VARIABILITY IN THE ARCHAEOLOGICAL RECORD OF NORTHEASTERN EARL Y PALEOINDIANS: A VIEW FROM SOUTHERN ONTARIO An overview is provided of the Early Paleo indian occupation of southern Ontario focussing on temporal variability in site types, site frequency, site locations, fluted point morphology, tool kits, and lithic exploitation patterns. Through the use of long-accepted archaeological techniques of age estimation, including typological dating and a variant of contextual seriation, it is clearly shown that the three well-documented point types from the region (Gainey, Barnes [Parkhill], Crowfield) represent a temporal series. Several significant temporal trends are recognized, including: decreases in actual site numbers and findspots, reduction in site sizes, increasing association of sites with the strand line of what is either the main stage of pro-glacial Lake Algonquin or post-Algonquin Lake Ardtrea, apparent design changes in projectile tips suggesting variation in specific use contexts, reductions in "distance" mobility (as measured by decline in the distances between sites and stone source used), and changes in directions of interaction with groups in surrounding regions. It is argued that many of these patterns are duplicated in immediately adjacent areas such as Michigan, and that they can be mainly explained as a combined product of a) the movement of groups into previously uninhabited regions; and b) changes in subsistence practices resultingfrom increasedforest closure over time. These trends seem to contrast with areas to the east in northernmost New England and the Canadian Maritimes, perhaps because of differences in the nature and extent of environmental change between the regions. This paper presents a synthetic overview of the Early Paleoindian (i.e., fluted point associated) occupation of southern Ontario estimated to fall in the period from ca. 11,000 to just before 10,000 B.P. (see Ellis and Deller 1990; Storck and Spiess 1994; Storck, in press). The major focus of the paper is on variability in these occupations over time within southern Ontario and between the southern Ontario record and that of other areas of the Northeast/Great Lakes - variability which can be used as clues to cultural variation, especially adaptive strategies, among these groups. Prior to 1980, most investigators stressed the homogeneity of the Early Paleoindian record over large areas of North America and downplayed variability. We have argued elsewhere that the focus on the widespread similarities of Paleo indian assemblages at the expense of variability is not solely a product ofthe normative approach to archaeological study which characterized much archaeological research prior to the 1960s. Rather, it is also a heritage of the view that these Paleoindian peoples were the first inhabitants of the New World who spread very rapidly over a large land mass and, as a result, were virtually identical everywhere they were found (Deller and Ellis 1992a:25; Ellis 1984:3-5, 1993:604, 1994a: 416; Ellis and Deller 1988). In essence, investigators simply ignored variation and stressed instead the similarities or homogeneity of Paleo indian materials. The few individuals who drew attention to variation (e.g., Bonnichsen 1977) did so in order to argue that it undermined that model- a position with which we disagree. Regardless, hints of variabilility between the western and eastern North American Paleo indian record were evident by the 1960s, such as the dominance of the western record by kill sites and the east by occupation sites (e.g., Mason 1962). By the early 1970s, investigators were becoming aware of variation within eastern North America and particularly between northern and southern areas (MacDonald 1971). Subsequently, this north to south variation became better documented and investigators presented more formal models which explained such variation in terms of differences in Paleoindian adaptations (e.g., Gardner 1976; Meltzer 1984, 1985, 1988). Meltzer (1988), for example, argued for a highly mobile tundra or parkland adaptation including the communal hunting of caribou in the more northern areas of the east in contrast to a less mobile, complex boreal/deciduous forest-adaptation among peoples of the south. The more southern groups were argued to have had a more generalized subsistence base. Building upon such constructs, recent investigators have focussed on variation at even smaller spatial scales such as within the Northeast per se, and as well, have begun to suggest variation over time in that area (e.g., Deller and Ellis 1988, 1992a, 1992b; Shott 1986, 1989a, 1989b; Spiess and Wilson 1987; Storck and Spiess 1994: 137). Deller and Ellis (1988) suggested that the earliest northeastern groups were adapted to a more open spruce parkland and practiced some communal hunting of caribou, whereas later Paleoindians were seen as adapted to more forested conditions and were less involved in the communal hunting of game. Indeed, and as stressed by Wright (1989:356) and Deller (1989), the suggestions of temporal variation are so well established in certain areas of the Northeast that it is probably misleading to treat the Paleoindian record of any area as representing one homogeneous cultural system, and most certainly so in areas of marked environmental change. For this reason, we summarize the Early Paleoindian archaeological record of southern Ontario within the context of three recognized, temporally sequential, "complexes" or "phases": Gainey, Parkhill (Barnes) and Crowfield. The term "phase" seems to be most widely used in referring to these developments, particularly in Michigan (e.g., Shott 1986, 1993), and we will employ that term in this paper. It is used by us as an equivalent of an "archaeological culture" or as the smallest archaeologically defined unit within which we can not currently recognize any significant temporal-spatial variation (see Clarke 1968:146). The main distinguishing criteria of the suggested southern Ontario phases is variability in projectile point form. Comparative analyses of material from a number of sites and locations in southern Ontario, and elsewhere in the Great Lakes drainage, have clearly demonstrated the presence of at least three point types. Each of these types is distinguished by a consistent clustering of certain attribute states and restricted variable ranges and are referred to as: Gainey, Barnes and Crowfield (Figures I and 2). Gainey points (also called Bull Brook or Gainey-Clovis points; see Figures la-b, 2a-b) are relatively wide and thick (Figure 3a), essentially parallel-sided (Figure 4) projectile points with single, relatively short, yet broad, fluting. They tend to have deep basal concavities (Figure 3b) which are formed as simple arcs, and sometimes have fishtails. They also appear to have been relatively long points with fore-sections which projected, at least prior to resharpening, much beyond the hafted or bound base (e.g., Figure Ib; for more detailed descriptions of Gainey and the other types, and supporting documentation, see Deller 1988; Deller and Ellis 1984, 1988, 1992a:35-48; Ellis 1984; Roosa 1965; Shott 1993:79-103; Simons et al. 1984; Storck 1983,1990; Wright 1981). Barnes type points (Figures 1c-d; 2c-f; Roosa 1965, 1977a; Wright and Roosa 1966) are diagnostic of the Parkhill phase. These are often fishtailed points which expand moderately from the base (Figure 4) to a position of maximum width at, or if unresharpened, just below, mid-point. Like Gainey points, the foresections on unresharpene<}points, or at least, what appear to be the earliest dating Barnes points such as those from the Thedford II site (Deller and Ellis 1992a), seem to have been quite long prior to resharpening (e.g., Figure 2e). Barnes points have thicknesses and basal concavity depths intermediate between those of Gainey and Crowfield points (Figure 3a, 3b) and have relatively narrow bases (under 20 mm; Figure 3c). The basal 5 eM concavities instead of being simple concave arcs, can have a straight top/distal margin. Fluting tends to be single and long, often extending to within 10 mm of the tip end on at least one face. The distinctive Crowfield points are characterized by being very thin «5 mm; Figure 3a) and yet wide (25-35 mm), with shallow (ca. 2-4mm; Figure 3b), often "squared-off' (e.g., straight top margin), basal concavities (Figures le-(2g-h). They expand markedly (Figure 4g) from a narrow base «20 mm; Figure 3c) to a point of maximum width at or above mid-point. Fishtails are lacking. Fluting is exceptional, given the thinness of the points, but it is often multiple with up to three flutes on a face. Judging from final stage Archaeology of Eastern North America Figure 2. Fluted bifaces from various locations in Ontario. a-b: Gainey Type; c-f: Barnes Type; g-h: Crowfield Type (h is a preform). a-b and f-g are reproduced courtesy of the London Chapter, Ontario Archaeological Society from Ellis (1987a, I987b) and Ellis and Deller (1990). e is reproduced courtesy of the Museum of Anthropology, University of Michigan from Deller and Ellis (1992a). i ! :r UJ 8 :;:6 0 >•... >4 ..• G' « u 0 z 161 preforms at sites such as Crowfield (Deller and Ellis 1984) and Bolton (Deller and Ellis 1996; see Figure 2h), fore-sections on Crowfield points were uniformly quite short even prior to resharpening. This characteristic contrasts with known examples of the other types. A very distinctive aspect of Crow field points is the resharpening of tip ends. The fore-sections can be resharpened diagonally to the long axis to form straight edges on each side immediately adjacent to the tip margin. When combined with the fact that lateral basal edges expand markedly from the base, the overall result is a pentagonal outline shape. Some Crowfield points are also known with a shoulder and resharpening on only one lateral fore-section edge (Deller and Ellis 1984:44-45, 1996). These shouldered items appear to be hafted knives as opposed to projectile tips. u ~2 « '" CD 0 2 i !24 :r •... 0 ~ 3' ...J ~20 1141 G2 « CD 16 Figure 3. Summaries ofthickness (A), basal concavity depth (B) and basal width (C) measurements for fluted point samples (after Deller and Ellis 1992a). Numbers in brackets indicate sample sizes. G 1: Gainey Points, Gainey Site (estimated from a figure in Simons et al. [1984] so only mean basal concavity depth given); G2: Gainey Points, Sites and Findspots in Ontario; B I: Barnes Points, Small Sites and Findspots in Ontario; B2: Barnes Points, Thedford II Site, Ontario; B3: Barnes Points, Parkhill Site, Ontario; B4: Barnes Points, Barnes Site, Michigan (after Voss [1977]; Wright and Roosa [1966]; some statistics omitted because of small sample size or unavailability of data); B5: Barnes Points, Fisher Site, Ontario; C: Crowfield Points, Crowfield Site, Ontario. Obviously, we would like to have extensive supporting stratigraphic data, as well as 14Cdates, to substantiate the idea the point types represent a temporal sequence. However, these are currently not available. As well, we have severe doubts that 14Cdates have the resolution necessary to track the time changes of concern to us here (see Ellis 1994a). This lack of resolution is especially a problem in the time period of interest (e.g., Late Pleistocene - Early Holocene). At that time it is becoming clear that there are long periods of real or sidereal years (up to 400+ years) within which there is little change in radiocarbon dates (see Becker et al. 1993; Lotter 1991:326-327). There may even be some reversals in radiocarbon dates during that period (e.g., sites later in real time could have earlier radiocarbon dates; see Fedje et al. 1995:9395). As a result, we have focussed on other means in order to evaluate whether or not the types represent a time series or are contemporary. For example, it has been shown that the differences between the types are not due to numerous potential sources of contemporaneous variation mentioned in the literature (see Gardner and Verry 1979; Judge 1973; MacDonald 1971; Roosa 1963, :01 ~~-9"-i~""i -ri[l] BARNES (ISOLATED FINDS; N=81 ••. ,D~98-i ~~i~lb~4~i ••... n""'i ~..--r-i'--IT""\O""" C 4 -~i ~. i 0 •.......•... 92~i~"""""" 21 n UolJ 98' riARNES (THEDFORDII SITE; N=BI , 164i i D do" Figure 4. Measurementof face-angle(degreeof expansionof pointsfrom base), various sites (reproducedfrom Deller and Ellis 1992a with permission of Museum of Anthropology, University of Michigan). Measurement records angle betweenlinedrawn along lateraledge and linedrawnacrossbaseof point (e.g" betweenapex of point ears).Measurement is averaged for each comer (if both are intact)per point. Parallel-sidedpoints are ca. 90°, whereas points which expand from base are >90°. 1965). These factors would include: resharpening or recycling; at a general level, use (e.g., all were used as projectile tips although this does not preclude their use for other purposes); "social" stylistic variation; idiosyncratic behaviour; discard of bifaces in manufacture instead of an exhausted used state; differences in the distance to raw material source employed; and variation in the raw material sources exploited. These causes can only be shown to account for variation within each type - not the differences between them (Deller and Ellis 1992a:34-48; Ellis 1993:605-607). As an example, since we have complete manufacturing sequences for all of the types, including late stage preforms which approximate closely the finished product (see Deller and Ellis 1984, 1992a, 1996; Simons 1996; Simons et al. 1984; Storck 1983), it is clear that each type was initially designed and made as such and not made by recycling or resharpening an example of one of-the other types. However, within each type, variation can be found which is due to recycling or resharpening such as the production of pentagonal outlines on some Crowfield points. As another example, all points were used as projectile tips based on the presence of distinctive tip impacts, a primary focus on tip as opposed to lateral· edge resharpening, and certain patterns or breakage and discard (Ellis 1993:607). Yet, within types there can be evidence of use variation such as the shouldered Crowfield "knives" noted earlier which differ from the normal unshouldered Crowfield projectile point. As a final example, since all types can occur on the same raw materials it is clear the types are not a product of raw material variation. Once again though, variation Barnes Basal Finishing X X Presence of Fishtails X X Medial Ridge to Guide Flutes X X Maximum Width Around Mid-Point Straight Margin to Top of Basal Concavity within a type can be explained using this variable, such as variation in length of Barnes points at some sites (Deller and Ellis 1992a:27-28). Another major approach to assessing the types as a measure of temporal variation is to use long-established and venerable archaeologically developed techniques of relative age estimation - techniques which, except where dendrochronology can be employed, have produced the most refined chronologies currently used by archaeologists. We have argued (Deller and Ellis I992a; Ellis I994a:420) that the fluted point types exhibit variation of a form which, using principles of typological dating (Montelius 1903), can only be explained as indicating change through time. In essence, comparison of the three types almost always results in the intermediate placement ofthe Barnes type between that of Gainey and Crowfield in terms ofform and manufacturing procedures. For example, and as alluded to in the brief descriptions of the types given above, Barnes points are intermediate in terms of certain discrete attributes between the other two types. In other words, they share several characteristics with either Gainey or Crowfield points but not both. As shown by some examples in Table I, Barnes points are like Gainey points and unlike Crowfield points in that they can exhibit the removal of a broad flake over the base of the flute ("Barnes Basal Finishing"), often possess fishtails and can have a medial ridge formed by lateral flake terminations which served to guide flute removals. At the same time, however, Barnes points can be like most Crowfield points in that they can have a straight apex to their basal concavities and can have a point of maximum width around mid-point - features rare to non-existent on Gainey points. Similarly, in terms of point variables such as basal concavity depth, point thickness, and degree of expansion of the lateral edges from the base, Barnes points always fall between the other two types in the Ontario samples (Deller and Ellis 1988: 255, 1992a:41-45; see Figures 3 and 4). Moreover, one can posit good reasons for cases of continuous variables where Barnes points do not occupy an intermediate position. For instance, although Gainey points have wider bases (>20 mm), both Barnes and Crowfield points have bases of a similar, albeit smaller, width (ca. 12 to 20 mm; Figure 3c). Crowfield points do not have a narrower base than Barnes points as might be expected, but this exception can easily be attributed to technical/functional constraints. In short, basal widths, which are undoubtedly related to shaft/haft diameter, can only become so narrow. With anything less, the shafts would be too weak to use as either projectiles or knives (e.g., they would snap too readily). Overall, Barnes points clearly are intermediate between the other two types in almost all aspects ofform and manufacture. These contrasts between the types could be a product of contemporaneous variation in space with one type grading into another. However, if this was the case, the types should exhibit regionalization and c1inal variation. Since the sites and findspots representing each type occur over the vast area of southern Ontario and adjacent areas as well as regions farther afield, this explanation is easily rejected. Thus, there seems to be no viable alternative explanation for the consistent ordering of the Barnes type between the other two forms except to argue they represent a time series. The Gainey point type is the closest Great Lakes analog to Clovis points, and Crowfield points most closely resemble Late Paleoindian unfluted points and particularly, those of the Holcombe type. On geoarchaeological grounds, these Holcombe points post-date fluted points (Ellis and Deller 1986). This information by itself supports the idea the types are a temporal series (e.g., it is the only way to explain why one type, and not all, more closely resemble, alternatively, the earliest dated fluted point type elsewhere, or the demonstrably later Holcombe type). It also strongly suggests Gainey is earliest and Crowfield is the most recent. Barnes points, and the Parkhill phase of which they are diagnostic, are intermediate in form and manufacturing procedure and thus, in time. The types are "stylistic" in the sense they can be used as time markers. However, we wish to stress that this conclusion does not mean the change in the types over time is due to stylistic (that is, random, selectively neutral) forces. In fact, we believe it is possible that the variation is "functional" in the sense of Dunnell (1978); that is, it had a selective value. As Shott (1989a) has suggested, one can argue that the changes in projectile point design relate to alterations in hunting strategies over time. If the types vary at least partially due to functional reasons, it follows that variation in projectile point form and design from area to area in the Northeast might provide clues as to adaptive variability and also, that points with similar'characteristics in widely divergent areas need not be contemporaneous. Besides typological dating, one can also use a variant of contextual seriation, first employed by Flinders Petrie (1899), in order to evaluate the idea that the types represent a time series. Petrie (1899) attempted to order a series of Egyptian graves through time. He noted that the graves each included a series of pottery types. Petrie (1899) assumed that each of these types had a different although overlapping period of use. On this basis, one could arrange the graves in time with graves closer together in time having a similar series of types while those farther apart in time shared fewer types because types would be added or deleted over time. In the present context, we are dealing with essentially single types of artifacts or are suggesting only one of the three point types was in use at one time. Therefore, we can not use Petrie's (1899) exact method but must modify it somewhat. As we have argued elsewhere (Deller and Ellis 1,988:255), the vast majority of fluted point sites in the eastern Great Lakes area, including even the largest ones, yield only points of one type (see Table 2). This pattern is best explained in terms of temporal variation. Ifthe point types were contemporaneous, we would expect more co-occurrences of the types. However, even the few sites with more than one fluted point type can be informative in evaluating the types as temporal markers. The Early Paleoindian occupations of the eastern Great Lakes coincided with very rapid and marked environmental changes. In turn, these changes would have caused marked variation over time in most of the specific locations with resources suitable for attracting human occupation. Therefore, most locations would be attractive for only part of the span of the Paleoindian occupation and hence, would yield points of only one type (Deller 1989: 199). However, one could expect exceptions and specifically, at sites which were in locations attractive to occupation for periods spanning the duration of use of more than one of the types. Thus, we should expect the types which co-occur together to be those which were closest in time. Conversely, and using the same logic underlying Petrie's (1899) contextual seriation, types whose time of use was widely separated should not co-occur or only very rarely co-occur. A corollary of this idea would be that areas of longer term land use (e.g., multiple phase representation at sites) would be ones of relative environmental stability. We will return to this idea later in this paper. Table 2 includes Early Paleoindian sites in the eastern Great Lakes area where multiple types occur. On the assumption there might be more continuity from the latest Early Paleoindian material into Late Paleoindian times, this table also shows sites where Late Paleo indian material occurs. It is clear based on this listing that the distribution of co-occuring types is not random. Indeed, point types which on the basis of typological similarity should be cleJser together in time are almost exclusively the ones which co-occur. For example, there are four sites where Gainey and Barnes points occur and none, excepting Udora, where Gainey and Crowfield types co-occur. Moreover, even at Udora the intermediate Barnes points occur, suggesting use of SITE Gainey, Michigan Butler, Michigan Leavitt, Michigan Nobles Pond, Ohio Lamb, New York Graham, Ontario Snary, Ontario Kolapore, Ontario Uniondale, Ontario Baker, Ontario Ward, Ontario Ferguson, Ontario Haunted Hill, Ontario Hunt, Ontario Potts, New York Banting, Ontario Weed, Ontario Udora, Ontario Corditaipe, New York Dutchess Quarry Caves, N.Y. Devil's Nose, New York Glass, Ontario Schoefield, Ontario Parkhill, Ontario Thedford II, Ontario Dixon, Ontario McLeod, Ontario Mullin, Ontario Mawson, Ontario Stott Glen, Ontario F. Wight, Ontario Fisher, Ontario Babula, Ontario Barnes, Michigan Hussey, Ontario Watpool A, Ontario Crowfield, Ontario Bolton, Ontario Alder Creek, Ontario Zander, Ontario Reagen, Vermont GAINEY X BARNES CROWFIELD LATE PALEO-INDIAN REFERENCE X X X X XC?) X X X X X XC?) X X X X X X X Simons et al. 1984 Simons 1995 Shott 1993 Gramly and Summers 1986 Gramly 1988 Storck 1982 Wortner and Ellis 1993 Storck 1984 Deller 1988 James Keron: pers. comm. Deller and Ellis 1988 Deller and Ellis 1988 Ellis and Deller 1990 Ellis and Deller 1990 Lothrop 1988 Storck 1979 Deller and Ellis 1988 Storck 1982 Funk and Wellman 1984 Funk and Steadman 1994 Tankersley 1994 Deller and Ellis 1992b Deller 1988 Roosa 1977a, 1977b Deller and Ellis 1992a Deller and Ellis 1992b Deller and Ellis 1992b Deller and Ellis 1992b Deller and Ellis 1992b Deller and Ellis 1992b Deller and Ellis 1992b Storck 1983 Deller and Ellis 1992a Wright and Roosa 1966 Storck 1979 Storck 1982 Deller and Ellis 1984 Deller and Ellis n.d. Timmins 1994 Stewart 1984 Ritchie 1953 that locality throughout the time of Early Paleoindian occupation. Similarily, and as expected, the Late Paleoindian items co-occur mainly on sites yielding Crowfield points. Exceptions are the Hunt and Butler sites which have yielded both Gainey and Late Paleoindian Holcombe points. Once again there are clear patterns here which we believe can not be ignored. There seems to be no viable alternative explanation for these patterns except that the types represent a time series, and especially because the ordering is exactly the same as that produced by the typological dating (e.g., we have multiple lines of independent evidence which point to the same conclusion). In summary, if one accepts typological dating and contextual seriation as valid constructs (and the last last one hundred years of research have only served to support their viability and validity), these lines of independent evidence, as well as several other arguments detailed elsewhere (Deller 1989; Deller and Ellis 1988, 1992a, 1992b), point to the same inevitable conclusion: the point types, and the phases of which they are diagnostic, represent a time series. Gainey Phase Gainey sites and findspots seem to be more Strandline Phase N common in southern Ontario than those representing 8(66%) 4(33%) Gainey 12 the other suggested phases (see Tables 3 and 4 and 12(75%) 4(25%) Parkhill (Barnes) 16 DeIler and Ellis 1992: Figures 32 to 34). Some 12 4(57%) 3(43%) eld 7 sites with Gainey components are now known from southern Ontario (Figure 5) and at least five other sites have been documented which, although they Table 3. Frequency lack the diagnostic fluted points, have wide channel southern Ontario. flakes and other characteristics which strongly suggests a Gainey affiliation (e.g., Ellis et al. 1992; Jackson 1994). However, with the exception of some of the latter, only presumed Gainey sites, none have seen extensive excavation and/or full publication. This situation limits our knowledge of this material, especially with regard to tool kit composition. Regardless, Gainey sites tend to be very widely distributed in southern Ontario. Some sites, such as Haunted Hill (Deller 1979) and Udora (Storck and Spiess 1994), are known in the modem Lake Huron basin on or in the vicinity of the real or inferred strandline of what was, depending upon which of two available geological model of lake levels one foIlows (see Karrow and Calkin, 1985; Storck, in press), either glacial Lake Algonquin or a post-Algonquin, Upper Group, Lake Ardtrea (hereafter, for brevity, Algonquin; in both models the lake dates to ca. 11,000 to 10,400 B.P. - only the name is different). However, Gainey sites seem to be rarer in the vicinity of that strandline than subsequent fluted point sites (see below). Most Ontario Gainey sites appear to be small, having only one or two clusters of artifacts. A possible exception is the Udora site. Udora may have as many as 10 discrete clusters of artifacts but since it is multicomponent, having also yielded Barnes and Crow-field points (see Storck 1982: Figure 7e, 7f, 7g; Storck and Spiess 1994: Figure 2), it is difficult to precisely interpret. Regardless, the clusters at all sites tend to be under 200m2.each in extent and artifact yields are low «60 tools/preforms). Large, multi-locus Gainey or comparable sites are not known in Ontario although, of course, they are known from surrounding areas such as at Bull Brook - Massachusetts, Nobles Pond - Ohio, Butler - Michigan, and of course, Gainey - Michigan. We strongly suspect that the absence of large Ontario sites is a product of sampling error and expect these sites will be reported in the future. Gf especial interest is the fact the Udora site in Ontario has yielded some preserved faunal remains including caribou (Rangifer tarandus), arctic fox (A/apex /agapus), and hare (Lepus sp.) in association with Gainey phase materials (Storck and Spiess 1994). ICmwfi Lake Huron SITES GAINEY PARKH I LL I BARN ES e • • •• CROWFIELD MULTI-PHASE CHERT OUTCROPS ONONDAGA KETTLE BAYPORT POINT COLLINGWOOD 1~0 Km 200 I 6. o Because of the small size of known asPhase semblages and incomplete reporting, it is difficult Gainey to characterize the stone tool kits of the Gainey Parkhill (Barnes) phase as it is represented in Ontario. Certainly, the Crowfield assemblages include pan-Paleoindian types or classes such as trianguloid end scrapers, concave Table 4. Frequency of fluted point findspots by type, side scrapers, beaks, and various forms of southern Ontario. gravers/micro-piercers. Some less widely distributed and more specific types also occur, albeit rarely. These include pieces esquillees, the occasional burin and, in at least two cases, fluted points which were subsequently recycled into drills (e.g., Storck and Spiess 1994: Figure 4a). The recycled drill forms are unknown on sites of other phases. Lacking in reported assemblages are a series of very distinctive types we have recognized in other Paleoindian asssemblages such as miniature points on channel flakes, large alternately beveled bifaces, backed bifaces and hafted perforators (see Ellis 1994c; Ellis and Deller 1988). However, these artifact forms are rare in those other assemblages. Moreover, given that our overall sample of Gainey phase artifacts in Ontario is very small, and consists of material from smaller sites which tend to yield more limited tool inventories than larger ones, this absence could easily reflect sampling error. In any case, we currently recognize 25 tool types and classes (Table 5) which we believe, as near as can be determined from formal characteristics (Ellis and Deller 1988), are monitoring as closely as possible use variation in Early Paleoindian assemblages. Only 13 or 52% of these tool types/classes Archaeology of Eastern North America are known from Gainey phase sites, as opposed to 68% or more from sites of subsequent X X phases (Table 6). X A number of chert sources were exploited X X by Early Paleo indian groups in southern OnX X tario (Figure 3). Although geological source X X assignments were originally based on detailed X study of macroscopic characteristics, chern ical and physical (e.g., thin section) analyses have X X only served to confirm the assignment of raw X X material to certain sources (Ellis 1984:42-45; X Storck and von Bitter 1989). As well, whenX ever original unflaked surfaces of the raw maX X terial are present on finished tools, they almost X X always indicate the use of primary rather than X secondary deposits. Gainey phase sites in OnX X tario seem to follow a pattern seen in other X X X Ontario sites in that specific assemblages are X X dominated (>80%) by material from one X X source. The specific source used (see Figure X 3) varies by site. Assemblages from Snary, X X Uniondale, and Ferguson are dominated by X Fossil Hill (Collingwood) chert whereas OnX X ondaga chert predominates at Ward, Selkirk X X chert at Baker, and apparently, Upper Mercer, Table 5. Distribution of artifact types/classes by phase. Ohio chert at Weed. Present data suggest three ways in which lithic source use of Gainey peoples differed from those of later groups. First, certain materials seem to have been employed only by Gainey peoples. These would include Upper Mercer, Ohio cherts and even, as demonstrated by two isolated finds from southcentral and southwestern Ontario respectively (e.g., Roberts 1984: Plate 5), quartz crystal. The latter material apparently originates in the eastern part of Ontario and is also found as debris and unifaces on some other probable Gainey sites in that area such as at the Halstead and Sandy Ridge sites (Jackson 1990: 125). Second, chert from certain sources which was used by later groups such as Kettle Point, Ontario and Bayport, Michigan, was rarely if ever used by Gainey phase peoples in Ontario. Indeed, the only known exception is a single Gainey point base on Bayport. Finally, in contrast to sites oflater phases, Gainey phase sites can be found where the primary raw material (>80% of assemblage) employed at the site had to originate over 200 km away. Examples include the Snary site which is largely on Fossil Hill chert from 260 km to the northwest (Wortner and Ellis 1993) and Weed, the Gainey component from which is apparently on Upper Mercer chert from some 350 km to the south (Deller 1988). Regardless of how one converts these distances to actual measures of mobility such as territory size, to date, sites of other phases'have not been found where the distance between the site and main chert source employed exceeds 200 km. Gainey Artifact Type X Fluted Points Shouldered Fluted Points Large Beveled Bifaces Backed Bifaces Channel Flake Points X Drills Leaf-Shaped Bifaces X Pieces Esquillees X Other Bifaces X Trianguloid End Scrapers Large Parallel-Sided End Scrapers Offset End Scrapers Narrow End Scrapers X Other End Scrapers Proximal End and SideScrapers X Concave Side Scrapers X Other Side Scrapers Backed and Snapped Unifaces Denticulates/RetouchedFlakes X Beaks X Hafted Perforators Micro-Piercers X Chisel Gravers NotchlBorer/Denticulates X Other Unifaces X Parkhill Crowfield Parkhill Phase The Parkhill phase is probably the best known of all Ontario fluted point phases because extensive excavations have been carried out at several sites such as Parkhill (Ellis 1994c; Roosa 1977a, 1977b), Thedford II (Deller and Ellis 1992a) and Fisher (Storck, in press). Overall, more sites of the Parkhill phase are known in Ontario than of the Gainey phase, with at least 16 sites being 0/0 Phase N reported (Figure 3; Tables 2 and 3). However, these figures are Gainey 13/25 52% biased. At least initially, much attention was directed to locating Parkhill (Barnes) 21/25 84% sites in Ontario along the abandoned strandline of Lake Algonquin Crowfield 17/25 68% (e.g., Deller 1979; Storck 1982). Parkhill phase sites seem very common in the vicinity of that lake (e.g., within one or two km Table 6. Percentageof tool types/classesby of the lakeshore) whereas Gainey sites are clearly relatively rare phase. in those locations. Only three (25%) of the Gainey sites are reported in the Algonquin vicinity whereas 12 (75%) of the sites with Parkhill phase components are in that area. Indeed, although certainly present, only four definite, yet very small, Parkhill phase sites are known in "interior" areas inland from the strand line area in Ontario: Mullin, Stott Glen, Glass, and Babula Farm (Deller and Ellis 1992b). If we were to assume that interior sites removed from the strand line of Algonquin are more representative of overall frequencies, Gainey sites are more common (Table 3) and would be even moreso if we include probable Gainey phase sites lacking the actual fluted bifaces such as Halstead and Sandy Ridge (Jackson 1990, 1994). Similarly, it is clear from maps of findspots that Gainey points are much more common than Barnes points in areas removed from the strandline (compare Deller and Ellis 1992a: Figures 32 & 33). As shown on Table 4, overall Gainey findspots are two to three times as common as Barnes or Crowfield point findspots. As was the case with Gainey phase sites, known Parkhill phase sites are predominantly small (<200 m2) with low artifact yields «40 tools/preforms). These sites include not only small domestic sites but also, apparently at least, one small kill/butchering site (F. Wight; see Deller and Ellis 1992b:31). However, one moderate-sized single locus site (ca. 700 m2) is known: Thedford II (Deller and Ellis 1992a).ln addition, there are two multi-locus sites with debris spread over several hectares: Parkhill (Roosa 1977a, 1977b) and Fisher (Storck 1990, 1994, in press). Evidence, such as differences in raw material profiles between the site areas, and a lack of overall patterned layout in the distribution of those areas, suggests the largest sites represent multiple rather than a single occupation (Deller and Ellis 1992b:29-31; Stewart, in press). Yet, at least Parkhill also has been suggested to be an aggregation site because of the large spatial extent of some individual occupation areas; that is, there are single areas at Parkhill which cover up to 2000m2 and seem to indicate a larger population occupying those particular areas (Deller and Ellis 1992b:29-31). Interestingly, and in contrast to Parkhill and Fisher, other large sites in the Northeast such as Nobles Pond (Gramly and Summers 1986), Vail (Gramly 1982) and Bull Brook (Grimes 1979), have an overall patterned arrangement of the different loci, similarities in raw material use between areas, and artifact refits, which suggest most areas at those sites were contemporaneous (Seeman 1994; Spiess 1984). The largest Parkhill phase sites, Parkhill and Fisher, are similar in at least three other respects of concern to this paper. First, they are both located on the real or inferred Algonquin strandline. Second, despite the large areal extent of these sites (and at least at Parkhill, the large extent of some of the individual site areas) artifact yields are relatively low versus the very large numbers of artifacts (and particularly hafted, curated artifacts with extended use-lives such as trianguloid end scrapers) recovered from site areas at other large northeastern sites such as Debert, Nova Scotia (MacDonald 1968), Bull Brook, Massachusetts (Grimes et al. 1984) and even Great Lakes area Gainey phase sites such as Nobles Pond, Ohio (Seeman 1994). Third, and unlike most other large, multi-locus sites in the Northeast with the exception of Debert (MacDonald 1968:131-133), there is evidence that certain of the loci at these Parkhill phase Ontario sites witnessed different, in some cases relatively specialized, activities. For example, at Parkhill at least some of the loci are dominated exclusively by points broken in use, and debris from the manufacture of the same; whereas others have a broader, unspecialized inventory (Deller and Ellis I992b:28-29). Similarly, at Fisher, some areas are notable for the presence of large amounts of primary flaking debris whereas others have little evidence of such lithic reduction activity (Stewart, in press). Largely as a result of the extensive excavations at several larger Parkhill phase sites, the stone tool kits are quite well known. Besides pan-Paleoindian tool forms, Parkhill phase sites have yielded a number of distinctive bifacial and unifacial tool forms which do not seem to occur on Gainey phase sites. This difference has led to the suggestion that flaked stone tool kit size and complexity increases over time (e.g., Deller and Ellis 1988:258; Simons et al. 1984:266; see Table 5). However, it is just as easy to argue that the absence of some types from Gainey sites is due to sampling error. One can also question the reality and significance of some of these types, particularly some simple unifacial ones such as offset end scrapers and backed and snapped unifaces (Ellis and Deller 1988), because sample sizes for certain types from Parkhill phase sites are small (Storck and Tomenchuk 1990; Storck, in press). Nonetheless, most of these types, such as well-made bifaces with a flat back along one edge (backed bifaces), large ovate bifaces with fore-section edges resharpened by unifacial beveling, hafted perforators, and narrow/nosed end scrapers, occur in other phases, and often appear in large numbers. A good example is the Crowfield site where no less than 14 backed bifaces and three large beveled bifaces were recovered in a single feature (Deller and Ellis 1984:46; Deller 1988). Also, several of these types are reported from areas much beyond the Great Lakes. For example, the large beveled bifaces and narrow(nosed end scrapers are widely reported on several western Paleoindian sites from Texas to Colorado (e.g., Frison and Bradley 1980: 33-38, 109; Irwin and Wormington 1971:30; Johnson and Holliday 1987: Figure 9.6a; Wilmsen and Roberts 1984: Figure 94f). In fact, excellent examples of beveled bifaces were initially recognized as a distinct tool form by F. H. H. Roberts (1935:24-25) at the first excavated Early Paleoindian occupation site: the Folsom site at Lindenmeier, Colorado. Undoubted examples of hafted perforators (which are often lumped in with limaceslflake shavers or simple beaks in descriptions but are clearly distinct) are also reported from sites as far away as Maine (e.g., Gramly 1988a: Plate 19, three artifacts at top left). There should be no doubt as to the existence of the majority of these types. Despite the large number and size of available assemblages, what are often considered typical northeastern Paleoindian tool forms are very rare or absent on Parkhill phase sites. In other words, it is unlikely their absence is due to sampling error. Notable examples include fluted d611s (whether they were initially made as such or simply made by recycling points) and pieces esquillees. Moreover, if the items called limaces (e.g., Gramly 1982:37-40) are typologically the same as what Grimes and Grimes (1985) call flake shavers, then we can state with assurance that such items are absent on Parkhill phase sites and in fact, are not reported from any Ontario sites. A final aspect of the Parkhill phase which deserves discussion is lithic raw material procurement patterns. Almost all known sites in Ontario are dominated (>80% of the particular assemblages) by Fossil Hill formation (Collingwood variant) cherts from southcentral Ontario. One possible exception is the small interior Glass site whose two artifacts are on Onondaga cherts (Deller and Ellis I992b). In no cases, and in direct contrast to the situation in the Gainey phase, does the distance between the main source employed and the site exceed 200 km. The Parkhill phase is also distinctive from Gainey in that there is no evidence of the use of Upper Mercer, Ohio cherts or of quartz crystal. Yet, the use of more northerly located sources such as Bayport, Michigan chert from outcrop sources and, to a much lesser extent, Kettle Point, Ontario chert from secondary deposits, becomes more common on Parkhill phase sites (Deller 1989; Deller and Ellis 1992b). Crowfield Phase Crowfield phase sites are the rarest of all known Early Paleoindian sites in southern Ontario with only seven sites being reported (Figure 3; Tables 2 and 3). Findspots are also rare versus those of earlier phases, particularly when compared to the number of known Gainey findspots (Table 4; see also Deller and Ellis 1992a: Figures 32 to 34). Over half of the known sites are on or near the inferred pro-glacial Lake Algonquin shoreline, but given the biases towards surveys in those areas this association may be fortuitous. All sites presently known are relatively small, covering less than 200 m2. Known sites have yielded less than ca. 30 tools each. An exception is the Crowfield site (Deller and Ellis 1984) which, in addition to a small amount of normal occupation debris, included a single ceremonial feature containing over 200 complete heat-fractured tools and preforms. The tool kits of the Crowfield phase seem similar to those of Parkhill in that they can include, besides pan-Paleoindian forms, large unifacially beveled bifaces and backed bifaces, and lack items such as pieces esquillees. In apparent contrast to the Parkhill phase, drills have been recovered from Crowfield phase contexts. Fragmentary drill fore-sections (but lacking tip ends) were recovered from the Crowfield site. The Crowfield site itself also yielded several small distinctive bifaces with "leaf-shaped" outlines and one somewhat thickened end (e.g., Deller and Ellis 1984: Figure 17a, 17b) which have so far not been reported from any other fluted point site in the area. However, if the Crowfield site feature assemblage was an individual's tool kit, these items could be a product of idiosyncratic behaviour. Ontario Crowfield phase sites are known to be dominated by either Fossil Hill (Collingwood) or Onondaga cherts, and no site is known which is located more than 200 km from the main source of chert used. Some isolated finds are known on Kettle Point, Ontario and Bayport, Michigan cherts, but these are rare, perhaps due to the small available sample of points (Deller 1989). Based upon the foregoing, we summarize in Table 7 the suggested changes/time trends in the Early Paleo indian archaeological record of southern Ontario. As stressed above, some of these changes, such as those involving differences in tool kit size, may be a product of sampling error. This possibility can only be resolved through continuing research. However, it is very unlikely that most of the differences in site and findspot distributions, the total number of findspots reported, differences in projectile point form, the use of different raw materials, etc., are due to sampling error. It is possible to explain some of the suggested changes as indicating a gradual process of colonization of southern Ontario. For example, the indications of reduction in ranges exploited from Gainey to the later phases could be explained as a result of the gradual filling up of the landscape by the earliest colonists and, with population growth, the addition of restrictions on movement and reduction in the size of the areas which could be exploited. As another example, there are shifts in the exotic chert materials (e.g., from sources outside the province) which occur in assemblages and presumably indicate changes in the patterns of interaction of Ontario groups with those in surrounding areas, as well as changes in the ranges exploited. Assuming the earliest groups entered the lower Great Lakes area from the south/southwest, it is perhaps not surprising that the only sites or findspots where southern-derived cherts (such as from Upper Mercer, Ohio) were employed, are the earliest. One could explain, for example, the fact the Weed site assemblage is almost all on this material as a product of groups from Ohio gradually extending their range into southern Ontario and adjacent areas such as Michigan. Similarly, the small amounts of Upper Mercer seen in other Gainey assemblages/findspots, but the general lack of Michigan cherts, suggests more interaction with the "parent" groups to the south from which these peoples presumably derived (Deller 1989:215; Deller and Ellis 1988). However, over time, as southern Ontario and adjacent areas to the west in Michigan were populated, there were shifts in interaction patterns towards these adjacent areas. One example is the increasing frequency of Bayport, Michigan chert use in Ontario by the time of the Parkhill phase. While it may be possible to explain some of the trends as due simply to the colonization and infilling of the landscape in Ontario and surrounding areas, other suggested changes can not be explained using such a construct. Elsewhere (Deller and Ellis 1988; Shott 1986, 1989a), possible alterations in subsistence patterns havebeen suggested as causes of these changes. While Paleoindians in other areas of the Northeast may have inhabited solely more closed boreal forest landscapes, as argued by researchers such as Custer and Stewart (1989), there is no indication this was the case in southern Ontario. Rather, the period from 11,000 to 10,000 # 1 2 3 4 5 6 7 8 9 10 Change Decline in number of "interior" sites removed from Algonquian/ Ardtrea strandline Decline in overall frequency of sites and findspots over time Reduction in frequency of large, multi-locus sites Suggestions of increased complexity of spatial organization of activities at larger multi-locus sites Decrease in use of "southern" chert sources; increase in use of "northern" chert sources Decrease in distances between sites and main chert sources used Increase in stone tool kit size (?) Disappearance of tool forms such as pieces esquillees and burins (?) Appearance of new tool forms such as backed bifaces, large alternately beveled bifaces, and narrow/nosed end scrapers (?) Changes in projectile point tip form from thick, parallel-sided items with long fore-sections and wide bases, to thin items short fore-sections whose sides expand from a narrower base towards the tip; appearance of "pentagonal" resharpening B.P., which most likely encompasses the period of Early Paleoindian occupation of the Northeast (see Levine 1990), was one of marked temporal variability in the southern Ontario landscape. At 11,000 B.P. evidence indicates almost all of southern Ontario was covered by an open spruce parkland or woodland environment (Julig and McAndrews 1994; McAndrews 1994). An open landscape is suggested by the low pollen influx values of the time (e.g., Mott and Farley-Gill 1978) and the presence of significant amounts of pollen from open ground plant forms such as sedges, shrubs and herbs (e.g., Gaudreau 1988; McAndrews 1981, 1994). It is also suggested by the recovery of fauna we would expect to be more indicative of an open landscape such as mammoth (McAndrews and Jackson 1988; Pilny et al. 1987), and of course, the Arctic fox from the Gainey phase Udora site (Storck and Spiess 1994: 132). Over time, this open landscape became a more closed boreal forest dominated by pine. The timing of such closure is not clear although it seems to coincide with the shift from spruce to pine pollen dominance which is dated to an average of around 10,500 B.P. at 11 sites in southern Ontario and adjacent areas (Karrow et al. 1975). The appearance of the period of pine dominance seems to coincide with the marked reduction in the pollen of open ground species, such as herbs and sedges, implying increasing tree closure (e.g., McAndrews 1981). A major implication of increasing tree cover is a greater amount of dispersion of game animals such as caribou as well as shifts in caribou ranges (Deller and Ellis 1988:261; Shott 1989a:23-24, 1989b:295). Moreover, pine forests have long had a reputation for being environments of a low biotic productivity (e.g., Fitting 1968). Because ofthe dominance of boreal species in the pollen record of southern Ontario, it is probable that resources such as game and fish were of most importance to the Paleoindian diet of that area (Fitting et al. 1966). Also, as the first inhabitants of the area, one would expect these peoples to have had abnormally low population densities versus environmental productivity. These low densities are strongly suggested by evidence for the exploitation of large annual ranges and the extreme rarity of Early Paleoindian sites in comparison to those of even the immediately following Early Archaic (Ellis 1993:599-600; Ellis et al. 1990:68; Seeman 1994:275). Such low population densities clearly favour more of a subsistence focus on larger game among ethnographically known foragers (Keeley 1988). Therefore, the changes towards a more closed environment, and gre~ter dispersion of major potential prey species, such as caribou, undoubtedly had a profound effect on Early Paleoindian cultural systems and particularly, on subsistence practices. We have suggested shifts over time in hunting strategies. Specifically, we have posited a shift from more of a reliance on intercept, communal hunting of larger game on earlier sites inhabited at the time of more open environments, to an-increasing reliance on more individualistic taking oflarge game animals in the later, more closed environments (Deller and Ellis 1988:261). The more individualistic taking of game could include, and probably did, both stalking of individual animals using weapons as well as the use of traps or various kinds of untended facilities. However, we stress hunting with weapons in some of the following since we only have archaeological evidence for the weapons and not the facilities. It is proposed that a shift from more communal to individualistic taking of game can explain many of the changes evident over time in the Paleoindian archaeological record of southern Ontario. For example, with the assumption that larger sites represent population aggregations to intercept game such as caribou, one can use the disappearance of these large sites over time as evidence of a decrease in communal hunting practices. We readily admit, however, that the interpretation of some large sites as aggregations for the primary purpose of communal hunting is questionable at some sites such as Fisher (Storck, in press) although it seems a perfectly viable explanation for sites such as Parkhill (Deller and Ellis 1992b:27). Similarly, one can argue that the suggestion of a restriction of larger sites over time to more northerly areas of southern Ontario, and specifically the Algonquin strandline, could be consistent with this explanation. One might expect the more closed environments to spread into the area over time from roughly southwest to northeast. It has even been suggested that the restriction of known large Parkhill phase sites to the area of the real or inferred Algonquin strandline, and the real rarity of any "interior" Parkhill phase sites, is due. to the fact that environments remained more open in those strandline areas over a more extended period of time -- an inference supported by some paleoenvironmental data (e.g., Morgan 1988:204-205; Morgan et al. 1983:252). For that matter, a greater frequency of any sized sites in the area of that strandline over time, and a corresponding decrease in frequency of interior sites, might suggest the continuing attractiveness of the shoreline areas to occupation over time. This trend could be due to greater concentrations of game in more open areas as well as, of course, the development over time of more sterile pine-dominated forests in the interior. In fact, it is interesting to note that all of the Ontario sites with evidence of continuous occupation through more than one of the hiter Early Paleoindian phases (e.g., Parkhill and Crowfield) and extending, in some cases, even into Late Paleo indian times, are in more northern areas and specifically, on the Algonquin shoreline in southcentral Ontario. These would include the Udora, Hussey and Zander sites. We suggested earlier that one would expect more occupation of the same location over time to be in areas of relative environmental stability. Therefore, the fact that all of these sites are assoCiated with the inferred Algonquin strand line supports the idea that the area remained relatively stable environmentally over time, or at least remained more attractive to occupation than interior areas. These sites are in the most northerly section of Ontario known to have been occupied by Early Paleoindian and one might expect this general area to have maintained the open vegetation cover for the longest period of time during the Early Paleoindian era. It is also worth stressing that larger Parkhill phase sites (e.g., Parkhill and Fisher), as evidenced by lower frequencies of artifacts and the fact they have evidence of multiple occupations, seem to represent shorter duration per year, and much less"intensive occupations, than earlier, large, Gainey phase sites (albeit sites outside Ontario) such as Nobles Pond, Ohio. A shorter span per occupation might also account somewhat for the presence of more specialized activity areas on later-dating large sites such as Parkhill and Fisher; that is, they were occupied for relatively short periods of the year, and for a limited range of activities, in comparison to earlier sites. Overall, suggestions of a decline in occupation span per year at potential aggregation sites might again indicate that the conditions promoting aggregation and its extended duration (such as communal hunting, or even, a diversity of resources in the same location) were declining in importance. It is even possible to explain' changes in projectile morphology over time as due to changes in hunting strategies associated with changing environments (Shott 1989a:22-35). As mentioned earIier, the fact one can use the various point types as stylistic markers does not necessarily mean the causes of these changes were purely stylistic or random. Instead, it is also possible such changes would be "functional" or evolutionary in terms of Dunnell's (1978) usage of these concepts. That is, the changes could have resulted from selective forces acting on variability in tool design to increase the frequency of certain attributes or lead to shifts in variable ranges - changes which would produce more superior weapons in a given and changing use context. Unlike virtually every other stone tool, points are directly used to procure resources from the environment and we might expect a priori that hunting strategies, and changing contexts of weapon use, certainly could impact on tool design and should be considered as a causative factor. Nonetheless, we do not pretend most projectile tips, or aspects thereof, change due to such "functional" factors. For example, where hunting is not that important to the economy, less attention may be paid to the relative superiority of certain designs as killing weapons in those contexts. In sum, the effort required to achieve superior weapons may be outweighed by other factors such as increased costs of procuring raw materials, shortened use-lives, higher maintenance costs, higher transportation costs, and so on. A good ethnographic example is the Northwest Coast. Historic data clearly demonstrate that stone projectile tips are more lethal than those made on organic materials. However, stone tips were not used over much of the Northwest Coast, probably due at least partially to the limited role land mammal hunting played in the predominantly maritime economy (Ellis 1994b). In direct contrast to the historic Northwest Coast situation, and as discussed earlier, one can make very compelling arguments as to the greater importance of large animal hunting to Early Paleoindian economies in the Great Lakes area because of the low population densities and environmental conditions inferred for the time period. Therefore, we might expect weapon design to be more closely related to subsistence changes among Great Lakes Paleoindian than it was in other areas of the time or in later time periods. In fact, achieving more optimum projectile designs may not have been as important among Paleoindian groups in the Southeast where, based on the number of reported findspots alone (see Mason 1962), there may have been greater population densities, and one presumes, less emphasis on large animal hunting. Also, in that same area, as well as in the Great Lakes in post-Early Paleoindian times, groups would have had more access to other resources not as plentiful in the Great Lakes, such as floral ones. Therefore, hunting was probably less important to the economy at those other times and places, and again less emphasis might have been placed on producing superior weapons. For the above reasons, we must consider the possibility that changes in Great Lakes fluted point form is evolutionary. It is extremely difficult to distinguish between that kind of change and random, stylistic changes (Maxwell 1995; O'Brien and Holland 1992, 1995: 187). However, as we will discuss below, experimental and ethnographic evidence has clearly shown that some design aspects of stone projectile tips do result in more efficient weapons under certain conditions, than do other alternative designs. Demonstrating this greater efficiency strongly implicates adaptation through selection rather than simple stylistic change as the causative mechanism (Maxwell 1995: 116). In addition, we believe a functional explanation has the potential to explain a wider range of variability in fluted bifaces within the Northeast than does a simple random change explanation. For example, as we will discuss later, one can use the arguments developed here to explain not only change in point form over time in the Great Lakes but also to explain differences in morphology in space and, particularly, between the Great Lakes area and areas such as northern New England. We suggested above that there was a shift over time from communal hunting toward more individual stalking of game animals. In intercept communal hunting the game is taken at close range, often in situations which impede the game animal's mobility such as at river crossings. In such situations, the weapons must be more optimally designed for use as thrusting weapons which can be used to repeatedly stab an animal(s), or, if launched weapons, ones which need be propelled only over short distances. However, in the more individualistic or small group hunting of game by stalking, other factors are clearly of major importance in determining weapon design. These factors would include: killing prey at a greater distance, killing more successfully through the use of more accurate weapons, and being able to easily trail wounded animals because of more extensive bleeding. Therefore, we might expect evidence that the earlier projectiles in southern Ontario were not as efficient as later projectile tips in these specific regards. The earliest inferreq points (Gainey) are essentially parallel-sided or even contracting from the base, have wide bases, and are relatively thick and long. At the opposite end of the time series (Crowfield) the points are markedly expanding from the base, have narrow bases, exhibit short fore-sections, and are very thin. Thick, robust points with a long fore-section like Gainey (or even Barnes points) may be well-suited to create deep and multiple wounds from several hand-held thrusts. In fact, the binding projecting from the parallel-sides on at least Gainey points would even inhibit complete penetration of the weapon tip (see below). Alternatively, the long point could penetrate more deeply if launched from close range with a heavy shaft to force deep penetration despite the projecting binding. Simply because of their weight, these heavier shafts only can be used over short distances (Ellis 1994b). Certainly, the wide bases of Gainey points in comparison to later forms indicate a heavier shaft more suitable either for launched weapons used only at short ranges, or use of a stouter shaft less prone to breakage which would be more suitable for repeated hand-held thrusts (see also Gramly 1984: 113). The deep basal concavities of the Gainey points with thick ears also seem more suited to thrusting use. If lodged in a split shaft, the long ears projecting down the sides of the shaft would be more useful in preventing the tip from being forced or turned sideways during penetration. Such turning in the haft would prevent use for subsequent thrusts. While Gainey points seem well-designed for thrusting spears or spears launched over short distances, they seem much less useful as throwing weapons intended to be propelled at prey from a long distance. As implied above, it is notable that experimental evidence has consistently demonstrated that parallel-sided points, or points with essentially fore-sections as narrow or narrower than their bases, are relatively inefficient weapons in killing from a distance (e.g., Frison 1991:293; Guthrie 1983:289; Huckell 1982; Pope 1974:56). Based on ethnographic data, it is at longer distances where one would expect lighter spears shafts to be used simply because they can be thrown further (Ellis 1994b; Gramly 1984:113). On parallel-sided lance-shaped points the binding projects from the sides ofthe point and prohibits penetration. This attribute can present a real problem with throwing weapons intended to kill at a distance because the combination of a required lighter shaft, and diminishing thrust as the projectile travels towards the game, would minimize penetration potential. One means of overcoming this problem is simply to increase the width of the fore-section versus the base so that the former cuts a hole wide enough to let the shaft and binding enter the prey more easily (see Frison 1991:293; Pope 1974:56). On this basis, Barnes and Crowfield points, with their expanding lateral edges from the base, are clearly superior weapons. One can also argue that the very broad Crowfield points in contrast to earlier forms would create an even larger wound which would lead to more excessive bleeding. More bleeding would not only increase the chance of successfully killing the prey at a distance but also, and this is especially important in individual stalking of game, in actually trailing and locating the wounded animals. Of course, the narrower bases of later point forms such as Barnes and Crowfield also suggest a less stout shaft/foreshaft, or a lighter weapon designed to be propelled longer distances. Still another means of increasing the killing power of points tipping launched projectiles is to increase the sharpness of their edges (Frison 1991:293) and in this regard, thinner points would be favoured. Therefore, the later dating fluted points are inferred to be more lethal weapons in this situation. As well, ethnographic and experimental evidence shows that one of the reasons (but not the only reason) for the greater killing power of stone points in comparison to those on other materials is the fact they break up in the wound (Ellis 1994b; Flenniken 1985:270). In this regard, thinner points which are more amenable to extensive fragmentation in the wound, such as Crowfield points, would be more lethal weapons. At the same time, of course, their thin delicate nature means they would be less effective as tips used in multiple thrusts (e.g., they would break too readily and thus, would be less reliable weapons in the sense of Bleed [1986]). Also, the lack of long tips beyond the bound area would mean they could not be thrust too deeply and withdrawn easily as is required for multiple thrusts. In other words, with a deep thrust the complete point could penetrate the game and the point's ears would prohibit the spear being withdrawn from the wound. It is even possible to explain the pentagonal resharpening of Crowfield points as being related to a projectile designed to be more lethal in attacking from a distance. Experiments have shown, perhaps not surprisingly, that a relatively narrow, sharp tip allows for better penetration than a broad, blunt one (Raup 1976:283). Crowfield points markedly expand from the base and are quite broad. If one was to create a narrow, .1\, / I I I , \ \ I I I , \ \ \ \ \ I I \ I I , I / / I I '" \ \ \ \ I I " \ \ / \ \ \ \ \ \ \ \ : I ./ I ) I \ I Figure 6. Potential tip resharpening strategies for Crowfield points. A-C are hypothetical and are not known from archaeological examples. D is the actual strategy seen on Crowfield points and results in a pentagonal outline. sharp tip during resharpening of damaged ends (as in the hypothetical example on Figure 6a), in line with previous discussions this strategy would result in a narrowing of the overall point such that the hole made by the projectile tip would be smaller. In turn, this would inhibit penetration of the shaft and prevent deeper wounds. Alternatively, this resharpening strategy might result in abrupt,changes in outline (as on the hypothetical Figure 6b) which would reduce penetration. Also, given the thinness of these points, it would produce a very narrow, weak tip prone to collapse on initial contact with the prey prior to any penetration. All of these factors would diminish the usefulness ofthe point as a launched weapon tip. At the opposite extreme, if one were to simply maximize point width in comparison to the base (as on the hypothetical example on Figure 6c), the broad blunt tip would be a drawback and particularly on a weapon launched over a long distance. The most ideal solution to maintaining both maximum width and a sharp tip is a compromise: one resharpens diagonally to the point's long axis each fore-section edge to produce a relatively straight margin adjacent to the tip. In this manner, one not only gets a sharp, relatively narrow and yet still strong tip, but also, maximizes fore-section width in comparison to basal width (Figure 6d). This compromise is exactly what one sees on actual resharpened Crowfield points and, as noted earlier, contributes to their often pentagonal outline shape. Finally, Shott (1989a:26-32) has presented arguments suggesting that points with narrower bases have improved range and accuracy and in this regard, points such as Barnes and Crowfield would be superior to those of the Gainey phase and favoured in individualistic stalking of game. Therefore, one can make numerous and strong arguments to the effect that later dating projectiles are more lethal and accurate in hunting at greater distances from the prey than earlier dating point tips. Conversely, the earlier tips seem more suited to short range or thrusting use than the later forms. In turn, these changes can be accounted for if one accepts shifts over time from more intercept communal hunting methods using thrusting or "short range" launched weapons employing heavy shafts, to more individualistic stalking of game using lighter weapons launched at the prey from some distance. As near as can be determined with existing evidence, the archaeological record for Early Paleoindians in southern Ontario is duplicated in immediately adjacent areas such as southern Michigan, northern Ohio and western New York state. In those areas, there is evidence based on reported sites and scattered surface finds that material resembling the three suggested phases is also present (e.g., comparable point forms). Moreover, the remainder of the tool kits seem to be quite similar in the sense that tool forms rare or absent in southern Ontario are also rare in those areas such as pieces esquillees, fluted drills, and flake shavers. Patterns oflithic raw material exploitation are also similar. For example, in Michigan, and exactly as is the case in Ontario, only Gainey phase peoples utilized Upper Mercer cherts; whereas later peoples of the Parkhill phase used more northern cherts such as Bayport, Michigan (e.g., Simons et al. 1984; Voss 1977; Wright and Roosa 1966). In addition, Gainey phase sites in Michigan such as the type site (Simons et al. 1984) can yield assemblages predominantly made on cherts from sources over 200 km away, whereas later Parkhill phase sites are always within 200 km from the outcrops (e.g., Voss 1977). It also appears based on reported findspots (e.g., Lepper 1986; Mason 1958; Prufer and Baby 1963; Ritchie 1957; Roosa 1965) that Gainey or Gainey-like points are much more common than Crowfield or Barnes type points in Michigan, northern Ohio, and western New York state. While there may be some major differences which are not yet apparent, the suggestions of widespread similarities over these areas argues that the cultural systems and adaptations of those groups were much like those of southern Ontario, and, in fact, the consistent presence of Ohio and Michigan cherts in southern Ontario assemblages suggests the peoples of these areas were in relatively direct contact. Expanding comparisons to areas farther east, it is clear that the eastern Great Lakes record is broadly similar to that of these adjacent areas. For example, fluted points similar to Gainey, Barnes and Crowfield do occur in southern and western New England (e.g., Grimes 1979; Grimes et al. 1984: Plate 6-1; Kraft 1973; Moeller 1980: Plate 7; Ritchie 1953; Spiess et al. 1995; Witthoft 1952). Moreover, Gainey-like points are much more common in those areas than presumed later forms; an~, as well, known large sites such as Shoop and Bull Brook solely yield Gainey-like forms. Large sites/assemblages yielding Barnes-like or Crowfield-like points are nonexistent. These Gainey-like sites, which Grimes et al. (1984) attribute largely to a Bull Brook phase, are also notable in that the main sources of stone materials used can be from large distances away (>200 km; see Spiess et al. 1995; Witthoft 1952). In this respect they also mirror Gainey phase assemblages in the Great Lakes. These similarities suggest broadly similar changes over time (reduction in range mobility, tendency for large possible aggregation sites to disappear over time, etc.). The southern/western New England assemblages are not identical to the Great Lakes assemblages. The main differences seem to be in artifact assemblages. In particular, pieces esquillees, fluted/twist drills, and limaceslflake shavers are very common in reported assemblages from the New England sites. As as noted earlier, these types are rare to nonexistent in the eastern Great Lakes. Similarly, certain Great Lakes tool forms such as large unifacially beveled bifaces and narrow end scrapers are virtually if not totally absent from the New England assemblages. In the absence of a good knowledge of the function of all of these tool types (and of course, a knowledge of whether or not pieces esquillees are tools or cores), the reasons for these contrasts remain obscure. Moving even farther afield to the "Far Northeast" (e.g., northernmost Maine and the Canadian Maritime Provinces), even more contrasts become apparent with Great Lakes assemblages. Besides the fact these far northeastern assemblages commonly include the tool forms seen in southern New England but rare in the Great Lakes area (pieces esquillees, fluted/twist drills, etc.) and exclude Great Lakes types such as large unifacially beveled bifaces and narrow end scrapers, a major difference with the Great Lakes area is the . absence of points even remotely resembling both Barnes and Crowfield points. In the Far Northeast, assem~ blages and findspots are either Gainey-like (e.g., Ellis and Payne 1995) or they are the similar large points with exceptionally deep basal concavities reported from sites such as Vail (Gramly 1982) or Debert (MacDon- Archaeology of Eastern North America aid 1968). Some actually consider the latter fluted point forms to be equivalent to Gainey (e.g., Gramly 1988b; Gramly and Funk 1990) but in the absence of detailed statistical comparisons to show their essential identity, and given suggestions they are not time-equivalent (see Deller and Ellis 1992a: 128), we continue to find the distinction valid. Regardless, there is a decided lack of pronounced variability in the fluted points of the Far Northeast which markedly contrasts with the diversity of forms seen in southern Ontario. It has been suggested that this lack of variability indicates a short time span for the Early Paleoindian occupation of the Far Northeast compared to the Great Lakes area (Ellis 1993:607). In essence, and on the assumption that the Far Northeast is the most removed from the area where fluted point peoples originated (e.g., western North America) and was the last to be colonized, it can be argued that it was only occupied for a short period of time before fluted points went out of use. As a result, there was little time for points to change in form. However, there are alternative explanations. It is quite clear that environments remained relatively open longer in the Far Northeast. For example, all evidence still seems to indicate that the Debert site was in an open parkland or perhaps even a tundra environment when it was occupied around 10,600 B.P. (Ammann et al. n.d.; Davis and Jacobsen 1985; Richard 1985). Moreover, locations in areas such as southeast New Brunswick suggest that an actual forest may have been in existence as early as ca. 11,000 B.P. but subsequently, and throughout most of the lIth millenium B.P., there was a return to quite open landscapes (Levesque et al. 1994:325; Mott 1975:284). Earlier, we argued that the large parallel-sided points of the earlier Gainey phase in the Great Lakes may have changed to forms which expand from the base as vegetation covers became more closed and there were pressures to design more effective weapons which were more useful in the individualized stalking of game. If so, one could suggest that the persistence of open environments in the Far Northeast meant there were few selective pressures to change point design such that groups could continue to make similar points, more suited for communal hunting, for an extended period. Such an explanation would also explain why the Debert and Vail sites seem to date relatively late in the expected time range of Early Paleoindians in the Northeast (e.g., closer to 10,000 B.P.) in comparison to the earlier age estimates for comparable large, parallelsided points elsewhere in the Northeast (see Deller and Ellis 1992a: 128). We have summarized evidence which clearly indicates at least three different Early Paleoindian manifestations, referred to here as "phases," occur in southern Ontario and more broadly, the lower Great Lakes region. Multiple lines of independent evidence also point to the same conclusion: these three manifestations represent a time series - there seems to be no viable alternative explanation. Patterned differences are also recognizable between these three manifestations which, except for some limited instances where the differences could also be explained by simple sampling error, indicate changes through time in the lifestyles of these Early Paleoindian groups. We have suggested that these changes can be explained by a combination of: 1) colonization of previously uninhabited areas, and 2) alterations in adaptive patterns in response to changing environmental conditions and specifically, a lessened role for communal hunting in Paleoindian subsistence. We have also suggested that the differences seen between the eastern Great Lakes and areas farther east can also be explained by invoking variation in subsistence strategies between the two regions. Obviously, we are just beginning to recognize patterned variation in Great Lakes/Northeastern Early Paleoindian assemblages, let alone to explain such variation. However, it is now evident that patterned variation does exist, that it needs to be carefully documented, and that it has a great, largely untapped, potential to enhance our understanding of Early Paleoindian cultural systems and environmental coping strategies. Our research on the Paleo indian occupations of southern Ontario has been supported by grants from the Social Sciences and Humanities Research Council of Canada (most recently by Research Grant #410-90-1642 which allowed us to examine some small sites mentioned in this paper) and from the Ontario Heritage Foundation. We gratefully acknowledge that support. We also wish to thank Andrew L. Christensen and Stephen Harding who helped us with references and figures. Lawrence Jackson, David Meltzer, the three paper reviewers, and Arthur Spiess, editor of AENA, provided helpful comments on earlier drafts. These comments forced us to rethink some precepts and rewrite certain sections of the paper. We believe they have strengthened the arguments in the paper but we stress we have not always agreed with the reviewers comments and we remain responsible for any faults which remain. Peter L. Storck invited us to participate in the symposium at the Minneapolis, Minnesota Society for American Archaeology Meetings for which this paper was originally written and allowed us to cite the in press manuscript on the Fisher site. The artifact drawings were prepared by Janie Ravenhurst and Ian Kenyon. Since much of our work in fluted point typology is built upon the pioneering ideas and efforts of the late John Witthoft and William Roosa,we respectfully dedicate this paper to their memory. Ammann, B. R., D. A. Livingstone and R. Stuckenrath n.d. Absolute Pollen Frequency Diagram from the Vicinity of the Debert Paleo indian Site, Nova Scotia. Manuscript on file, Archaeological Survey of Canada, Canadian Museum of Civilization, Ottawa, Ontario. Becker, B., B. Kromer, and P. Trimborn 1993 The Continental Tree-ring Record - Absolute Chronology, 14CCalibration and Climatic Change at Ilka. Palaeogeography, Palaeoclimatology, Palaeoecology 103:67-71. Bleed, Peter 1986 The Optimal Design of Hunting Weapons: Maintainability or Reliability. American Antiquity 45:4-20. Bonnichsen, Robson 1977 Models for DerivingCultural Information From Stone Tools. National Museum of Man, Archaeological Survey of Canada, Mercury Series Paper No. 60. Clarke, D. L. 1968 Analytical Archaeology. Methuen and Co., London, England. Custer, Jay and R. Michael Stewart 1989 Environment, Analogy and Early Paleo-Indian Economies in Northeastern North America. In Early Paleoindian Economics of Eastern North America, edited by K. Tankersley and B. Isaac, pp. 303-322. Research in Economic Anthropology, Supplement 5. Davis, Ron and George L. Jacobsen 1985 Late Glacial and Early Holocene Landscapes in Northern New England and Adjacent Areas of Canada. Quaternary Research 23:341-368. Deller, D. Brian 1979 Paleo-Indian Reconnaissance in the Counties of Lambton and Middlesex. Ontario Archaeology 32:3-20. 1988 The Paleo-Indian Occupation of Southwestern Ontario: Distribution, Technology and Social Organization. Unpublished PhD Dissertation, Dept. of Anthropology, McGill University, Montreal, Quebec. 1989 Interpretation of Chert Type Variation in Paleoindian Industries, Southwestern Ontario. In Eastern Paleoindian Lithic Resource Use, edited by C. Ellis and 1. Lothrop, pp. 191-220. Westview Press, Boulder, Colorado. Deller, D. Brian and Christopher 1. Ellis 1984 Crowfield: A Preliminary Report on a Probable Paleo-Indian Cremation in Southwestern Ontario. Archaeology of Eastern North America 12:41-71. 1988 Early Palaeo-Indian Complexes in Southwestern Ontario. In Late Pleistocene and Early Holocene Paleoecology and Archeology of the Eastern Great Lakes Region, edited by R. Laub, N. Miller and D. Steadman. Bulletin of the Buffalo Society of Natural Sciences 33:251-263. 1992a Thedford II: A Paleo-Indian Site in the Ausable River Watershed of Southwestern Ontario. Memoirs of the Museum of Anthropology, University of Michigan 24. 1992b The Early Paleo-Indian Parkhill Phase in Southwestern Ontario. Man in the Northeast 44(2): 15-54. 1996 The Bolton (AfHj-89) Site: An Early Paleo-Indian Crowfield Phase Site in Southwestern Ontario. Ontario Archaeology 61, in press. Dunnell, Robert C. 1978 Style and Function: A Fundamental Dichotomy. American Antiquity 43: 192-202. Ellis, Christopher 1. 1984 Paleo-Indian Lithic Technological Structure and Organization in the Lower Great Lakes Area: A First Approximation. Unpublished PhD Dissertation, Dept. of Archaeology, Simon Fraser University, Burnaby, B.c. 1987a Gainey Fluted Points. Kewa 87(5):4. 1987b Barnes Fluted Points. Kewa 87(5):5. 1989 The Explanation of Northeastern Paleo indian Lithic Procurement Patterns. In Eastern Paleoindian Lithic Resource Use, edited by C. Ellis and 1. Lothrop, pp. 139-164. Westview Press, Boulder, Colorado. 1993 Premieres Industries Lithiques PaIeo-Indiennes de la Region N-E de L'Amerique du Nord Dans Leurs Principaux Contextes (Temps, Espace, Paleo-environnement, Sites, Culture). L'Anthropologie 97(4):586-622. 1994a Some Unanswered Questions Concerning Early Paleo-Indian Subsistence and Settlement Systems in Southern Ontario. In Great Lakes Archaeology and Paleoecology: Exploring Interdisciplinary Initiatives for the Nineties, edited by R.I. MacDonald, pp. 413-429. Quaternary Sciences Institute, University of Waterloo, Publication No. 10. 1994b Factors Influencing the Use of Stone Projectile Tips: An Ethnographic Perspective. Paper presented in symposium: Projectile Technologies: Archaeological, Experimental, and Ethnoarchaeological Perspectives, organized by Dr. Heidi Knecht, University of Miami. 59th Annual Meeting, Society for American Archaeology, Anaheim, California. Manuscript on file, Department of Anthropology, University of Western Ontario, London, Ontario, Canada. 1994c Miniature Early Paleo-Indian Stone Artifacts from the Parkhill, Ontario, Site. North American Archaeologist 15(3):253-267. Ellis, Christopher 1. and D. Brian Deller 1986 Post-Glacial Lake Nipissing "Waterworn" Assemblages from the Southeastern Huron Basin Area. Ontario Archaeology 45:39-60. 1988 Some Distinctive Paleo-Indian Tool Types from the Lower Great Lakes Region. Midcontinental Journal of Archaeology 13:111-158. 1990 Paleo-Indians. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. Ellis and N. Ferris, pp. 37-63. Occasional Publications, London Chapter, Ontario Archaeological Society, No.5. Ellis, Christopher J., D. Brian Deller and Lawrence 1. Jackson 1992 Investigations at Small, Interior, Early Paleo-Indian Sites in Southwestern Ontario. Annual Archaeological Report for Ontario for 1991, New Series 2:92-96. Ellis, Christopher J., I. T. Kenyon and M. Spence 1990 The Archaic. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. Ellis and N. Ferris, pp. 65-124. Occasional Publications, London Chapter, Ontario Archaeological Society, No.5. Ellis, Christopher 1. and James H. Payne 1995 Estimating Failure Rates in Fluting Based on Archaeological Data: Examples from NE North America. Journal of Field Archaeology 22:459-474. Fedje, D ..W., 1. White, M. Wilson, D.E. Nelson, 1.S. Vogel, and J.R. Southon 1995 Vermillion Lakes Site: Adaptations and Environments in the Canadian Rockies During the Late Pleistocene and Early Holocene. American Antiquity 60:81-108. Fitting, James E. 1968 Environmental Potential and Post-Glacial Readaptation in Eastern North America. American Antiquity 33:441445. Fitting, James E., Jerry DeVisscher and E. Wahla 1966 The Paleo-Indian Occupation of the Holcombe Beach. Anthropological Papers, Museum of Anthropology-, University of Michigan, No. 27. Flenniken, U. 1985 Stone Tool Reduction Techniques as Cultural Markers. In Stone Tool Analysis. Essays in Honor of Don Crabtree, edited by M. Plew, 1. Woods and M. Pavesic, pp. 265-275. University of New Mexico Press, Albuquerque. Frison, George C. 1991 Prehistoric Hunters of the High Plains, Second Edition. Academic Press, San Diego. Frison, George C. and B. Bradley 1980 Folsom Tools and Technology and the Hanson Site, Wyoming. University of New Mexico Press, Albuquerque. Funk, Robert E. and David Steadman 1994 Archaeological and Paleoenvironmental Investigations in the Dutchess Quarry Caves, Orange County, New York. Persimmon Press Monographs in Archaeology, Buffalo, N.Y. Funk, Robert E. and Beth Wellman 1984 The Corditaipe Site: A Small Isolated Paleo-Indian Camp in the Upper Mohawk Valley. Archaeology of Eastern North America 12:72-80. Gardner, William 1976 Paleo-Indian to Early Archaic: Continuity and Change in Eastern North America During the Late Pleistocene to Early Holocene. In Habitats Humains Anterieurs a L 'Holocene en Amerique, edited by 1.8. Griffm, pp. 19-46. IXth International Congress of Prehistoric and Protohistoric Sciences, Nice, France. Gardner, William and R. Verry 1979 Typology and Chronology of Fluted Points from the Flint Run Area. Pennsylvania Archaeologist 49{1): 13-46. Gaudreau, Denise C. 1988 The Distribution of Late Quaternary Forest Regions in the Northeast. In Late Pleistocene and Early Holocene Paleoecology and Archeology of the Eastern Great Lakes Region, edited by R. Laub, N. Miller and D. Steadman. Bulletin of the Buffalo Society of Natural Sciences 33:215-256. Gramly, R.Michael 1982 The Vail Site: A Palaeo-Indian Encampment in Maine. Bulletin of the Buffalo Society of Natural Sciences 30. 1984 Kill Sites, Killing Ground and Fluted Points at the Vail Site. Archaeology of Eastern North America 12:110-121. 1988a The Adkins Site: A Palaeo-Indian Habitation and Associated Stone Structure. Persimmon Press Monographs in Archaeology, Buffalo, N.Y. 1988b Paleo-Indian Sites South of Lake Ontario, Western and Central New York State. In Late Pleistocene and Early Holocene Paleoecology and Archeology of the Eastern Great Lakes Region, edited by R. Laub, N. Miller and D. Steadman. Bulletin of the Buffalo Society of Natural Sciences 33:265-280. Gramly, R. Michael and Robert Funk 1990 What is Known and Not Known About the Human Occupation of the Northeastern United States Until 10,000 B.P. Archaeology of Eastern North America 18:5-31.' Gramly, R. Michael and Gary Summers 1986 Nobles Pond: A Fluted Point Site in Northeastern Ohio. Midcontinental Journal of Archaeology 11:97-123. Grimes, John R. 1979 A New Look at Bull Brook. Anthropology 3: 109-130. Grimes, John R., William' Eldridge, Beth Grimes, A. Vaccaro, F. Vaccaro, 1. Vaccaro, N. Vaccaro and A. Orsini 1984 Bull Brook II. Archaeology of Eastern North America 12: 159-183. Grimes, John R. and Beth Grimes 1985 Flakeshavers: Morphometric, Functional and Life-Cycle Analyses of a Paleo indian Unifacial Tool Class. Archaeology of Eastern North America 13:35-57. Guthrie, R. D. 1983 Osseous Projectile Points: Biological Considerations Affecting Raw Material Selection and Design Among Paleolithic and Paleoindian Peoples. In Animals and Archaeology I: Hunters and Their Prey, edited by 1. CluttonBrock and C. Grigson, pp. 273-294. British Archaeological Reports International Series 163. Archaeology of Eastern North America Huckell, B. B. 1982 The Denver Elephant Project: A Report on Experimentation with Thrusting Spears. Plains Anthropologist 27:217-224. Irwin, H.T. and H.M. Wormington 1971 Paleo-Indian Tool Types in the Great Plains. American Antiquity 35:24-34. Jackson, L. 1. 1990 Interior Paleo indian Settlement Strategies: A First Approximation for the Great Lakes Region. In Early Paleo indian Economies of Eastern North America, edited by K. Tankersley and B. Isaac, pp. 95-142. Research in Economic Anthropology, Supplement 5. 1994 Gainey Phase Occupation in the Southern Rice Lake Region, Canada. PhD. Dissertation, Department of Anthropology, Southern Methodist University, Dallas, Texas. University Microfilms, Ann Arbor, Michigan. Johnson, E. and V. Holliday 1987 Lubbock Lake Artifact Assemblages. In Lubbock Lake: Late Quaternary Studies on the Southern High Plains, 'edited by E. Johnson, pp. 100-119. Texas A&M University Press, College Station, Texas. Judge, W. 1. 1973 Paleo-Indian Occupation of the Central Rio Grande Valley in New Mexico. University of New Mexico Press, Albuquerque. Julig, Patrick 1. and 1. H. McAndrews 1993 Les Cultures PaIeoindiennes dans la Region Des Grand Lacs en Amerique du Nord: Contextes Paleoclimatique, Geomorphologiques et Stratigraphiques. L'Anthropologie 97:623-650. Karrow, Paul F. and P. E. Calkin, editors 1985 Quaternary Evolution of the Great Lakes. Geological Association of Canada, Special Paper No. 30. Karrow, Paul F., Thane W. Anderson, A. Clarke, L.D. Delorme and M. Sreenivasa 1975 Stratigraphy, Paleontology and Age of Lake Algonquin Deposits in Southwestern Ontario, Canada. Quaternary Research 5: 49-87. Keeley, Lawrence E. 1988 Hunter-Gatherer Economic Complexity and "Population Pressure": A Cross-Cultural Analysis. Journal of Anthropological Archaeology 7:373-411. Kraft, Herbert C. 1973 The Plenge Site: A Paleo-Indian Occupation Site in New Jersey. Archaeology of Eastern North America 1:56117. Lepper, Bradley T. 1986 Early Paleo-Indian Land Use Patterns in the Central Muskingum Basin, Coshocton County, Ohio. PhD Dissertation, Dept. of Anthropology, Ohio State University, Columbus, Ohio. University Microfilms, Ann Arbor, Michigan. Levesque, A. 1., L. C. Cwynar, and I. R. Walker A Multiproxy Investigation of Late Glacial Climate and Vegetation Change at Pine Ridge Pond, Southwest New Brunswick, Canada. Quaternary Research 42:316-327. Levine, Mary A. 1990 Accomodating Age: Radiocarbon Results and Fluted Point Sites in Northeastern North America. Archaeology of Eastern North America 18:33-64. Lothrop, Jonathan The Organization of Paleoindian Lithic Technology at the Potts Site. PhD Dissertation, Department of Anthropology, SUNY-Binghamton, Binghamton, New York. University Microfilms, Ann Arbor, Michigan. Lotter, A. F. 1991 Absolute Dating of the Late-Glacial Period in Switzerland Using Annually Laminated Sediments. Quaternary Research 35:321-330. MacDonald, George F. 1968 Debert: A Palaee-Indian Site in Central Nova Scotia. National Museums of Canada, Anthropology Papers No. 16. 1971 A Review of Research on Paleo-Indian in Eastern North America, 1960-1970. Arctic Anthropology 8(2):32-41. 1988 1994 Mason, Ronald 1958 Late Pleistocene Geochronology and the Paleo-Indian Penetration into the Lower Michigan Peninsula. Anthropological Papers, Museum of Anthropology, University of Michigan No. 11. 1962 The Paleo-Indian Tradition in Eastern North America. Current Anthropology 3:227-246. Maxwell, T. D. 1995 The Use of Comparative and Engineering Analyses in the Study of Prehistoric Agriculture. In Evolutionary Archaeology - Methodological Issues, edited by P. A. Teltser, pp. 113-128. University of Arizona Press, Tucson. McAndrews, J. H. 1981 Late Quaternary Climate of Southern Ontario: Temperature Trends from the Fossil Pollen Record. In Quaternary Paleoclimate, edited by W. C. Mahaney, pp. 319-333. Geoabstracts, Norwich, England. 1994 Pollen Diagrams for Southern Ontario Applied to Archaeology. In Great Lakes Archaeology and Paleoecology: Exploring Interdisciplinary Initiatives for the Nineties, edited by R. I. MacDonald, pp. 179-195. Quaternary Sciences Institute, University ofWaterioo, Publication No. 10. McAndrews, J.H. and LJ. Jackson 1988 Age and Environment of Late Pleistocene Mastodont and Mammoth in Southern Ontario. In Late Pleistocene and Early Holocene Paleoecology and Archeology of the Eastern Great Lakes Region, edited by R. Laub, N. Miller and D. Steadman. Bulletin of the Buffalo Society of Natural Sciences 33: 161-172. Meltzer, DJ. 1984 Late Pleistocene Human Adaptations in Eastern North America. PhD Dissertation, Dept. of Anthropology, University of Washington, Seattle, Washington. University Microfilms, Ann Arbor, Michigan. 1985 On Stone Procurement and Settlement Mobility in Eastern Fluted Point Groups. North American Archaeologist 6:1-24. 1988 Late Pleistocene Human Adaptations in Eastern North America. Journal of World Prehistory 2:1-52. Moeller, Roger 1980 6LF2I: A Paleo-Indian Site in Western Connecticut. American Indian Archaeological Institute, Occasional Publication No.2. Washington, Connecticut. Montelius, O. 1903 Die Typologische Methode: Die Alteren Kulturperioden im Orient und in Europa, Volume I. Svelbstverlag, Stockholm, Sweden. Morgan, A.V. 1988 Late Pleistocene and Early Holocene Coleoptera in the Lower Great Lakes Region. In Late Pleistocene and Early Holocene Paleoecology and Archaeology of the Eastern Great Lakes Region, edited by R. Laub, N. Miller and D. Steadman. Bulletin of the Buffalo Society of Natural Sciences 33:195-206. Morgan, A.V., A. Morgan, A. Ashworth and J.V. Mathews 1983 Late Wisconsin Fossil Beetles in North America. In Late Quaternary Environments of the United States. Volume 1: The Late Pleistocene, edited by H.E. Wright Jr., pp. 354-363. University of Minnesota Press, Minneapolis. Mott, R. J. 1975 Palynological Studies of Lake Sediments from Southern New Brunswick. Canadian Journal of Earth Sciences 12:273-288. Mott, RJ. and L.D. Farley-Gill 1978 A Late-Quaternary Pollen Profile from Woodstock, Ontario. Canadian Journal of Earth Sciences 15:1101-1111. O'Brien, M. J. and T. D. Holland 1992 The Role of Adaptation in Archaeological Explanation. American Antiquity 57:36-59. 1995 The Nature and Premise of a Selection-Based Archaeology. In Evolutionary Ar.chaeology - Methodological Issues, edited by P. A. Teltser, pp. 175-200. University of Arizona Press, Tucson. Petrie, F. 1899 Sequences in Prehistoric Remains. Journal of the Anthropological Institute 29:295-301. Pilny, J., A. V. Morgan, and A. Morgan 1987 Paleoclimatic Implications of a Late Wisconsinan Insect Assemblage from Rostock, Southwestern Ontario. Canadian Journal of Earth Sciences 24:617-630. Archaeology of Eastern North America Pope, S.T. 1974 Bows and Arrows. University of California Press, Berkeley. Prufer, OlafH. and Robert S. Baby 1963 Paleo-Indians of Ohio. Columbus, Ohio Historical Society. Raup, J.W. 1976 Some Experiments with Replica Projectile Points Used as Arrow Points. APE (Experimental Archaeology Papers) 4:281-303. Richard, P. J. H. 1985 Couvert Vegetal et Paleoenvironnements Du Quebec Entre 12000 et 8000 Ans B.P., L'Habitabilite Dans Un Milieu Changeant. Recherches Amerindiennes au Quebec 15(1-2):39-56. Ritchie, W.A. 1953 A Probable Paleo-Indian Site in Vermont. American Antiquity 18:249-258. 1957 Traces of Early Man in the Northeast. New York State Museum and Science Service, Bulletin 358. Roberts, A.B. 1984 "Paleo Indian on the North Shore of Lake Ontario. Archaeology of Eastern North America 12:248-265. Roberts, Frank H. H. Jr. 1935 A Folsom Complex: Preliminary Report on Investigation at the Lindenmeier Site in Northern Colorado. Smithsonian Miscellaneous Collections 94(4). Roosa, William B. 1963 Some Michigan Fluted Point Types and Sites. Michigan Archaeologist 9(3):44-48. 1965 Some Great Lakes Fluted Point Types. Michigan Archaeologist 9(3): 89-102. 1977a Great Lakes Paleo-Indian: The Parkhill Site, Ontario. Annals of the New YorkAcademy of Sciences 288:349-354. 1977b Fluted Points from the Parkhill, Ontario Site. In For the Director: Research Essays in Honor of James B. Griffin, edited by C. Cleland, pp. 87-122. Anthropological Papers, Museum of Anthropology, University of Michigan No. 61. Seeman, Mark F. 1994 Intercluster Lithic Patterning at Nobles Pond: A Case for "Disembedded" Procurement Among Early PaleoIndian Societies. American Antiquity 59:273-288. Shott, Michael 1986 Settlement Mobility and Technological Organization among Great Lakes Paleo-Indian Foragers. PhO Dissertation, Dept. of Anthropology, University of Michigan. University Microfilms, Ann Arbor, Michigan. 1989a Stone Tools and Economics: Great Lakes Paleoindian Examples. In Early Paleoindian Economies of Eastern North America, edited by K. Tankersley and B. Isaac, pp. 3-43. Research in Economic Anthropology, Supplement 5. 1989b Diversity, Organization and Behavior in the Material Record: Ethnographic and Archaeological Examples. Current Anthropology 30:283-315. 1993 The Leavitt Site: A Parkhill Phase Paleo-Indian Occupation in Central Michigan. Memoirs of the Museum of Anthropology, University of Michigan 25. Simons, Donald B. 1996 The Gainey and Butler Sites as Focal Points for Caribou and People. In Caribou and Reindeer Hunters of the Northern Hemisphere, edited by L. Jackson and P. Thacker. Avebury Press, Glasgow, Scotland, in press. Simons, Donald B., Michael 1. Shott, and Henry T. Wright 1984 The Gainey Site: Variability in a Great Lakes raleo-Indian Assemblage. Archaeology of Eastern North America 12: 266-279. Spiess, Arthur E. 1984 Arctic Garbage and New England Paleo-Indians: The Single Occupation Option. Archaeology of Eastern North America 12:280-285. Spiess, Arthur E., James Bradley, and Deborah Wilson 1995 New England-Maritimes Region Early Paleo-Indian Occupation and Great Lakes Region Parallels. Paper presented in Symposium: Early Paleo-Indian Cultural-Ecological Adaptations in the Late Pleistocene Great Lakes Region and Northeast: Common Strategies or a Diversity of Approaches, organized by Dr. Peter L. Storck. 60th Annual Meeting, Society for American Archaeology, Minneapolis, Minnesota. Spiess, Arthur E. and Deborah Wilson 1987 Michaud: A Paleo indian Site in the Maine-Maritimes Region. Occasional Publications in Maine Archaeology No.6. Stewart, A. 1984 The Zander Site: Paleo-Indian Occupation of the Southern Holland Marsh Region of Ontario. Ontario Archaeology 42:45-79. n.d. Site Structure: Inter-Area Comparison of Debit age, Artifacts and Raw Materials. In The Fisher Site: Archaeological, Geological and Paleobotanical Studies at an Early Paleo-Indian Site in Southern Ontario, edited by P. L. Storck. Memoirs of the Museum of Anthropology, University of Michigan, in press. Storck, Peter L. 1979 A Report on the Banting and Hussey Sites: Two Palaeo-Indian Campsites in Simcoe County, Southern Ontario. National Museum of Man, Archaeological Survey of Canada, Mercury Series Paper No. 93. 1982 Palaeo-mdian Settlement Patterns Associated with the Strandline of Glacial Lake Algonquin in Southcentral Ontario. Canadian Journal of Archaeology 6: 1-31. 1983 The Fisher Site, Fluting Techniques and Paleo-Indian Cultural Relationships. Archaeology of Eastern North America 11:80-97. 1984 Glacial Lake Algonquin and Paleo-Indian Settlement Patterns in Southcentral Ontario. Archaeology of Eastern North America 12:286-298. 1990 Imperialists Without a State: The Cultural Dynamics of Early Paleo indian Colonization as Seen from the Great Lakes Region. In Clovis Origins andAdaptations, edited by R. Bonnichsen and K. Turnmire, pp. 153-162. Centre for the Study of the First Americans, Peopling of the Americas Publications, Corvallis, Oregon. 1994 Case Closed: The Fisher File. Rotunda 27(4):34-40. Storck, Peter L., editor n.d. The Fisher Site: Archaeological, Geological and Paleobotanical Studies at an Early Paleo-Indian Site in Southern Ontario. Memoirs of the Museum of Anthropology, University of Michigan, in press. Storck, Peter L. and Arthur E. Spiess 1994 The Significance ofNew Faunal Identifications Attributed to an Early Paleo indian (Gainey Complex) Occupation at the Udora Site, Ontario. American Antiquity 59:121-142. Storck, Peter L. and John Tomenchuk 1990 An Early Paleo-Indian Cache of Informal Tools at the Udora Site, Ontario. In Early Paleo-Indian Economies of Eastern North America, edited by K. Tankersley and B. Isaac, pp. 45-93. Research in Economic Anthropology, Supplement 5. Storck, Peter, and Peter H. von Bitter 1989 The Geological Age and Occurrence of Fossil Hill Formation Chert: Implications for Early Paleoindian Settlement Patterns. In Eastern Paleoindian Lithic Resource Use, edited by Christopher 1. Ellis and Jonathan C. Lothrop, pp. 165-190. Westview Press, Boulder, Colorado. Tankersley, Kenneth 1994 Devil's Nose: A Parkhill Complex Site in Western New York. Current Research in the Pleistocene 11:51-53 Timmins, P. 1994 Alder Creek: A Paleo-Indian Crowfield Phase Manifestation in the Region of Waterloo, Ontario. Midcontinental Journal of Archaeology 19(2): 170-197. Voss, 1. A. 1977 The Bames Site: Functional and Stylistic Variability in a Small Paleo-Indian Assemblage. Midcontinental Journtil of Archaeology 2:283-305. Wilmsen, E. N. and F. H. H. Roberts Jr. 1984 Lindenmeier, 1934-74: Concluding Report on Investigations. Smithsonian Contributions to Anthropology No. 24. Witthoft, John W. 1952 A Paleo-Indian Site in Eastern Pennsylvania: An Early Hunting Culture. Proceedings of the American Philospohical Society 96(4):464-495. Wortner, S. and C. 1. Ellis 1993 The Snary Early Paleo-Indian Site. Kewa 93(2):3-15. Wright, H. T. 1981 Technology and Economy: Recent Research on Early Hunters of the Central Great Lakes. Paper presented at the 46th Annual Meeting of the Society for American Archaeology. San Diego. California. 1989 Late Glacial Foragers in Eastern North America. In Easlern Paleo indian Lithic Resource Use. edited by C. Ellis and 1. Lothrop, pp. 345-351. Westview Press, Boulder, Colorado. Wright, H.T. and W.B. Roosa 1966 The Barnes Site: A Fluted Point Assemblage from the Great Lakes Region. American Antiquity 31 :850-860. The preceding article was subjected to formal peer review prior to pub Iication.