtained between Juntunen ceramics and CI, T2 21+6 projectile points. Beyond these relationships, little temporal definition of chipped stone artifacts was possible. Due to the large number of high 3 values obtained between artifact classes, these data have been treated in a more general manner than above. For instance, grouped projectile points and unifaces covary highly (f. .589“), suggesting a functional association between the two. Within the projectile point and uniface categories there was also a great deal of high internal correlation. These artifacts behave similarly both within and between groups, and are distributed primarily in the vicinity of Feature 1. These materials are very probably associated with butchering activities. All core categories correlated with each other to at least £52200, perhaps an indication of specific core preparation areas or areas commonly used to discard cores. Low significance levels, however, tend to diminish the credibility of these hypothesis. "Constellations" of variables (Freeman and Brown 1961+; Brose 1970c), in this case tool types, which could be defined from the intercorrelation matrix, could be defined by two overlapping groups. Both constellations _ were also recovered in the vicinity of Feature 1. The first consists of eight chipped stone artifacts, CI, T2 projectiles, CI, Tl} projectiles, fragmented projectiles, gravers, drills, ovate bifaces, retouched flakes, and utilized flakes. Utilized flakes link this group and small triangular hafted knives, small triangular unhafted knives, and awls. The first cluster, mnctionally, could well have been a wood or bone working tool combination if it were notifor the presence of projectile points. It is possible that at times projectiles were used in butchering activities. Both gravers and drills are assumed to have functioned in such manners. The second constellation, with an awl, could have served for hide - r.- 21F? preparation. The high intercorrelation with artifact types not included in the first cluster and of unknown function other than perhaps as cutting implements would argue for a functionally specific group of tools. It is, of course, possible that projectile points were included in the first constellation, along with both utilized and retouched flakes, due to their high frequencies in the Feature 1 area; Feature 1 seemed to serve as a locus for a number of activities. Itiis due to this point, as well as its position within the longhouse (near the NE-SW wall), that it is proposed that Feature 1 is not contemporaneous with the structure but might: possibly predate it. The lack of both early and late Late woodland . ceramics in the vicinity of both structures, with the attendant difficulties in dating theselhabitation features, would suggest maintenance activities that take place away from dwellings.- If this is the case, the maintenance activities performed adjacent to Feature 1 would not relate to the longhouse but more likely“would be a focus of activity on the lakeward side of the earlier, circular structure. Discussion Patterns of intrasite spatial utilization that emerge from the inter- correlation matrices, as well as from the discussion of both chipped stone and ceramic distributions, are in many ways unclear. Ceramic distributions within both Area A and Area B point to localized areas of high sherd frequency, suggesting short term, small group use of the O'Neill Site at different periods of time. These data also demonstrate a lakeward shift through time. During the late Late woodland Juntunen and Traverse occupations, the small dwelling was placed in back of the partially stabilized dunes in Area A (Occupation Zone III). In the case of food processing, at least part of the activities connected with this process 2#8 were performed in the lakeshore dunes. Similar activities may'have been practiced by Traverse and Skegemog occupants as well, although the data only'support an association with core processing activities. The association of early Late woodland ceramics in what may'well be a functionally specific situation in the dune area could reflect a similar back-beach dwelling/lakeshore processing dichotomy as was generated for the later occupants. This argument would be strengthened if Feature 1 and the varied activities which seem to be associated'withiit do, indaed, date from this earlier occupation sequence. As a result of the multiple occupation of Area B by both early and late Late woodland populations, practicing both northern and southern ceramic decorative traditions, it is difficult to isolate activity areas that can be definitely assigned to a single occupation. The presence of both early and late structures on this part of the site, both constructed ' in different manners, is a good indication of use as a dwelling area. Isolation of the use of these houses by northern Juntunen/Mackinac or southern Traverse/Skegemog groups isznot possible; the ceramic distributions could argue for either. The local settlement pattern seems to be one in which most activities other than specific household tasks are accomplished around the peripheries of whatever structures are present. During Mackinac and Skegemog, or Mackinac/Skegemog, occupations (ca. A.D. 1000-A.D. 1200), the features that can be assigned with some certainty to this time period occur north and west of the circular structure: Features 1, 3, 35, the flint concentrations adjacent to Feature 35, and the early core preparation locus in Occupation Zone III. Somewhat tenuous, but compatible with the ceramic distributions, would be assignment of‘Features 26 and 27, to the northeast, to this time span. If Feature 1, in front of 2#9 the entrance to the early structure, was indeed the center for a :multiplicity'of activities, it could well have served as a communal center, while individual activities such as flint knapping took place on the peripheries of the campsite. Similar patterns are displayed by the Traverse and Juntunen occupations. The major noticeable change is the extended length of the dwelling, with two definitely associated hearths, sterile of artifacts, along the center (Features lb and 5). This is most likely a multi-family' 'dwelling, perhaps of two or more families. Although the increase in house size through time iSjprobably'illustrative of changing social structure as well, the data are too limited to draw fOrceful conclusions. Due to small sample size, it is impossible to test the seeming increase in homogeneity of ceramic style through time at the O'Neill Site. Spatial organization of the occupation closely-parallels that of the earlier occupations. That few ceramics were found in the longhouse is again compatible with peripheral activity areas. It has been argued that at least the people making Traverse ceramics used the lakeshore area as a core preparation and knapping area (Occupation Zone II). The later Juntunen Phase populations carried out either food processing or consumption activities in the grassy lakeshore hollows (Occupation Zones II and III), in the vicinity of Features 10, 31 - 34, 36 - 39. Feature 39,:a clay concentration, could relate to food processing, but more likely attests to ceramic manufacture. High Juntunen Ware frequencies and a radiocarbon date of A.D. lt55 from Feature 8, adjacent to the longhouse, indicate contemporaneity with the house, and once again illustrate what seem to be activity areas located near, but not in, the dwellings. 250 'No structures can.be assigned to the most recent occupationséof the O'Neill Site. The only possible part of a structure is the stone cairn or footing, Feature 40. If it is, indeed, a house footing then it is not part of the aboriginal occupation, although a stone cairn could occur with either European or aboriginal occupations. The singlegunflint frcm this stratum of N360/WllO, along with some eroded smooth ceramics, would date this area to the late eighteenth century, when both European or indigenous peoples could have used the site. The possible hearths in Occupation Zone I of Area A are within an area that yielded historic trade goods. A gun- flint (spall type), two tinkling cones, and a white bead were recovered from this vicinity. The nearby bone concentration would support a late date, especially since it contained bone fragments with possible iron tool butchering marks. This information would suggest a late seventeenth to early eighteenth century butchering area. The stone cairn or footing, Feature 40, may well be coeval with this occupation. Trade material distributions from Area B and Area A demonstrate that if this is a single short-term site, it was not highly localized. It may be a multiple occupation. Both O'Neill Curvilinear and Chevron Incised ceramics were recovered adjacent to Feature 12, a hearth. There is a flint concentration, with plano-convex cores, in close proximity to both the ceramics and the hearth. Three bone awls were recovered from nearby'N390/Wh0, and there was a high correlation between fish bone and ceramics in this occupation zone. The “utilization of space is therefore localized. Food processing, flint lenapping, and possibly sewing Or other household'tasks were performed near er hearth area. Although later in time, this locus of materials is reminiscent of the limited distributions of all ceramic types at the (J'Neill Site, attributable to short term, small group occupations. 251 In summary then, the spatial patterning of settlement throughout the O'Neill Site's use reflects separate dwelling and.maintenance activity areas. The latter are most often located around structures, and there is a distinct possibility of a communal hearth area during the early occupations. Ceramic distributions reflect confined occupation areas attributed to short term camps by small groups. Most commonly, the specific activity areas deal with feed preparation and flint knapping; while butchering activities, hide preparation, and the working of wood and bone took place in the same area, at least during the early occupations. Conclusion and Summagy Synthesis of the varied types of information set forth in the preceding sections allows a rather complete portrayal of the occupational sequence and functional specificity of the O'Neill Site. Primary among the varied activities which took place within the excavated areas was flint knapping. High corezartifact.and artifactzflake ratios demonstrate that the manufacture of chipped stone tools, or at least the preforms for future finishing, was of primary concern. Large amounts of raw material derived from the Piawan-go-ning Quarry Site at Norwood or from Hayden Point on Lake Charlevoix would suggest extraction from one of these primary sources. As a whole, the chipped stone industry is a flake tool industry, many of the artifacts having been produced on flakes less than 3 cm in length. Given the specialized activity which occurred at O'Neill, it seems to have been one of a series of sites in the vicinity of the quarries which functioned in such a manner. Other sites of similar function would include the Eastport Site, as well as Whiskey Creek, and ‘the Neff's A.H. Map Site. 252 Preparation of cores and the production of flakes involved a;series of options based on the size of the raw material and the types of fracture which occurred. Thus, the aboriginal flint worker, no doubt well versed in the production of chipped:stone implements, could make specific decisions based on a variety of criteria. Since the indistry is a flake tool industry, the options did not involve the discarding of smaller prepared, but fractured, cores. There is, further, some evidence of the use of indirect percussion for blade-like flake detachment. Extraction of raw material and the preparation of chipped stone implements was not, as might be expected, performed at the expense of maintenance activities necessary to the survival of the group. Faunal remains were abundant on certain parts