NORTHEAST PROPERTY TYPES
In the ten northeastern states, fourteen NRHP listings mention Paleoindian resources; no established Historic Districts are devoted specifically to them. This situation is partly due to the loss of integrity of most prehistoric sites in the region, given its lengthy history of Euro-American development and landform destruction and the intensity of private collecting on sites. Compromised integrity will continue to limit resource interpretation, listing, and protection in the region.
Sites and Districts
Among these sites, the Dutchess Quarry Caves must be reconsidered, as the association between a Paleoindian artifact and the caribou bone in Cave #1, dated to 14,300-15,200 years ago, is no longer considered supportable (Steadman and Funk 1987).
Two National Register Districts and one Landmark include Paleoindian sites and find spots: the Riverside Archeological District in central Massachusetts, the New Hampshire Veterans' Association Historic District at the Weirs in New Hampshire, and the Abbott Farm National Historical Landmark district in New Jersey. All three listings recognize extensive areas of dense prehistoric sites, including Paleoindian artifacts.
The record of non-listed Paleoindian sites, expanding at an unprecedented rate, defies counting. Publication, descriptive or interpretive, lags badly, but news of finds is quickly disseminated.
Sites listed on the NRHP, inventoried at the state level and/or represented in the literature, exemplify several categories, displayed on Table 16. Quarry sites and lithic workshops are the least problematic categories for this review. The ‘occupations' category, however, is problematic; few sites have been investigated and interpreted with sufficient thoroughness to demonstrate specific domestic activities on site. Here, only sites with domestic features or clearly distinct tool clusters are listed as occupation sites. Small sites interpreted provisionally as ‘hunting stands' may be in fact small camps, and therefore occupation sites.
* = Possible NHL candidate
designated an NHL
Isolated Finds and Small Sites
Small sites in the Northeast might be dismissed lightly as stray finds except that when opportunity to confirm the finds is taken, additional artifacts are typically found—flakes of exotic cryptocrystalline rock, uniface tools, bifacial trimming flakes of characteristic forms, channel flakes, scrapers, or pièces esquillées. Some few have demonstrable stratigraphical integrity and association with datable organics (e.g., Sanger et al. 1992). Distribution maps of such small finds provide a richer documentation of Paleoindian spatial activities than do accounts of areas called "sites" (e.g., Custer 1984b, 1986; Lantz 1984; Lyon 1989). In Rhode Island (SHPO office), Long Island (Saxon 1973), Delaware (Custer 1984a), and coastal Massachusetts (Mahlstedt 1987), isolated finds constitute the entire record of Paleoindian activities (Table 17). In Ohio, Lepper (1989a) showed that information about land use can be teased out of inventories of isolated finds, carefully examined and recorded. No comparable study exists in the Northeast, where identifying and contacting owners is difficult.
* = Possible NHL candidate
designated an NHL
Four instances of Paleoindian caching behavior are reported in the Northeast, three of them by a single investigator. At the DEDIC/Sugarloaf site in Massachusetts, a deposit of biface preforms and raw material was discovered close to other artifacts (Gramly 1998:34). The Lamb site in western New York was originally interpreted as a dwelling site with bifaces dispersed among a restricted inventory of other tools (Gramly 1988b: 273) but later reported as a biface cache (Gramly and Funk 1990: Figure 1 caption; Tankersley et al. 1997:31). At the Adkins site in Maine a boulder pile on the shore of an artificial lake was declared a "meat cache" (Gramly 1988a, d). Despite the resemblance to an arbitrary subset of a massive boulder deposit on an ice-plowed shore, no chemical or other critical tests were applied, nor were artifacts associated to attest to caching behavior. Selected boulders were airlifted to the Maine State Museum and installed in an attractive display. A cache of twelve lanceolate points of Late Paleoindian (Early Holocene) style is reported from Yarmouth on Cape Cod (Dunford and O'Brien 1997:34).
Bone beds and kill sites
Uncalcined boney materials preserve poorly outside of northeastern bogs or caves. Mastodont remains have been found in Massachusetts, New York, New Jersey, and Pennsylvania bogs, none associated with artifacts or other clear evidence of human intervention (e.g., Hoff 1969; Kurten and Anderson 1980: 344-346; Lepper and Meltzer 1991; Moeller 1984c). Mastodont and mammoth remains are dragged from the inner continental shelf, also without artifacts (Edwards and Emery 1977; Edwards and Merrill 1977; Oldale et al. 1987). In Pennsylvania and New York, limestone caves and crevices have yielded late Pleistocene fauna, all without evidence of human intervention in the life cycle (Guilday 1982, 1984). The association of caribou bone and fluted point at the Dutchess Quarry Cave #1 has been dismissed as fortuitous (Steadman and Funk 1987). At the Hiscock paleontological site in NY some Paleoindian tools have been found among Pleistocene fauna, without evidence of direct kills (Laub et al. 1988, 1996). The claimed caribou kill site near Vail meets the Criteria for such a site despite lack of organic remains: numerous tip fragments of fluted points, some of which match bases from the domestic site across the river (Ewing 1981; Gramly 1984a). Lepper and Meltzer suggest (1991) that searching for kill sites in the East may be a futility supported only by hopes of matching dramatic western slaughters.
Within the region, no human remains of great age have been found because of poor preservation conditions. However, the late Paleoindian Crowfield site in Ontario may be a cremation burial (Deller and Ellis 1984); others might be found.
Rock art and other representations
No Paleoindian rock wall art has been recognized in the Northeast, although late prehistoric examples are numerous. Engraved steatite pendants at the Reagen site in Vermont are unique and dubious (Ritchie 1953: Fig. 89). A soft pebble with scratch marks has been noted at the DEDIC/Sugarloaf site in MA (Gramly 1998:19, 63). An equally enigmatic scratched quartzite pebble was found in the Paleoindian context at West Athens Hill, NY (Funk 1973:27, 30). Drilled stone beads, conventionally considered items of personal adornment, possibly talismanic, are reported in poor context from DEDIC/Sugarloaf and Hiscock (Gramly 1998; Laub 1995a); their recovery at two sites commands attention.
Quarries and workshops
Northeastern stone quarry sites are more diverse than those of the Southeast and rarely closely associated with living sites. Research interest in bedrock has swelled lately (Hatch 1994; King et al. 1997; La Porta 1994; Luedtke 1987, 1993; Pollock et al. 1996; Pollock et al. 1999), powered by interest in tracing Paleoindian movements; the inventory cited here will likely expand. Sources of raw materials utilized by Paleoindians are known approximately, without direct evidence of Paleoindian quarrying, notably in Vermont (Mt. Independence, Hathaway/St. Albans, Colchester ‘jasper,' and Cheshire quartzite), Mt. Jasper in New Hampshire, Flint Mine Hill in NY, and numerous small quarry areas in Pennsylvania, Delaware, and elsewhere (cf. Wray 1948). Paleoindians in the Northeast quickly learned to identify non-cherty siliceous rocks: fine-grained metarhyolites, crystal quartz, silicified siltstones. Some of these have point sources certified by lithological Criteria alone, but lack evidence of Paleoindian mining. West Athens Hill in New York, the Saugus ‘jasper' quarry area in Massachusetts, and bedrock sources in the Munsungun Lakes region of Maine do have Paleoindian quarry debris (Bonnichsen et al. 1980; Funk 1973; Grimes et al. 1984). Additionally, in glaciated and periglacial areas appropriate rock available in colluvium, outwash gravels and alluvium was apparently utilized; more analysis is needed to confirm this (Crissel 1999; Custer et al. 1983; Holland and Dincauze 1999; Moeller 1980). Boulder-field quarries are entering awareness (Boisvert 1998).
Major known quarry sites are listed in Table 16. Sources at the scale of the Onondaga escarpment and Flint Mine Hill required no great skill for recent immigrants to find, but some are truly cryptic, requiring not only an eye for lithologies (e.g., Nevers), but effort for recovery (e.g., Gramly 1984b). The actual scale of Paleoindian quarry activity may be beyond knowing, given the millennia of recurrent human activities at the sites; studies of long-distance transport will be informative as petrological studies expand (e.g., Pollock on Mt. Jasper and Munsungun sources). Gravel quarry sources in alluvium of the Delaware and Susquehanna rivers, and the Housatonic River (La Porta 1994:55), remain to be studied, along with suspected sources in western Connecticut (Moeller 1980).
Débitage from tool-making and repairing is expectable on any Paleoindian site. Criteria for recognizing ‘workshops' as a special class of site vary with the reporter. Famous spreads of Paleoindian flaking debris characterize sites in proximity to quarries, among which are the Chase-Windy City group in Maine (Payne 1987), West Athens Hill, Kings Road, Corditaipe, Emanon Pond, and Arc in New York. Other prehistoric debris-spreads (e.g., Flint Mine Hill NY, Conklin RI, Vera Cruz PA) cannot be associated specifically with Paleoindian activities, although close to bedrock known to have been used in that remote time.
Given the high percentage of damaged sites in the Northeast and the infrequent professional excavation accorded them, attribution of occupational function dominates the literature as a default category. Some sites such as Bull Brook I, Arc, and Shoop clearly qualify by size and redundant clusters of tools. Adkins, Bull Brook II, Lamb, Michaud, Neponset, Templeton, Vail, and Whipple most probably are. DEDIC/Sugarloaf may be of this class but that has not been clearly demonstrated. Lanceolate point occupation sites are especially elusive (Nicholas?). Careful excavation at small camps such as Hidden Creek in Connecticut (Jones 1997) can reveal domestic sites replete with other functions. A large area mapped in Vermont (Mahan site, VT-CH-197) is notable for the numerical dominance of scraper forms in a site yielding only one fluted point (Thomas et al. 1998); it has not been demonstrated to have been uncollected prior to investigation. Plowed sites, such as Potts, Twin Fields, and Plenge can be interpreted functionally only by analysis of the tools. Functional analysis of stone tools is a research approach grossly underutilized in the Northeast. The largest northeastern sites, clearly with some residential function, have been variously interpreted as aggregation sites for seasonal hunts and/or for seasonal gatherings for socializing people otherwise separated during the year (Curran and Grimes 1989; Spiess 1984). They are also proposed as a special kind of colonizing site (Dincauze 1993b). All have point styles in Groups I and II as defined below (Table 18).
Small sites yield point styles of all groups, most typically Group III (Jacobson 2001; Table 11). This is probably significant; large sites with Group III points are known only in Ontario, where they are convincingly modeled as seasonal aggregation sites (Ellis and Deller 1997; Storck 1997). In northeastern US, activity evidence tends to be less diverse at the small sites than at the large ones, indicative of shorter, more focused, residence. Several sites with Group III and later points are multi-component compilations: e.g., Wapanucket 8, Plenge, Potts, Reagen, and others. Numerous small sites in Pennsylvania are considered short-stay camps, but the point styles are not emphasized in the literature (Carr et al. 1996; Lantz 1984).
No northeastern cave occupations are reported with fluted points. Most paleontological caves are predator collections or natural traps (e.g., Dutchess Quarry Cave #1, a fissure). Rockshelters typically lack evidence of Paleoindian use, and may be mainly Holocene phenomena. Two rockshelter finds of fluted points are reported in Connecticut (Moeller in Brennan 1982:41) without details. The unique Pleistocene lanceolate biface at Meadowcroft emphasizes this point; Clovis-style bifaces occur in the Cross Creek vicinity, but not in the shelter.
Sub-regional variation in relief, altitude, hydrography, and climate complicates efforts to generalize about site locations favored by Paleoindians. Isolated fluted points are found widely across the region, restricted only by available data which varies directly with the density of archeologically informed searchers. The Paleoindian Northeast, as summarized above, was bounded by sea on the east and northwest, as well as by glacial lakes northwest. The northern boundary is political. The grain of the country follows subparallel mountain ranges trending mainly NE-SW; the slopes of mountain ranges and coastal plains define the routes of rivers. Sea level was low throughout the period in reference to today, except for a Late-Glacial intrusion into Central Maine that may have been partly visible to the earliest Paleoindians. Late-glacial marine eustatic transgression, slowed by the Younger Dryas climatic reversal, was much less dramatic than it had been and would become; in places it was reversed by isostatic rebound (Belknap et al. 1987). After 13,000 years ago, prevailing winds relaxed from the Late-Glacial northwest and blew more equably from the west (Thorson and Schile 1995). Climate and seasonality in the Late Glacial period became responsive to altitude and proximity to the sea. Land surfaces tilted and rose with isostatic rebound that followed the ice melt and drainage of the large lakes. Continental and mountain glaciers and meltwater had scoured valleys in the north of the region, bringing river channels to very low elevations that were buried anew by late-glacial lake deposits and/or rising sea levels backing up major drainages such as the Susquehanna, Delaware, Hudson, Connecticut, and Penobscot rivers. Rivers draining toward the Mississippi, such as the Ohio, Allegheny and Monongahela systems, similarly cut deeply during glacial time, leaving high old terraces on their valley flanks, later raising the channels and bringing early Holocene alluvium over the lowest Late-Glacial terraces (e.g., Broyles 1966). The dynamism of Late-Glacial alluvial deposition may explain why rockshelter utilization is so rare among Paleoindians in North America.
Distributions of resources that sustained life for Paleoindians—plant and animal foods, potable water, well-drained living surfaces, and tool-quality lithics—were not static throughout the deglacial millennia. Boulders and gravel, moved southward by ice and water, made tool materials available at various distances from their bedrock sources. North of glacial moraines, the accessibility of bedrock outcrops buried under ice, glacial sediments, or lakes gradually improved. The dynamism of the northeastern Late-Glacial environment makes prediction of Paleoindian distributions particularly difficult. The following discussion is framed in terms of environmental categories rather than modern political divisions.
Coasts and shores
Seacoast defined the eastern limit of the Paleoindian Northeast from Delaware to Maine, and the northwestern limit in Vermont, where the Champlain lowland was briefly an arm of the sea. West of the Champlain Sea lay the freshwater Great Lakes, whose shores shifted markedly throughout the millennia, from early extensive overflows of the land to later lowstands that permitted Paleoindian exploration on surfaces now underwater (Karrow and Calkins 1985; Laub et al 1988; MacDonald 1995). We only glimpse Paleoindian life near New York shores, in contrast to the rich record in Ontario. Marine transgression began in the south, where the isostatic depression was shallowest and briefest. The transgression is still being expressed north of the Gulf of St. Lawrence. In Maine the uplifted Paleoindian coastal zone, now being eroded (Kelley et al. 1992), accounts for the southern Maine sites on deltas and dunes such as Michaud, Lamoreau, and Hedden (Petersen 1995; Spiess et al. 1998). The generalizations offered by Spiess and Wilson (1987:130-132) apply only to coastal southwestern Maine, which has marine deltas and associated landforms of the appropriate age. Elsewhere, coastal and inland dune building was earlier (Thorson and Schile 1995) and the marine deltas older and lower. Two of the rare northeastern large sites—Bull Brook and Debert—are situated near, but not on, their coeval coastal plains, which were exposed by the Younger Dryas low sea (Oldale 1985, 1986). There may have been unique resource attractions on the expanded plains, but huge herds of migrating caribou are unlikely to be among them (see "Game" section below).
South from Maine, site densities decline coastwards, reflecting the loss of sites directly on the transgressing coast. Lower courses of rivers that ran into the Late Glacial sea are now deep in alluvium or marine muds, any sites buried with them. Off the south coast of Connecticut at Hammonasset State Park, an "eastern Clovis" point appeared in sand dredged from under 16-18 feet of water (Glynn 1969:70), lost on land since inundated by Long Island Sound. Rhode Island has very few known Paleoindian find-spots and no sites, although the range of point styles is representative. Massachusetts' Cape Cod and islands show the same situation (Mahlstedt 1987; Richardson and Petersen 1992). Long Island has many finds, but no sites (Saxon 1973). In New Jersey and Delaware, known sites and find spots cluster on uplands landward of today's coast (Custer 1984a, 1996; Custer et al. 1983; Marshall 1982; Mounier et al. 1993). In contrast to the stray finds elsewhere, the Inner Coastal Plain of New Jersey, with rivers draining west toward the Delaware River, has most of the sites (Custer 1996; Grumet 1990).
Significant numbers of Group I and II points in the dunes of the Champlain seacoast of Vermont are noteworthy (Loring 1980). Marine environments had marine animals which could have been taken by the Paleoindians. Utilization of the fresh water Great Lakes shores of New York and Pennsylvania is not attested in the literature, perhaps because of subsequent inundation by rising Holocene water levels (Anderson and Lewis 1985). The Late-Glacial Younger Dryas period Champlain Sea probably possibly featured icebergs, though unlikely as the glacial front was farther north. The Late Paleoindian Reagen site at the north end of the Champlain basin is below the elevation of the marine incursion; but within the shore area for brackish or freshwater successors. The Davis site on the New York shore of the basin has yielded a few broken fluted points that may belong in Group III. The findspot lies below the Champlain Sea water level (Ritchie 1965:19-21). The age of the site and its setting requires clarification (Snow 1980:142).
Much has been made of Paleoindian sites clustering in river valleys (e.g., Anderson 1990a; Anderson and Faught 1998), but in the Northeast large valleys have find spots and small sites, whereas upland low-order streams have the larger sites (Arc, Bull Brook, Vail, Shoop, etc.) and numerous small sites (e.g., Lantz 1984). If DEDIC/Sugarloaf on the Connecticut River is really a big site, as claimed (Gramly 1998), it will be an exception to a well-established rule. The big river valleys, and valley passes through mountains, were travel routes for both people and game, but not preferred places for domiciles.
Significantly, no Paleoindian sites or findspots are reported on the immediate eastern shore of the Hudson River trench, where the topography appears to have been appropriate (Haviland and Power 1981: 35; Ritchie 1957:11; Funk, pers. comm. 1999, concurs with this). This observation bears on initial colonizing routes (see Research Needs and Questions for the Northeast). Hudson Valley cherts on Paleoindian sites in Massachusetts correlate strongly with the presence of Group III points. The Hudson trench was formidable during the low marine levels (Dineen 1996); people might well have been willing to walk north to find better crossing places. The known distribution of Group Ia points in New England is heavily skewed toward the north (Table 18; Dincauze and Jacobson 2001).
Like their prey or because of it, Paleoindians were attracted to springs, including mineral springs in western New York (Arc, Hiscock), and to the shores of lakes and swamps (Wapanucket 8, Bull Brook). Even Shoop in its mountain valley may reflect an attraction to nearby upland wetlands (Custer 1996:120). Low-order (headwater) streams are implicated as chosen places on Long Island (Saxon 1973). Wetlands in drained glacial lakes may have been attractors, as clusters of isolated finds imply (Research Needs and Questions for the Northeast; Curran and Dincauze 1977; Nicholas 1988); however, large Paleoindian sites have not yet been found near them. Drained lake basins do appear to have been favored during the drier climates of the Early Holocene, especially by users of early notched points (Nicholas 1988; Webb et al. 1993:111).
Because some Paleoindian sites are buried, inundated, ablated, or eroded away, accessible sites cannot be the full record of original site locations and preferred landscapes. Schuldenrein (1994) has properly employed the Schumm (1977) model of stream basin development to argue that Paleoindian and later sites along the middle reach of the Delaware are on alluvial surfaces of appropriate age, which are knowable. Along major river valleys Paleoindians spent their time and lost their signature artifacts on well-drained elevated surfaces such as delta plains, dune fields, and high terraces (e.g., Curran and Dincauze 1977). These may be the areas where Post-Pleistocene erosion has removed less of the surface, and buried less of it, than in areas neighboring higher-order streams. What is clear from the upland sites we know throughout the region is that they lie predominantly on droughty soils, possibly chosen for edge habitats as well as for warmth and dryness. McWeeney (1994) has discussed and speculated about Criteria for locating Paleoindian sites by analysis of landscape age; just such an analysis sent a survey team to DEDIC looking for Paleoindian remains. Many such areas are fire-prone, which might explain some of the visibility problems and troubles with radiocarbon dates (e.g., Spiess et al. 1995). Fires in Paleoindian times may explain how Paleoindians found and quarried boulders at New Hampshire's Nevers site (Boisvert 1998), which otherwise seem to be small phenomena, easily overlooked.
Some Paleoindian find spots are located away from living places favored by people who occupied the region after 11,500 years ago, probably because of the reorganized hydrography that followed incision of the major rivers into glacial deposits. This is not an iron-clad rule: Plenge, Twin Fields, Bull Brook, Potts, Davis, Shoop, Wapanucket, Neponset, Port Mobil, Kings Road, DEDIC and others also yield scatters of later artifacts. The deeply stratified Meadowcroft and Shawnee-Minisink sites are exceptional in this respect, with their long histories of post-Paleoindian use.
No fluted point sites have been reported north of the St. Lawrence east of Ontario, where the glacial front halted at the St. Narcisse moraine for much of the Younger Dryas (Dincauze and Jacobson 2001). Nor have any been reported from the tundra zone immediately south of the St. Lawrence trench; all are within the zone of coeval parkland vegetation. However, the northernmost sites, Vail in Maine and Debert and smaller areas in Nova Scotia, were not far from the narrow tundra belts of their time, especially during the Younger Dryas period (Dincauze 1988; Mayle et al. 1993).
McWeeney's research into the vegetation of northeastern Paleoindian landscapes is the richest to date. Relying more on macrofossils than on pollen, she was able to show the absence of coeval tundra near sites, their situation within spruce parkland mosaic landscapes, an unexpected diversity of hardwood species, and great instability in plant associations during the thirteenth millennium (McWeeney 1994). Her research confirmed a major warming episode prior to the Younger Dryas, cooling during it, and a subsequent rapid and early return of hardwoods and temperate forest elements in southern New England. She speculates that the northeastern Paleoindians arrived prior to fourteen thousand years ago (1994: 146), that their distributions contracted south during the Younger Dryas cold, and that shortly after 12,000 years ago, at the time of Group III styles, Holocene climate regimes with hardwood tree species prevailed into central New England and farther south (1994: 139-141; Kutzbach and Webb 1991). This radical revision of cherished notions fits well with the absence of evidence for caribou in Group III and younger sites in the United States. A critical test will require direct radiocarbon dating of organics strongly associated with Paleoindian features, evidence that has been elusive to date.
Recent AMS dates from Shawnee-Minisink fireplaces set that site, with temperate forest plant remains and a Group Ia point, into the early Younger Dryas at 12,900 years ago, in a southern location (Dent 1999). The generalizations we have lived with clearly mislead. It is critical to control for time, latitude, and altitude when modeling paleoecology for Paleoindians (Gaudreau 1988; Kutzbach and Webb 1991); the extant literature falls below that standard.
Radiocarbon complications make difficult also the evaluation of claims for big-game hunting in the Northeast. Mastodont remains are numerous from Massachusetts south and west, and east onto the continental shelf (Edwards and Merrill 1977). The latest proboscidian dates fall into the Younger Dryas millennium, but are there hung up on a radiocarbon plateau with some of the early site ages. No mastodont finds in the Northeast have Paleoindian artifacts associated, although a case has been made for a mastodont tusk artifact dated 13,000 years ago (Laub et al. 1996). New information implies that the elephants were among the last Pleistocene fauna to go extinct, close to 13,000 years ago (Elias 1999). The convergence of fluted point finds and wetlands, reflecting shared needs for water between prey and predators, guaranteed encounters as long as those were chronologically possible.
The allopatric Late Pleistocene assemblage of animals at the Hiscock site, including extinct and exotic species, is well dated to the thirteenth and fourteenth millennia, although the human associations are loose within the envelope (Laub et al. 1988). The conventional notion of Paleoindians as maniacal hunters of big game is being abandoned for more realistic expectations for adaptive behavior appropriate to the lightly forested terrains of the Late Pleistocene Northeast (Dincauze and Curran 1983; Lepper and Meltzer 1991). Huge herds of caribou are a Holocene arctic development (Loring 1997), and therefore not available as Paleoindian prey concentrations. Spiess argues eloquently to the contrary, but unpublished. A few Group Ia sites have yielded calcined caribou bone (Grimes et al. 1984; Spiess et al. 1985); a cervid at Group III Templeton may be Odocoileus. Pleistocene cervalces has not been found in human association, while beaver and fish appear in Group Ia sites (Arc, Bull Brook II, Whipple, Shawnee-Minisink).
We can populate the regional woodlands by analogy, but we have little direct evidence for what was taken from among what was likely present (Curran 1999; Curran and Dincauze 1977). Proboscidian attraction to wetlands and rivers, especially when ill, increased encounter probabilities and opportunities for scavenging carcasses. By the time of Paleoindian residence in the Northeast, there had been several millennia during which fish could return to inland lakes and rivers. Late Pleistocene flyways terminated northward at the Great Lakes and ice front, potentially creating high concentrations of birds in season (Dincauze and Jacobson 2001). The lesson from scant evidence is that Paleoindians were taking animals that were available; the degree of selection is beyond calculating. Again, latitude, altitude, and age are the critical variables for modeling.
The defining role of bedrock quarries in the distribution and adaptation of Paleoindians is less obvious in the Northeast than it is in Virginia (Gardner 1977, 1981), with the possible exception of the Delaware coastal plain. Given discontinuous distributions of high quality cryptocrystalline rocks favored by Paleoindians, exploration and discovery was likely initially intensive, revealing outcrops and boulder fields of knappable volcanic and sedimentary rocks used throughout the remaining span of fluted point manufacture. Prior to awareness of the need for petrographic support for source attributions, it was easy for archeologists to claim, without controls, that Paleoindians depended upon long-distance transport or trade in high-quality exotic lithics.
Research is moving from ‘eye-ball' identification of bedrock sources in terms of color and grain into more responsible intensive physical-chemical research that will ultimately reliably indicate the directions of movement of Paleoindian groups, and help delineate the range sizes characteristic of different time periods (Hamilton and Pollock 1996; La Porta 1994; Pollock et al. 1999; Tankersley 1988). Exotic distant lithics occur most often in sites with Group I spearheads; by the time of Group III, local or regional sources of stone dominate assemblages throughout the region (Jacobson 2001). Sources for the materials of Group II points remain notably enigmatic, although local sources are becoming more likely. Continuing work on the identification of both bedrock quarries and sequential styles of Paleoindian spear points will provide details of diachronic change and exploration patterns.
As a consequence of the sobering new realities, models of social and economic groups and ranges for Paleoindian hunters based on lithic utilization patterns need reevaluation or at least caution (e.g., Curran 1999; Dincauze 1993a; Gramly 1988b; Spiess and Wilson 1987a; Spiess et al. 1998). In contrast to Gardner's ‘lithic tethering' model of Paleoindian territoriality, north of Virginia Custer's serial model (1986) is a better fit to available data on Paleoindian use and distribution of lithic raw materials (Carr 1998). This supports the ‘embedded procurement' model of lithic exploitation propounded by Seeman (1994) in Ohio.
Distributions of point styles
The observed density of fluted point sites and find spots, low in comparison to later cultural periods (except for late Paleoindian and EA), is impressive given the intensity of land disturbance and artifact collecting in the northeastern US (Table 17). One can conclude that fluted points are widely distributed in the Northeast, controlled by the distribution of rivers, lakes and wetlands, with secondary control by outcrops of useful stone types and relatively stable landforms. While suggestive, these observations are not yet supportive of strong research orientations. Many authors try to infer patterns of movement by Paleoindian groups coming into the region or, alternatively, pursuing resources in seasonal cycles (Curran and Grimes 1989; Custer 1996; Dincauze 1993b; Dincauze and Jacobson 2001; Ellis et al. 1998; Gramly 1988b; Gramly and Funk 1990; Lantz 1984; Spiess and Wilson 1987a; Spiess et al. 1998; Tankersley 1991, 1995). Recently, such studies appear close to exhausting the information potential of the regional data, where organic materials and intact patterning internal to sites are lacking. However, very different results are obtainable when the stylistic sequence in fluted points is added. For example, Lantz's classic study (1984) of Paleoindian immigration through western Pennsylvania to western New York maps all fluted point styles as equals. Now that Group I styles are acknowledged older than Group III, and Group IV is seen to belong to the Pleistocene/Holocene boundary, while Groups V and VI are Early Holocene, lumping them all together as immigrants clearly distorts the information.
The style groups of Table 18 are not cleanly discrete. There are intermediates between all of them, and more than one may appear in a given site. These two observations support the inference that Paleoindian use of the region was essentially continuous following the initial colonization. However, the fact that none of the fluted point style groups after II appears in the Canadian Maritimes or the area directly south of the international boundary may indicate a withdrawal of occupation from the northern fringe during the coldest YD times (this inference depends on the ultimate dating of Group II). New England, New York and western Pennsylvania fluted points fit well within the Great Lakes style clusters, but individual specimens are not always close to named styles defined there. The intermediates, especially, are difficult to classify. This variable degree of independence between regions weakens any arguments for more than one pioneering episode in the greater Northeast, while implicating the maintenance of wide-ranging communication and mating networks throughout the thirteenth millennium.
Group Ia points are distributed continuously from northwestern Pennsylvania through New York, especially along the lake fronts, northeasterly across northern Vermont, New Hampshire, and Maine, and south into Massachusetts, Connecticut, Rhode Island and the coastal plain, located on well drained Late-Glacial landforms near watercourses. They leave a trail of Ohio cherts in western Pennsylvania (Lantz 1984) and western New York (Gramly 1988b; Tankersley 1994c, 1995), of Onondaga cherts across north-central New York, eastern New York and Vermont cherts in the Hudson and Champlain valleys, and Munsungun cherts south along the coastal plain into eastern New England (Pollock et al. 1999), along with a few exotic lithics (i.e., Spiess et al. 1998). Vermont cherts occur in western New England beyond the limits of the Munsungun distribution (Dincauze and Jacobson 2001). In the study area, Group Ia points are the most frequently reported and most widespread style of fluted point.
Group Ib, the Shoop-like style group, has proven difficult to define with certainty; the manufacturing sequence is still debated (Callahan 1979; Carr 1989; Cox 1986:128-129; Fogelman 1986; Gardner and Verrey 1979; Painter 1973; Witthoft 1952, 1954, 1962). The distribution is mainly south of the New York/Pennsylvania border (Table 18). If it does indeed represent a population intrusive from the south, using locally deposited cherts rather than bringing bedrock from New York, its slight degree of stylistic strangeness in the Northeast may be actually unproblematic (cf. Moeller 1989).
Group II points dominate in northern New England, being possibly more numerous there than the Group Ia styles. In the main sites they appear to be made of fairly local cherts. The type is not yet well defined, limiting the incisiveness of comparisons, but there seems to be a wide distribution of single specimens with the typical deep basal indentation southwest in Vermont and across New York along the lakes, with a trickle south toward New Jersey and interior Pennsylvania (Custer 1996:121, 130). Whether this tail-off distribution represents more than stylistic insouciance or trading in exotics remains for rigorous analysis, and more secure ages, to determine (Morrow and Morrow 1999).
Group III points have a distribution more southerly, being rare to absent in the international border area where Group II may be partly coeval with them, and extending south into Pennsylvania and New Jersey but becoming rare in that direction (Custer 1996:130). They do not occur on sites larger than Michaud, in contrast to the situation in Ontario. They appear in small sites near water or wetlands; their spatial relationships in the coastal zone are not yet clear. Hudson Valley cherts appear with Barnes points in southern New England, but local sources of stone are more typical. Barnes points may ultimately be recognized widely at small, dispersed quarries throughout the region.
Groups IV and V styles, dated at the Pleistocene/Holocene boundary, are considerably less frequently reported than are the earlier styles. They retain a stronger presence in the north of the region, but occur as strays into Pennsylvania and New Jersey. Crowfield and Holcombe distributions have been extended easterly from the Great Lakes states, on good grounds (Spiess et al. 1998). Holcombe points, with slight or no fluting, have been reported mostly in Maine and Vermont, and seem not to occur south of New England, in contrast to Crowfield which occur south to New Jersey. Spiess et al. (1998), however, see Holcombe points at sites otherwise considered to have stubby lanceolates, supported by recent work at the Nicholas site in Maine (Wilson et al. 1995). In the Middle Atlantic states, New Jersey has the Turkey Swamp group (Cavallo 1981); New York has them at Piping Rock (Brennan 1977) and Port Mobil. The very similar Pennsylvania Miller point style, at Meadowcroft Rockshelter and along the west-flowing Cross Creek drainage (Adovasio 1983; Carr et al. 1996; Lantz 1984), may be earlier, around fourteen thousand years ago (Boldurian 1985).
Group VI, a heterogeneous group labeled "lanceolate," extends northeastward to the St. Lawrence lowland where it replaces or displaces the apparently coeval Early Archaic notched points. Sites of this tradition in Québec are found on the south shore of the St. Lawrence along the coast of the post-glacial Goldthwait Sea that succeeded the Champlain Sea of the fluted point sites farther west (Dumais 2000:105). Until they are studied comparatively, the many names by which lanceolates are tagged in the northeastern literature will be subject to revision. Agate Basin, Dalton, Eden, Hell Gap, Hi-Lo, Plano, and Scottsbluff have been invoked, while the type name "Ste. Anne" is gaining currency well beyond its origin area on the south shore of the St. Lawrence into northern Maine (Cox and Petersen 1997; Dumais 2000; Petersen et al. 2000). While the Western typology may be questioned, some specimens could be rare trade goods. Except for reaching farther north, lanceolate biface distributions are remarkably similar to the earlier Group I points, being clustered in the Great Lakes area, and northern New England (Doyle et al. 1985:Figure 23). These latest Paleoindian styles are less numerous southward from central New York and Massachusetts (Cox and Petersen 1997; Doyle et al. 1985; Fowler 1972; Funk and Schambach 1964; Jones 1997; Maymon and Bolian 1992; Petersen and Putnam 1992; Sanger et al. 1992, Spiess 1992; Stanzeski 1996; Tankersley et al. 1996), and are so far scarce in Vermont (Thomas 1992:188). Rarely solidly dated, they are nevertheless dependably Early Holocene in age (Table 6). There is a serious need for analysis and subdivision of this category.
Several different styles of Early Archaic notched points trickle into the Northeast between 11,500 and 10,000 years ago, contemporary with style groups V and VI. In Delaware, New Jersey, and coastal New York, notched Early Archaic points (Hardaway-Dalton, Kirk, Palmer) dominate at the time when the lanceolates occur to the north (Funk and Wellman 1984; Levine 1989; McNett 1985; Ritchie and Funk 1971; Stanzeski 1998). All have strong southern relationships, and are increasingly rare northward (Carr 1998; Custer 1996; Dincauze and Mulholland 1977; Pfeiffer 1986; chapters in Robinson et al. 1992). Geographically between the lanceolate points in Canada and the notched point Early Archaic in southern New England are early Holocene sites dominated by quartz tools and debitage and apparently lacking bifacial weapon tips; unless their ages remain in the ninth millennium they may force revision of current notions of early Holocene cultural dynamics (Curran 1994: 46; Robinson et al. 1992; Sanger et al. 1992). While diminishing in numbers north of New Jersey, notched Early Archaic points have been reported as isolates into northern New Hampshire and the central coast of Maine. Such strays may be trade items among coeval peoples, in parallel to the western styles of lanceolate points noted above.
In the Delaware valley at Plenge and Shawnee-Minisink, Paleoindian and Early Archaic components occur at the same sites, as is normal farther south. Farther north, in glaciated terrain, Paleoindian and Early Archaic sites are typically separated. This change occurs near the southern limit of Wisconsinan moraines and probably records nothing more cultural than the relative stability of landforms south of the moraines, where uplift ended earlier.
Large Sites and Clusters
Within the Northeast, some areas stand out as relative concentrations of Paleoindian and Early Archaic sites. Counts of Paleoindian points by state are skewed by such large sites as Vail and Bull Brook with their large artifact populations. Other clusters are groups of isolated point finds or small sites. Spiess et al. (1998:232) note differences with states south of New York, and affinities with the Great Lakes area, in both site types and distributions, but these may be superficial given the weakness of analytical studies to date. The Magalloway cluster near the Vail site includes also Adkins, two sites near Wheeler Dam, and the Cox, Morss, and Wight sites, inter alia (Gramly 1988a; Spiess et al. 1998:213). The Munsungun cluster is undoubtedly related to outcrops of good lithic material; at least ten sites are listed in the area in the Maine State inventory (Spiess et al. 1998:213). The sites are not yet well published, however (Bonnichsen 1981, 1984; Payne 1987; Pollock 1987a, Pollock et al. 1999).
Vermont Paleoindian sites and finds cluster west of the Green Mountains near the Late Glacial Champlain Sea (Loring 1980; P. Thomas pers. comm.), where both coastal biotic resources and good cherts were to be found. The large Mahan site should encourage research in this area (Thomas et al. 1998). Little evidence of Paleoindians has been reported in the northern Connecticut Valley in Vermont. New Hampshire has isolated finds clustering in the lower Merrimack drainage near Manchester and the Massabesic lakes (Curran 1994), with the Israel River group dominating the north (Boisvert 1999) and little between. The Whipple site accounts for most Paleoindian finds in the upper Connecticut Valley, where the valley floor is to date devoid of sites of that age.
Massachusetts has produced more fluted points than the literature now shows, with the Bull Brook/Ipswich group (Bull Brook I and II and Paisley) leading all of the northeast for sheer numbers of fluted points (Speiss et al. 1998). In Concord, within the basin of Glacial Lake Sudbury, at least six fluted points have been reported (Lyon 1989:10). The basin of Glacial Lake Hitchcock in the northern Connecticut Valley of Massachusetts has a cluster in the towns of Deerfield, Greenfield, Gill and Montague (Curran and Dincauze 1977; Lyon 1989:10), whether or not DEDIC/Sugarloaf is a large site (Gramly 1998). In the central part of the state is another sizable group including the unpublished Flanders/Winnimisset site and finds related to the Ware River drainage (Johnson and Mahlstedt 1985:26-28; Lyon 1989:11). In the southeastern corner, strays and the Wapanucket 8 site indicate early use of the area of the large lakes.
New York state offers an important cluster near the Onondaga escarpment in Genesee County, including the Arc, Hiscock, the unpublished Bear Paw/Bush sites, and others (Gramly 1988b; Holland 1994; Smith 1995; Tankersley et al.1997; Vanderlaan 1986). Finds along the Seneca River have indicated regular Paleoindian use at locations such as Baldwinsville, Cross Lake, Haiti Island, and Van Buren (J. Bradley pers. comm.; Spiess et al. 1998), and the Oneida River and Lake have produced a few finds (Caughdenoy Creek, Oneida Lake outlet, Muskrat Bay; J. Bradley pers. comm.). In eastern New York, near famous chert outcrops and the swampy beds of several glacial lakes, are Paleoindian sites such as Kings Road, Swale, and West Athens Hill, and the scattered finds near the Dutchess Quarry Caves (Funk 1973; Funk et al. 1969a,b,c; Funk and Steadman 1994; Kopper et al. 1980). Farther south on the Hudson estuary is the Port Mobil group (Kraft 1977a).
New Jersey Paleoindian sites cluster in the Delaware River drainage; the multi-component Plenge site and Zierdt (Kraft 1977b: 267; Werner 1964) are on the upper drainage. Isolates tend to be in the same drainage, or on higher elevations on the ocean side of the divide (Bello and Pagoulatos 1995; Grumet 1990).
On the Pennsylvania side of the Delaware River, a few known sites occur: Poirier, Shawnee-Minisink. More should be found when appropriate search methods are applied to landforms of Late Glacial age. An important group, including the Shoop site, occupies the Susquehanna River basin. In western Pennsylvania, Lantz (1984) has proposed the Allegheny Paleoindian Corridor, but the claim needs to be confirmed with disciplined studies of lithic sources and point styles. He noted that Crawford County has 24 sites/finds recorded and six are in Erie County south of Waterford (Lantz 1984:219). His survey found that sites "are twice as dense in the glaciated Allegheny Plateau than on the unglaciated Plateau" (1984: 213-4), where sites were "thickest" in Washington, Armstrong, and Indiana Counties (1984: 215). Carr et al. (1996) confirm that site distributions vary among the physiographic provinces of the state. The Meadowcroft site stands in splendid isolation, claiming neither related assemblages with early dates, nor fluted points (Adovasio et al. 1988; 1992).
Delaware has no single-component Paleoindian sites, but finds cluster on the central uplands and toward the Delaware River (Custer 1984b, 1986; Custer et al. 1983). The Delaware Chalcedony complex and Hughes Early Man Complex are quarry-related clusters of finds of both Paleoindian and Early Archaic points.
The Northeast emerges as an important area for Paleoindian sites and distributions, and continues to contribute information as the analytical vocabulary and theoretical perspectives mature and stabilize. Site types and distribution patterns differ significantly from those characterizing areas to the south and west. Glaciation and isostatic landform adjustments, glacial lakes, mountain ranges, and the sea contribute, with a dynamic Late Glacial climate, to the uniqueness of the area and the experience of its initial human populations.
The aesthetic specialness of fluted points of cherts and chalcedonies is so obvious to the casual observer that before there was any awareness of the issue of ‘first comers' or ‘big game hunters' there were names for these things. Among collectors in the Northeast they were called ‘Seneca points,' apparently in recognition of a Great Lakes connection. After the Folsom clarification, that name took over (Anon. 1936), although Edgar Howard in Pennsylvania was refreshingly careful (1934, 1942). As the Clovis terminology diffused, both ‘Clovis' and ‘Folsom' were used casually as envelope terms in the Northeast, preventing awareness of that diversity in time and space that now provides crucial insight into processes of colonization and adaptation. After mid-century, field investigations were dominated by museum professionals, among whom J. Witthoft (1952), W. A. Ritchie (1953, 1957), and D. Byers (1954), have pride of place by virtue of publishing and publicizing the Shoop, Reagen, and Bull Brook sites. Before those influential reports, these same archeologists had contributed to the growing avocational literature on fluted point distributions. Byers' experience with the Bull Brook site led to his involvement in excavation of the Debert site in eastern Canada that provided impetus for anthropological perspectives on early sites (MacDonald 1968) and, along with Jordan's attempt at interpreting the Bull Brook site (1960), set the tone for frustrations in dating sites by either radiocarbon, organic associations, or geochronology. An even more complicated, multi-component, site was reported about the same time (Robbins and Agogino 1964), raising interest nationally. Survey reports and site counts began to be available (e.g., Mason 1959). Emery and Edwards (1966) called attention to the importance for early cultural histories of the inundated continental shelves. Environmental determinism crept in as Ritchie joined J. Fitting in Michigan to ponder the chronological gap both perceived between fluted points and the nearly ubiquitous mid- to late-Holocene cultural remains in their respective research areas (Fitting 1968; Ritchie and Funk 1971). Neither at that time was able to identify elements of Early Archaic cultures, despite clear demonstrations to the south (Broyles 1966; Coe 1964). Mason (1962) saw the crucial difference between the frequency of fluted points in the East, versus their comparative rarity in the West, and sounded the call for explanation.
Prosperity and the boom in graduate education rapidly expanded the corps of researchers in the United States and Canada, initiating the era of University-based research. In Michigan, Roosa (1965), Fitting et al. (1966) and Wright and Roosa (1966) established a vocabulary for fluted point variation that reached the Northeast only later, after being expanded and grounded on geochronology in eastern Canada (Ellis and Deller 1988). By the middle of the 1970s northeastern Paleoindian studies, firmly wedded to paleoenvironmental research, were raising enduring issues: Callahan (1979) laid out a formal sequence for fluted biface manufacture; Dent (1985) and Eisenberg (1978) countered the Big Game Hunting model with evidence and a model of generalized foraging; Grimes (1979) elaborated the Debert issues of single vs. multiple occupations at big sites. New research was showcased at a seminal conference sponsored by the New York Academy of Sciences (Newman and Salwen 1977). Nearly every issue that has enlivened Paleoindian research in the region since then was presented or discussed there. Basic research on Paleoindian sites and site distributions expanded and engaged audiences, putting the region on national maps.
Building on Mason (1962), Brennan called for a fluted point census of eastern states (Brennan 1982). The response and the publication that followed galvanized research, albeit on weak grounds given the poor controls on sampling and the hasty conclusion that large numbers demonstrate an eastern origin for fluted points (see caveats in Cotter , Griffin , Haynes , and Purdy ). Northeastern researchers were aware of the problems and potentials (Adovasio 1983; Funk 1983; Gramly 1983; Grimes 1983; Moeller 1983; Ritchie 1983), but the tone of Paleoindian research had changed, not entirely for the better, as a strain of fanaticism appeared in academic and avocational communities alike (Table 15). Market forces came into play, inciting increases in both scale and intensity of looting.
Since 1980, Paleoindian research in the Northeast has intensified with contributions from compliance surveys and site evaluations, and more academic attention. Anthropological interpretive issues have been added to the old questions of ‘how old?' and ‘how many?.' The special complications of eastern archeology—small, selective, and scattered samples, competition with looters and market pressures on fluted points, and the predominance of private lands—explain some of the problems that have beset Paleoindian studies. Tankersley (1989:264) highlighted the problems with interpreting private collections. Chronological ambiguities related to radiocarbon dating and poor preservation bedevil everyone (Fiedel 1999; Levine 1990), while the disturbed state of most sites prevents good interdisciplinary and ecological studies (Dincauze 1996; McWeeney 1994). The scarcity of petrographical studies of lithic raw materials complicates and burdens models of mobility. These issues color discussion of the topical categories and research directions that follow.
In this context, it is necessary to keep in mind that Native Americans strongly hold that their ancestors emerged anciently on the continent, not being derived from any elsewhere. Anthropologically, such special creation is difficult to deal with, but the idea is powerfully held.
The oldest site question. The Meadowcroft Rockshelter burst into national awareness in 1975 (Adovasio et al. 1975), with claims for a tightly dug sequence of cultural deposits dating back to 20,000 years ago. At first euphorically welcomed, the claims were soon challenged on several issues. Taphonomic and contamination problems have been addressed in various reports and discussions over the years. As early publications contain some erroneously cited ages that were corrected in 1984 (Adovasio et al. 1984: 355, 357), only selected earlier publications are cited or discussed here, with recent reviews emphasized (Adovasio et al. 1980, 1981, 1982, 1984, 1988, 1990, 1992; Carlisle and Adovasio 1982).
The age of the ‘Miller’ point biface from Meadowcroft Zone IIa is bracketed between two radiocarbon dates that represent a conservative estimate of the radiocarbon age of cultural materials: 12,800±870 (SI-2489) and 11,300±700 (SI-2491) (Adovasio et al. 1988: 48). At face value, this indicates an age for the Miller point between thirteen and fifteen thousand years ago. It implies entry into southwestern Pennsylvania, at the western edge of the Appalachians on the unglaciated Allegheny Plateau, prior to the Younger Dryas while the glacier was shrinking toward the Great Lakes basins and the Ohio and Mississippi rivers were carrying heavy loads of outwash and meltwater.
There are several reasons why these radiocarbon ages are controversial, some technical, others contextual (Dincauze 1981a,b; Haynes 1980, 1991; Mead 1980; Tankersley and Munson 1992). For instance, the Miller point resembles the styles of Early Holocene Group V (Lantz 1984; Table 3), and no closely comparable style has been found to the west, the putative direction of origin (Boldurian 1985). Additionally, the rockshleter is in coal country, and in the past has been suspect for contamination by ancient carbons (Haynes 1980, 1991; Tankesley and Munson 1992). These issues are crucially important for the question of the initial peopling of the Northeast (Adovasio et al. 1992, 1999). The final report is eagerly awaited.
Trails east and north. Because Clovis sites in the West carry ages older than any yet reported in the East for fluted points, the conventional understanding of human settlement in the East is people moving from west to east and then dispersing northward and southward (e.g., Anderson 1990: Figure 4; Dincauze 1993b; Ellis et al. 1998; Funk 1972; Gramly and Funk 1990; Kelly and Todd 1988; Steele et al. 1998; Tankersley 1994c). During the rapid meltback of the Laurentide ice sheet in the Late Glacial warming preceding the Younger Dryas, the Mississippi River could have presented a barrier to west-east movement while it was carrying torrential meltwater with gravel (Kennett and Shackleton 1975). Consequently, it is not irresponsible to posit some delay between the ages of Clovis sites in the west and those to the east (but see Fiedel 1998), despite the closely analogous lithic technology of what seem to be the earliest fluted points in the East (e.g., Morrow 1995). Anderson and Faught (1998) show that the heights of the southern Appalachians were essentially unutilized and may represent another barrier; people apparently tracked around the Blue Ridge, splitting into southern and northern groups there or at the confluence of the Ohio and Tennessee Rivers. Whatever the routes, evidence currently available shows that people were in the East at least as early as the beginning of the Younger Dryas, perhaps sooner.
From the perspective of the Northeast, it is worth considering some alternative scenarios of pioneering movements: (1) from the west into the Allegheny Plateau and southern shores of the Great Lakes, thence along the lake fronts and southern edge of the Adirondacks into northern New England; and (2) north along the Great Valley or Piedmont and Coastal Plain from Virginia. These two directions are not mutually exclusive, but considering them separately illuminates some observations about eastern fluted points that have been puzzling to date (Boisvert 1999; Dincauze and Jacobson 2001). This distinction is becoming possible because of the addition to the data base of the sequential style series of fluted points and more reliable lithic identifications (Table 18).
At least one influx of people directly from the prairie states is strongly supported by the distribution of Gainey (Eastern Clovis) style points and the western lithics with which they appear in the Allegheny River basin of Pennsylvania (Lantz 1984; Lepper and Meltzer 1991). Lantz plots evidence for movement from both the southwest and the northwest into the upper Ohio River drainage. The evidence is, however, complicated by the fact that his fluted point compilation includes all style groups, so cannot represent only early, initial immigrants; two-way traffic is implicated. Gramly (1988b) maps many artifacts of Ohio cherts in the Lake Erie area of New York state, supporting Lantz's inference of movement from the west, as well as some stone that seems to come from as far as North Dakota. Tankersley (1988) also demonstrated some movement from the northwest. Better petrographic data are essential for development of this topic.
The earliest northeastern fluted point sites (Clovis/Gainey style points of Group Ia) occur frequently in the band of country immediately south of the Great Lakes, congruent with assumptions of movement from west to east. The Ohio River route east of the Tennessee River naturally bifurcates at modern Pittsburgh, with the Allegheny leading thence northeastward, the Monongahela leading southeast (Carr et al. 1996). Paleoindians initially following the southeastern route could subsequently have utilized Appalachian passes and moved north along river valleys among the mountains or the Coastal Plain through Delaware and New Jersey. Equally, people entering by the northern routes could have moved south from New York.
In the Mid-Atlantic states the southern routes to the Northeast are not challenged. Gardner and his students see many connections between the Thunderbird sites in Virginia and fluted points in the Pennsylvania valleys and the coastal upland sites of Delaware (Carr 1989, 1998; Custer 1986, 1996; Custer et al. 1983). Such southern affiliations could be evaluated by an intensive examination of fluted point manufacturing styles in southern Pennsylvania, New Jersey, and Delaware (e.g., Callahan 1979; Painter 1973), perhaps also Staten Island. Distributions of lithic types are less useful here, as the numerous potential bedrock sources are themselves difficult to differentiate and subject to over-generalization on that account (Hatch and Miller 1985; Kraft 1977b; La Porta 1994).
These alternative routes bear importantly on issues of chronology, demography, adaptation, and stylistic trends. For instance, the enigmatic Shoop site in Pennsylvania may be (1) a southern outlier of a northern population, established from the northwest, or equally (2) a northern outlier of a southern population, and thus indicative of entry from a second direction. While no critical study of Shoop technology has been done to compare it to assemblages northwestward, the Shoop assemblage is discussed in the literature as stylistically close to early fluted points in Virginia (Group Ib on Table 18; Cox 1986; Gardner and Verry 1979).
During research for his 1984 dissertation, D. Meltzer became aware of north-south differences in Paleoindian tool assemblages. Trying to explain this without benefit of the serial typology, he posited a paleoecological contrast, in terms of tundra environments north of the Wisconsinan end moraines and forest south of them, with corresponding differences in prey. This proposition, although developed later on ecological principles (Meltzer and Smith 1986) was insensitive to the available paleoenvironmental data for the region (see Dincauze 1988). The assemblage differences, however, are likely real (i.e., Group Ia vs. Group Ib fluted points and other details), and may have a socio-cultural basis in subsets of Late Glacial populations. Paleoindian movement early into the Southeast likely resulted in lifeways different from those adapted to the Ohio River/Great Lakes corridor. Colonization movement out of the south, east of the Appalachian peaks, brought northward an adaptive tradition different from that established near the Great Lakes, along with distinguishable lithic techniques and point styles, site location Criteria, group sizes, seasonal activities, and possibly even dialect differences (see chapters and discussions in Anderson and Sassaman 1996c).
The Shoop site near the Susquehanna River in central in Pennsylvania seems less odd when it is reunited with its relatives to the south, rather than being connected to the Onondaga escarpment in New York (Holland and Dincauze 1999; Moeller 1989). Sites in eastern Pennsylvania, most of New Jersey, and all of Delaware (Mid-Atlantic bight drainage) seem to belong to the southern set which is typified by Shoop/Williamson and Thunderbird. The sheer dynamic of northern environments, especially in the Younger Dryas period, was likely an important contributing, if not defining, factor in the differences. Although the Delaware River could have been followed northward from Chesapeake Bay, early Paleoindian fluted points in the Upper Delaware basin, including Zierdt and Shawnee-Minisink, seem at home in the northern set (Ia).
Paleoindians carrying fluted points of Groups Ia and II into the Northeast may have turned north west of the Hudson gorge, following the lake shores with their abundant wildlife (Dincauze and Jacobson 2001). In Vermont, fluted points are notably rare east of the Green Mountains, but numerous near the Champlain Sea (Loring 1980). There is little evidence in southern New England for Group Ia points made of Hudson Valley cherts, which was expected for the model bringing the initial movement from due west. Instead, we find in Group Ia points lithic materials from Maine, Vermont, and northern New Hampshire, and a variety of local and exotic lithics among the Groups II and III points and later styles (Jacobson 2001 and diverse sources). The Wapanucket 8 site in southeastern Massachusetts has Barnes points (Group III) with Hudson Valley cherts, while the Munsungun outcrops are represented at the earlier Bull Brook site (Group Ia) along with stone types originating in Vermont (Curran 1999). The Neponset site (MA) has Munsungun chert and Mt. Jasper rhyolite with Group III points. The situation at DEDIC/Sugarloaf (Gramly 1998) is too confused to support interpretation; reports of lithic materials conflict from observer to observer.
The history and origins of Paleoindian Group II points is unexamined. They are conventionally associated with the extreme northeastern distributions of fluted points, in Maine (Vail, Adkins) and Nova Scotia (Debert and others). More recently their deeply indented bases are recognized south of the Great Lakes at the Lamb site (Gramly 1997), as far as western Pennsylvania in Lantz's sample (1984), and east to Martha's Vineyard. Did they come into the Northeast on the heels of the Gainey type and end up on the northern margin? If the style developed originally in northern Maine and Nova Scotia, as the use of regionally specific lithics implies, how do we explain the wide distribution of strays? Radiocarbon dates place them among the earliest points in the Northeast (Levine 1990: Table 1), but their relationship to the development of the Younger Dryas cold period remains unclear. The rare, small, triangular late fluted points of the eastern Maritime Provinces and northern New England, with their deep basal indentations, seem to derive from this style (Keenlyside 1985).
From the time of Group III bifaces in the Northeast, the patterns of site distributions and lithic utilization indicate some territorial restrictions on the populations (Jacobson 2001: ‘settling in'). The distribution extends into Ontario and the Midwestern Cumberland style of the same relative age. Group IV points are less widely distributed, centered near the eastern Great Lakes and St. Lawrence, into Maine. As strays they show up in New Jersey and coastal New York. Lanceolate bifaces of Groups V and VI are not typically made of exotic raw materials, although there clearly continued to be selection for excellent flaking properties; dark to black cherts were notably favored, just the rocks that are notoriously difficult to trace to ground.
There is no evidence supporting the long-held notion that Paleoindians moved north out of New England to follow the glacial margin. North of the St. Lawrence River, there are no fluted points (Keenlyside 1985: Figure 1). Late Paleoindian tool kits of recognizable New England styles do not spread northward (Keenlyside 1985); they do not appear in higher latitudes even with a migration lag (Dumais 2000; Tuck 1984).
The lack of such evidence northward, therefore, makes enigmatic the change of biface styles that marks the beginning of the subsequent Early Archaic period. The earliest Early Archaic tool styles—the notched Hardaway-Daltons, Palmers, and Kirks—seem, like the first fluted point styles, to be intrusive into the Northeast, this time definitely from the south. There are no antecedents to these styles north of West Virginia, and no unique regional point styles in the eleventh millennium of the Northeast. The clearly southern suite of Early Holocene styles is also discontinuously distributed; more numerous in the south, rare to vanishing in the north (Robinson et al. 1992). Until investigators in the Northeast were forced to recognize the southern types and understand their distant origins, they depended on an environmentally determined depopulation hypothesis to account for a ‘gap' in the human presence in the Northeast after the time of fluted points (Ritchie and Funk 1971:56; see Dincauze and Mulholland  for a discussion of this). Dumont (1981) posited a continuum from Paleoindians on the basis of ecology and the bare outline of settlement patterning, as did Dincauze (1990), using as Criteria significant continuity in selective uniface tools from the Paleoindian to Early Archaic periods; only the bifaces change.
The origins and connections of human societies are not equivalent to those of projectile points. B. Robinson surprised New England archeologists by demonstrating Early Archaic cultural assemblages near the Gulf of Maine that lack bifacial stone weapon tips and are thus incomparable to Early Archaic complexes in southern New England (Robinson 1992). Northern New England Early Archaic cultural remains show as quartz debitage and ground stone rods, themselves not traceable far to the south. Relationships with southern-derived bifaces are therefore unclear.
The highly visible Bifurcate-base points, horizon styles of the early Holocene, are attributed variously to either the Early or Middle Archaic, depending on latitude. Their ages seem to decrease to the north, where the styles diverge from those of the southeast. Because they date younger than 10,000 years ago, too late to contribute substantially to issues of interest here, they are not considered further.
In the Northeast, the problem of explaining the distribution and origins of Iroquoian and Algonquian speaking peoples and their purportedly ‘distinct' cultures has exercised anthropologists for generations, and continues to do so today. Consequently, the region offers no models or analogs helpful for understanding or testing processes of cultural replacement or socio-cultural blending or separation (Dincauze and Hasenstab 1989; Snow 1992). Indeed, the North American continent is similarly bereft of explanatory models (Hall 1995) beyond the Native American origin tales.
Thinking about the formation of social institutions should be easier in the vacuum of Paleoindian time than in the more complex case of the exotic Early Archaic material culture, but it is no simple challenge. Late Glacial oscillations between cold and warm periods, and the Pleistocene-Holocene environmental transition were apparently very rapid as measured in the real time of Greenland ice layers (Mayewski et al. 1993; Taylor et al. 1993, 1997). While pollen sequences do not express rapid change well, they do show us highly dynamic biotic mosaic environments responding to the far more rapid climatic change (Gaudreau and Webb 1985; Jacobson et al. 1987; Peteet et al 1993), and macrofossils reveal more of the details (FAUNMAP Working Group 1996; McWeeney 1994; Miller and Thompson 1979; Morgan 1987). Hardwood mast species were established in New England as early as Group III fluted points and continued to increase in density and diversity. What kinds of social institutions can we expect in such circumstances? How do we understand change in such situations among small, mobile, human societies (see MacDonald  and Wright  with references cited there)? Early models of seasonal cycling relied on inadequate paleoenvironmental reconstructions and are now mostly discredited, along with the conclusions drawn about settlement sizes, durations, and functions. The archeological record for this time remains impoverished in the Northeast, where organic remains are rare and the lithic assemblages—scattered, collected and looted, sold, illustrated, and speculated about—are the greater part of the evidence.
Anthropological kinds of questions about social change should be based on demographic data at minimum (Laughlin and Harper 1988; Straus et al. 1996; Wright 1989:348), and reasonably good control of time. Notwithstanding, issues of coresident group size, frequency of moves, openness of social groups, economic specializations, gender roles, and technological innovation have been raised for the Paleoindian Northeast (Chilton 1994; Curran 1984; Curran and Grimes 1989; Custer 1996; Custer and Stewart 1990; Dincauze 1993b; Dincauze and Curran 1983; Eisenberg 1978; Fitting 1977; Funk 1991; Gramly 1988b; Gramly and Funk 1990; Grimes and Grimes 1985; Kuhn 1994; Lothrop 1989; Meltzer 1984a; Moeller 1984a; Spiess 1984; Spiess and Wilson 1987; among others). Several authors in the Ellis and Lothrop collection (1989) use exotic lithics as evidence for, variously, exchange, social signaling and banking, marriage networks, and residential mobility. Postulated differences between collectors and foragers are explored as controls on movement. The conclusions, elusive at best, must be pursued.
In the Northeast, shrinkage of the resident populations apparently occurred in the Early Holocene, between the end of fluted point manufacturing and the establishment in the Northeast of the Bifurcate base point style after 11,000 years ago. Paleoindian biface styles of Groups IV, V, and VI are as rare as the coeval Early Archaic notched points. This observation, and the dynamics of replacement or population regrowth, are not yet firmly established as research topics in the region. This neglect is crucial, as our concepts of relative or absolute population sizes are partly artifacts of radiocarbon warping of time—the Late Pleistocene and Early Holocene chronological plateaus (Fiedel 1999). The long spans of real time constricted in the radiocarbon envelopes of the twelfth and fourteenth millennia encourage us to imagine unrealistically greater population densities than were in fact the case. When we open the envelopes to deal in calendar time, the apparent populations shrink severely.
The reigning assumptions that enable speculations and arguments about social groupings are that artifact styles and lithic raw material classes reflect ideas and knowledge shared within a bounded social group. These concepts, well grounded in ethnology and social anthropology, lack analogical support for thin, mobile populations such as those of Late Glacial North America. We should be prepared for surprises here. In the meantime, biface styles, the only clues we have, are utilized as talismans of social groups. We classify them into discrete time units, despite observing that styles intergrade, changing as morphing forms. Gross point form and lithic material suites have been used by Curran (1999), Custer and Stewart (1990), Gramly (1988), Grimes et al. (1984) and Spiess and Wilson (1987) to define archeological ‘phases' which are also conceived as territorial ranges, as is customary for later time periods in the archeological record. Even if the assumptions are correct, the social meaning of style change frequency and distributions is probably not constant in differing demographic densities and various times.
The practice of lumping all fluted points as ‘Clovis' obscured until recently the diversity in point forms and distributions. If such differences do have social correlates, one can play with interesting patterns. The fundamental one is whether continuous or intermittent stylistic variation occurred, or whether, indeed, major changes represent separate group immigrations into the Northeast. For example, was the Debert style (Group II) a native development from Gainey in the far Northeast, or brought separately from the West, with its share of exotic lithics, into western Pennsylvania and New York? Similar questions arise at the Paleoindian-Early Archaic transitions in the region.
Custer et al. (1983) present lithic procurement as a control on site distributions, using the cyclical model of procurement developed by Gardner in Virginia. Custer found it necessary to posit an alternative, "serial," system for upland and inland areas of the Middle Atlantic States, where dispersed outcrops are readily available during the annual round. It is not uncommon to find archeologists working with lithic distributions in a social-system mode to overlook diversity in artifact styles within ‘social' spaces defined by stone types.
Issues of settlement mobility and seasonal mobility were raised early in the Northeast, with the publication of the Debert site (MacDonald 1968), and have received continuing attention. They are prominent in all efforts to understand the unique size of the Bull Brook site. These large sites have been modeled as seasonal destinations used repeatedly by small groups returning over a period of years (Curran and Grimes 1989), or as special-purpose aggregations for (1) intercepting game, (2) seeking mates and information, or (3) scouting newly encountered lands (Curran 1999; Dincauze 1993b; D. MacDonald 1997; G. MacDonald 1968; Spiess 1984; Sullivan 1992; Tankersley 1995). Spiess and Wilson (1987) make a good case for more than one visiting group at Michaud, but can't tell whether they were there together or separately. Behind the inconclusiveness of these models is the assumption of fixed ranges. If we conceive of ranges as social arrangements that developed in place over time, the meaning of aggregations will change. Our task is to know when and where.
From the perspective of a resident of New England, the transition zones that reflect distributions of culturally contrastive assemblages are eerily familiar. The differences in Paleoindian assemblages of the twelfth and thirteenth millennia between northern and southern New England, and between the northern Northeast and southern Pennsylvania, are paralleled in later prehistoric as well as modern social affiliations and even in the citing circles of Northeastern archeologists. How much the modern phenomena affect our perceptions of the ancient ones demands examination.
Several regional authors agree that the Early Archaic period shows little change, socially and culturally, from Paleoindian (e.g., Dincauze  groups both together; Carr 1998:48; Custer 1996: "Hunter-Gatherer I period"; etc.). This correspondence derives from the fact that sites with the early notched bifaces are small, scattered, rare, and not infrequently in the same places as Paleoindian sites, as well as sharing many lithic tool types except for the bifaces (e.g., Logan, Plenge, Reagen, Twin Fields,). This consensus is subject to change whenever archeologists agree about recognizing Early Archaic artifacts, pace Funk (1996). In the Northeast, the similarities end in the period of Bifurcate base points, when subregional diversification of Archaic cultures gains speed. Quarry sites are not important in the Early Archaic period north of Pennsylvania, where local rocks from rivers, ground moraine, and bedrock were used without much discrimination. No long-distance transport of lithic materials has been demonstrated for Early Archaic. In that respect, the societies seem more settled than were the Paleoindians. More attention should be paid to finding and examining Early Archaic sites as if they mattered.
Expressing Cultural Values
There is very little among the Paleoindian sites of the Northeast that can be considered evidence of cultural value systems. Clearly, there was culture: an effective set of shared behaviors to maintain communication, nurturance, physical reproduction, and social groupings in the region for nearly two millennia in very challenging times of environmental lability. We stand in awe of the achievements. But of ephemeral things like values, we have little in hand. The typological standardization of the fluted bifaces and later lanceolates suggests to anthropologists that people used them not only as tools but also as social symbols (Wobst 1977). Nevertheless, the broad spatial scale at which early styles were distributed cannot indicate a very intimate level of symbolization, either to members of groups or to practitioners of particular economies. The spatial scale is comparable to those of language groups today. The earliest Clovis-like styles, defined closely by southwestern Criteria, have a continental scale of distribution. Second-tier style units are at best regional and sub-regional in scale, as the Gainey/Bull Brook, Shoop, Debert/Vail, Barnes and later northeastern types show.
Some artifacts that may have carried symbolic meanings significant to Paleoindian observers are miniature fluted points that accompany fluted point assemblages (Ellis 1994), enigmatic and possibly spurious talc pendants reputedly from the Reagen site in Vermont (Ritchie 1957), and stone beads from the DEDIC/Sugarloaf and Hiscock sites, probably of the thirteenth millennium (Gramly 1998:27; Laub 1995a:28). Imputing original cultural meanings to these ancient, rare and scattered items can be no more than unbounded speculation. For example, the miniature fluted points have been interpreted variously as children's toys and shamans' paraphernalia—very different scales of signification. In the study area, they occur at the Arc, Bull Brook, Neponset, Port Mobil, Templeton, and Vail sites, and at sites in Ontario and Nova Scotia, as well as from Clovis and Folsom sites in the West (Storck 1991:157). The doughnut-shaped stone beads, with no known correlates elsewhere, have only context-free, intrasite proveniences, precluding interpretation.
No Paleoindian burials are known in northeastern United States, nor are art forms that communicate to us. The scratched rocks mentioned in the Property Types section are not obviously representational. In the absence of cave utilization, parietal art cannot be expected to survive. Whether lithic materials as such were considered representative of their places of origin, as is the case among Australians (Gould 1980), is unknown and so far untestable. The visual beauty of some fluted points made of exotic stones that must have been carefully selected, for example the Lamb site bifaces and the famous Intervale point from New Hampshire, strongly implies some meaning less mundane than tools for the hunt. Whether social status, magical significance, aesthetic delight, or genealogical claims invested them with meaning, we have at the moment no incisive clues (Ellis 1989).
In Ontario, the Crowfield site appears to have been a cremation burial with burned offerings (Deller and Ellis 1984). In this context, then, it may be significant that an Early Archaic cremation burial has been reported in New Jersey, with Kirk Corner Notched points (Stanzeski 1998). In New England, the first known mortuary sites are dated within the eighth millennium, outside the range of this study (Doucette and Cross 1998; Robinson 1992).
Shaping the Political Landscape
"No one has succeeded in devising an acceptable direct method for assessing population variability for Pleistocene foraging groups, knowledge of which is critical for assessing questions of alliance and communication" (Wright 1989: 348). Wright continues by offering suggestions for improving this situation, which are daunting and not yet feasible. We must open by restating that political units develop only after the achievement of a degree of sedentism; in the particular case, that would have been during and after the time of Group III points, at earliest. We can take comfort in noticing that by that time we can demonstrate the existence of sub-regional styles with increasing numbers and decreasing sizes of stylistic provinces. So, it is likely that political landscapes were taking form in the twelfth millennium; research to define them awaits.
If weapon tips are considered icons of social or political affiliation (Wobst 1977), then the several style groups discussed above may be tentatively considered elements of successive political landscapes, alternative to the social groups discussed above in Creation of Social Institutions. At any one time within the Late Glacial Northeast there were at most two or three such. Finer subdivisions have been posited by Curran (1999), Custer (1996), Gramly (1988), and Spiess and Wilson (1987), on the basis of equivalences in biface forms and types of stone utilized for tools. These groupings are smaller than style provinces and geographically more realistic as territories or ranges for human communities. However, there are problems: (1) some of the spatially defined groups are based on sites belonging to more than one time period or style cluster, and (2) the size and borders of the units, as defined, must be revised every time a new technique is applied to the identification and sourcing of utilized rocks.
Large-scale patterns within the regional set of sites may indicate territorial affiliations more fundamental and, at the same time, short-lived. If the occupants of the Shoop site in central Pennsylvania can be absolved from the necessity of obtaining Onondaga cherts from the bedrock outcrops of western New York, and be allowed to collect from nearby outwash gravels of the Susquehanna River (Holland and Dincauze 1999), then the southern affiliation of the point style can be taken seriously (see above, Group Ib). In such case, Shoop may lie near the northern limit of immigration from the south within the chain of the folded Appalachians, and sites to the north and west of that can be seen as the southern limit of immigrants from those directions. Pennsylvania may, therefore, be the location of an important political/demographic watershed between the Northeast and the Southeast, as discussed above. This notion is supported by Lantz's observation in western Pennsylvania that despite a very eclectic mix of lithic materials in the area, some from great distances away, "Almost none of the available lithic sources in southwestern Pennsylvania were finding their way up to the northwestern counties" (1984:213). A suite of subtle behavioral traits may be interpreted the same way; these lie behind Meltzer's recognition of differences north and south that he explained by positing a major environmental boundary (Meltzer 1984a).
Early in Paleoindian times, contrasts in social and stylistic entities east and west of the Appalachians may be explained by migration models, not fundamentally political. Increasingly thereafter, differences emerge on opposite sides of the Appalachian mountains that set the stage for normalcy in the Archaic periods. The first Early Archaic artifact styles in the Northeast patently arrive, whether from hand-to-hand exchange or by pedestrian traffic, from the south. By the end of the Early Archaic period, Bifurcate base points had spilled from the Tennessee Valley over the mountains north of the Blue Ridge, and Dalton-related side-notched points had arrived from the mid-continent into Vermont (Thomas and Robinson 1983). Thereafter, socio-cultural and economic contacts occurred within the Northeast and from the west, south, and north.
Politically, both Maine and Pennsylvania appear as important border zones in Paleoindian and Early Archaic times. A strong cultural demarcation between southern-derived and northern-derived cultural groups crosses Maine throughout most of prehistory. Northern Maine, arguably less heavily surveyed than any comparably-sized area in the region, nevertheless shows sites of the fourteenth millennium with Gainey and Debert-style points (Groups Ia and II), and a Maine presence to the south in numerous sites exhibiting Munsungun cherts (Pollock et al. 1999). These recent realizations premise widespread revision in the region. The area seems to have been more lightly utilized in the early twelfth millennium than it was either earlier or later. Group III points do not occur in the Maritimes or adjacent northeast Maine, an observation delayed by the famous Intervale, NH, point (Boisvert 1999). Groups IV and V are rare everywhere, so that their relative frequency in Maine cannot be closely interpreted, except to note that they are strongest in the southwest. Group VI points are increasingly reported in Maine, while their numbers to the south are augmenting much more slowly. By Early Archaic time a divide lay again in central Maine (Dincauze and Mulholland 1977; Robinson et al. 1992), although it was even then a permeable border with lanceolates coming south, notched points going north, as strays.
Far more precise data on artifact traits, lithic sources, site distributions, and chronologies of site occupation and resource use are needed before any of these arguments can be carried much further (cf. Dincauze 1991:6). Imputation to ancient communities of territoriality and political boundaries carries heavy freight in terms of social organization and human interaction—too heavy to be cavalierly imposed on scant data. Research into these issues is important and, with the fragility of private collections, urgent.
Developing the American Economy
Human beings have a fundamental set of economic needs: food, raw materials, finished goods, shelter. Shelter has last place here because for short times and in special cases, humans are ingenious enough to get along with minimal formal shelter. The absence of shelter evidence in the Paleoindian record should not be taken as evidence of absence; it is more likely to be a product of loss of site integrity in the long time spans between deposition and recovery. Investigation of Paleoindian economies in the Northeast has been devoted mainly to the search and utilization of raw materials for tools, the best preserved set of economic evidence. Tankersley and Isaac (1990) lay out a well conceived research strategy for Paleoindian economies, based on descriptions of production, consumption, distribution, and exchange. However, analyses such as they recommend require a data base with more integrity than is now available in the Northeast, for both sites and lithic sources.
Tankersley and Isaac (1990) argue for reducing the importance given to paleoenvironmental modeling in Paleoindian economic studies. However, they go on to praise Weissner's (1982) essay on risk management as a complement to acquisition strategies. Paleoenvironments set the contexts within which risks are expressed and assessed—scarcity or plenty, stability or instability are crucial elements in strategizing; sound paleoenvironmental modeling is essential to economic studies. Curran and Grimes (1989), Meltzer (1984a), and Meltzer and Smith (1986) have produced some of the most ambitious efforts at economic modeling south of the Canadian border, but all err when modeling seasonality and game populations (Custer and Stewart 1990; Dincauze 1996).
Discussions of food selection and procurement are inseparable from models of settlement patterns, which, in the Northeast, are founded on lithic procurement. Full consideration is delayed to below. Evidence for food choices is elusive. Organic remains have been recovered at the Shawnee-Minisink, Bull Brook, Nevers, Hiscock, Whipple, Neponset, and Templeton sites, mainly as fragments of calcined bone (Spiess et al. 1985). The faunal richness of the Hiscock site cannot be interpreted to represent human collection (Laub 1995b). Seeds at Shawnee-Minisink and Meadowcroft may be either human food remains or insect caches. Organics are even rarer at Early Archaic sites. Pollen studies are usually done off-site, with dating essential for relating humans and vegetation.
The idea that Paleoindians predominantly hunted and ate Big Game is fading from northeastern archeology (Custer and Stewart 1990; Dincauze 1981c; Meltzer and Smith 1986). There is no denying that caribou were present, but not in huge herds (Spiess is preparing a refutation of this idea). Caribou are not larger than other cervids that were present in the forest edges: moose, elk. No doubt cervids were taken when offered, but to posit specialization on such solitary and moderately sized game is to distort available data without justification. Small animals were also important (e.g., beaver at Bull Brook, fish at Shawnee-Minisink), and edible plants in season were crucial to maintaining healthy human body functions. Large-scale processing of seed plants is unlikely in the absence of heavy tools for such activity. In general, Paleoindians can be seen as living fairly high on the food pyramid, where predation comes into its own.
Wear studies on endscrapers show both woodworking and hide working functions, and possibly work on bone and ivory (Funk 1976:214-215). Division of labor is fundamentally an economic question: the frequency of scrapers in sites has been proposed as an indicator of women's activities (Chilton 1994). Wooden hafts, containers, house poles, and perhaps boats (Engelbrecht and Seyfert 1994) were essential goods. Wood, bone and ivory small hand tools could fill several essential functions from sewing to trapping. Prepared hides and skins were needed for clothing, houses, and boats. The discovery of twine and net impressions in European Upper Paleolithic sites (Adovasio et al. 1996; Pringle 1997) reminds us that plant fiber industries contribute to life support, and were likely used in the northeastern US as well. The supposition is supported by finds at Fishbone Cave in Nevada, associated with radiocarbon ages in the late Pleistocene (Orr 1956, 1965).
The economics of lithic material procurement dominates the Paleoindian literature to a degree that must over-represent its prehistoric importance. The sheer diversity of lithics in most northeastern sites belies our assumptions about the difficulty of acquiring suitable materials. Nevertheless, lithics are what we have dependably to work with, so their importance is central.
Bedrock sources of the fine-grained siliceous materials favored for fluted point manufacture are not ubiquitous in the Northeast, but they are massive and were apparently discovered early. Gainey points arrived as finished goods made from Midwestern cherts of very high quality (Dincauze 1993b; Gramly 1988b; Lantz 1984). People who moved into New York state utilized the Onondaga outcrops in the west (Ennis et al. 1995; Tankersley et al. 1996) as well as the Normanskill/Deepkill outcrops west of the Hudson River trench (Funk 1973; La Porta 1994). Moving north into Vermont along the Hudson-Champlain lowlands, they could select from Fort Ann and Mt. Independence cherts (Wray 1948; Haviland and Power 1981) and outcrops of Normanskill-like stone of the Mount Merino formation near the New York/Vermont border (Brumbach 1987). In the Green Mountains they found Cheshire quartzite, Colchester jasper, and St. Albans/Hathaway cherts (Haviland and Power 1981; Loring 1980). Farther east they encountered siliceous rhyolites in New Hampshire and the extensive colored beds of Munsungun cherts in Maine (Boisvert 1992; Pollock et al. 1996; Pollock et al. 1999). Those who went into Pennsylvania and New Jersey had bedrock sources of dark cherts in the Kittatinny, Helderberg, and Kalkberg mountains in the east (Kraft 1977b; La Porta 1994), and various jaspers in central and eastern Pennsylvania (Hatch 1994; Hatch and Maxham 1995; Hatch and Miller 1985). Bedrock jasper has been located in Rhode Island (Waller 1999) and suspected in the Berkshire hills in western Massachusetts and Connecticut (Moeller 1984a). Chalcedonies were available in northern Delaware (Custer 1984b; Griffith 1982). Materials in the western and northern Appalachian mountains and the Allegheny Plateau occur in various Paleozoic marine formations. More attention is needed to learn whether Paleoindians opened quarries at these sources. Most source attributions are to formation only, but with intensified fieldwork, Paleoindian quarry sites are now recognized in New York, Pennsylvania, and New Jersey (Funk 1973; Hatch and Maxham 1995; P. La Porta, pers. comm.).
Non-cherty bedrock with good flaking qualities occurs in the fine-grained Paleozoic volcanics (rhyolites and metavolcanic tuffs and silicified siltstones) of the Green and White Mountains, the Boston and Narragansett Basins, and locally in the Mesozoic formations of central Massachusetts and Connecticut (e.g., Calogero and Philpotts 1995; Gramly 1984b; Jones 1997; Luedtke 1993; Pollock et al. 1996). Quartz crystal in pegmatite dikes was used, as was the very pure quartzites of the Green Mountain and Berkshire ranges (Calogero and Philpotts 1995; Dickson 1967; Loring 1980; Haviland and Power 1981:29; Spiess et al. 1998:Table 7). Delaware chalcedonies were important (Custer 1984b; Griffith in Brennan 1982).
Long distance transport of lithic raw materials, widely acknowledged in the region, is now under revision as new studies of lithic sources refine attributions (Calogero and Philpotts 1995; Hatch and Maxham 1995; Hatch and Miller 1985; King et al. 1997; La Porta 1994; Luedtke 1987; Pollock et al. 1999; Spiess et al. 1998). As indicated above (Property Types), these will revolutionize the picture of migration into the region and subsequent patterning of settlement and distribution. For now, patterns of distribution of raw materials and finished goods among Paleoindian societies are matters of controversy (see especially papers in Ellis and Lothrop  and Tankersley and Isaac ). Debate typically takes the form of dichotomous categories: direct vs. indirect procurement, cyclical vs. serial quarry visits, embedded vs. logistical procurement, acquisition of raw material vs. finished goods. Opinion is moving strongly away from trade as a mechanism for the distribution of lithic materials among Paleoindians. Meltzer (1989), arguing that not all transport is exchange, concluded that exchange would have been unreliable among populations so thin on the land. Curran and Grimes (1989), Moeller (1984a), and Spiess and Wilson (1987) also dismiss exchange, making exceptions for rare exotics. No analyst has been able to discern distance-decay patterns in the Northeast, in contrast to what Anderson (this volume) reports in the Southeast. Lothrop (1989) argued for exotic lithics representing band member fluidity. In this context it is relevant to note that finished goods, purposely modified for transport, may travel separately from quarried raw materials, being distributed as biface preforms or finished tools. At the Dam site in Maine, represented lithic materials differed between tools and flakes (Spiess et al. 1998:214; Wilson and Spiess 1989). It is also important that utilization patterns varied in time. Gainey style points were frequently made of exotic stones, while Barnes and later styles tend to appear in local materials (Jacobson 2001). Munsungun chert is a notable exception, appearing in New England and to the south in style Groups I, II, and III. Black cherts, whose provenience is rarely established, seem to have been favored especially for the Paleoindian lanceolate points of Group VI.
No intensive study has been made of the economic roles of lithic tools. Lothrop's analysis of the Potts assemblage (1989) is a first effort at interpreting tool kit organization, building on good work to the west. Use and discard studies are rarely undertaken and even more rarely published, and use-wear analyses are rare (see Grimes and Grimes  on limaces for the potential of these). A recurrent problem in comparative studies of northeastern Paleoindian assemblages is the lack of standard analytical or descriptive vocabularies (Curran 1984). This problem is exacerbated by sub-regional typologies for cultural periods that fragment archeological communication in the region.
Paleoindian settlement patterns are not well understood; in fact, they are controversial. Settlement patterning is the basis for understanding mobility and acquisition of essential goods, a first step toward modeling economies. Paleoindian settlement pattern proposals in the Northeast, when not simply transferred analogies (Levine 1997), are based either on lithic acquisition or predation pattern models; most include speculations about seasonal mobility (e.g., Carr 1998; Curran and Grimes 1989; Custer 1986; Custer et al. 1983; Eisenberg 1978: 138-9; Gramly 1988b; Grimes et al. 1984; Meltzer 1984b; Spiess et al. 1998; Spiess and Wilson 1987). In the southern parts of the region, Gardner's Thunderbird model of lithic tethering and cyclical settlement moves has been influential. "The use of a staged biface reduction lithic technology and a curated tool assemblage are also significant characteristics of this adaptation" (Carr 1998:48). For areas north of Maryland, this model does not work. In the course of sequential settlement moves emphasizing locations optimal for acquiring foods (wetlands and river terraces), Paleoindians used many small quarries or other lithic sources (e.g., McCracken 1986; Wilson and Spiess 1989). Base camps are not preferentially located close to quarries, and the distribution of a single lithic type does not trace a group's annual round. Several types of stone are typically found together on sites. Efforts to define band territories stumble on this difficulty. More thought and evidence must be devoted to understanding Paleoindian settlement patterns in the Northeast; success requires control of both lithic sourcing and time, which are only now coming within our grasp.
Too few Early Archaic sites are known or excavated to support regional-scale study of economies (Bolian 1980; Carr 1998; Dincauze and Mulholland 1977; Funk 1996; Michels and Smith 1967; Moeller 1985; Robinson and Petersen 1993; Spiess et al. 1983; Turnbaugh 1980). The sites are everywhere small, and the tool classes include uniface styles carried over from Paleoindian. Organic materials in Early Archaic sites are almost unknown, leaving diets unreported. By the Early Archaic period utilized lithic raw materials are typically very local throughout the region.
Expanding Science and Technology
The exploration of a continent has been extolled for centuries as a heroic undertaking, even though the mere fact of extolling indicates that it is done by people from literate civilizations, most of whom had native guides. Prehistorically, it was truly exploration, and truly historic, as the explorers were occasionally entering areas without any human presence; this is comparable to a moon walk, or settling islands in the Pacific Ocean, relying entirely on one's own collective ingenuity. Paleoindians, as survivors, showed themselves equal to the challenge. They created mental maps and gazetteers for naming and organizing knowledge about landforms, sources of useful lithics, vegetation, and fauna. Although Beringian fauna were familiar wherever encountered, most of the biota were unprecedented and needed names and behavioral study. The biota changed in both time and space at short durations, because of rapid climate change and topographical diversity. Paleoindians established a cultural geography while themselves dealing with the consequences of rapid climate change. That's science. The absence of permanent records doomed these achievements to anonymity.
Tool kits and technological organization have been addressed only recently and lightly in the Northeast (Dincauze  reviewing papers in Ellis and Lothrop ). Tool kit organization is addressed by Lothrop (1989) at Potts, building on the work of Bamforth (1986), Ellis (1984, 1989), Knudson (1983) and Shott (1989b, c). The critical variables in this aspect of economics are not yet understood, with the result that there is little control on speculative scenarios. The growing literature on mobility scheduling and toolkit diversity (Kuhn 1994) has been used in the Northeast by Curran and Grimes (1989), Custer and Stewart (1990), Ellis et al. (1998), and Spiess et al. (1998).
The use and preparation of fluted points has been addressed from Ontario, where the sequence of biface styles appears "to represent temporal markers in a fairly continuous and incremental evolution of Palaeoindian point styles in the region" (Deller 1989:193). Ellis and Deller (1997) see styles varying with changes in the environment and economy, linking them through changing availability of prey species and hunting fields. Other more cultural reasons have not been explored, although Lepper's argument for multiple uses for fluted points (Lepper 1986c) could be embellished with a model of change over time. Tankersley (1994a) considers the effects of raw material and technology on point form. His argument supports conscious technological capabilities and effective heuristics in Paleoindian tool manufacture and use, which should be more widely acknowledged.
Ellis and Deller (1997:21) have noted that "pièces esquillées, fluted/twist drills, and limaces/flake shavers are very common in reported assemblages from the New England sites… these types are rare to nonexistent in the eastern Great Lakes." Shott, working mostly in the Midwest, is on the frontier of use and curation studies for Paleoindian tool kit elements, with studies of pièces esquillées and scrapers (1989a,b,c; 1990; 1995). Shott (1989a) presented an ethnographic survey supporting the origin of pièces esquillées as exhausted cores in situations of raw material shortage. A detailed study of pièces esquillées at the Vail site by Lothrop and Gramly (1982), with a review from a global perspective, concluded that they were not cores for usable flakes, that they work as wedges for splitting bone and wood (see also Goodyear 1993), but that a unique functional interpretation is not possible. Scrapers (Shott 1995a), twist drills (Jordan 1960), limaces (Grimes and Grimes 1985), and flake tools with denticules (Lothrop and Gramly 1982: note 1; Tomenchuk and Storck 1997) have been given close attention in the Northeast. This clearly productive line of research should be pursued; it demands exhaustive collection during excavation at discrete sites.
Selected Paleoindian tool kit elements continue in use after the end of fluted point making. Some forms of unifacial scrapers, and pièces esquillées, continue with lanceolate points and endure into Early Archaic assemblages. This observation does not contribute to the debate about evolution in place or renewed immigration to explain the changes that demand recognition with a new cultural period. Lanceolate point use may develop in place, later than in the Southeast, although the styles in the Northeast relate to large Dalton-affiliated bifaces in the southern Prairies and lanceolate styles in the Great Lakes. Early Archaic notched bifaces are clearly introduced into the region, directly or indirectly, from the south and west (Ellis et al. 1998; see citations to review articles for Early Archaic). The rarity of such styles in coastal Maine and eastern Canada is an enigma; presently implying a cultural frontier to the north.
Transforming the Environment
Paleoindians entering the Northeast from the Plains confronted daunting topographical contrasts and unfamiliar plants and animals. From the southeast, the terrain was not more diverse, but the biota would have been increasingly different with latitude. Reliable paleoenvironmental and paleoecological reconstructions have appeared only recently; in the 1970s researchers depended upon analogues of climate change borrowed from western Europe, and descriptions of Contact period vegetation (Funk 1972). As a consequence, debates about paleoecology and adaptations were initially misleading; efforts to import a Big Game tundra-hunting model to the forested, mountainous Northeast were accepted (Funk 1972; Kelly and Todd 1988; Meltzer 1984a). Unsuitable ethnographic analogies were chosen from the Barren Grounds and Boreal Forest peoples of the nineteenth century (see critiques by Dincauze  and Levine ).
Well informed current paleoenvironmental modeling is essential to the task, but still premature as information on dating and climate is changing radically (e.g, Clark et al. 2001; Dincauze 1996). Only generalities can be limned at present; models of Paleoindian ecology and behavior must be provisional, and should be specific to special types of landscape throughout the Northeast, with an eye to differences in climatic expression by latitude and altitude. Alert to the influence of topography on climates at human scales, Gaudreau (1988:245) warned that lack of fine-scale resolutions in time and space prevent any observation of human impacts on vegetation in the Paleoindian time period. Given the dry cold of the Younger Dryas period, human fires likely increased the rate of natural fire stress on both vegetation and fauna at the time.
Humans entering the Northeast during the retreat of Wisconsin glaciers (Dyke and Prest 1987a) represented to the resident fauna a major new predator and competitor. Acrimonious debates about generalist vs. specialist resource strategies by Paleoindians enlivened the literature in the 1970s and into the late 1980s, before a consensus emerged that in rapidly changing, diverse and challenging environments, one ate whatever was available (Curran 1987; Dent and Kaufman 1985; Dincauze 1981c, 1993a; Eisenberg 1978; Gramly and Funk 1990; Meltzer 1988; and others). To this extent we can confidently say that environment affected culture.
Travel itself was likely affected by the extraordinary climates of the Younger Dryas and melting glaciers. The trail along the southern shores of the Great Lakes, with those at either high or low water stages, would have been facilitated by the use of boats (Engelbrecht and Seyfert 1994). Boats may not have been enough to conquer the challenges of the Hudson Trench, which was apparently a barrier for the earliest immigrants carrying Gainey and Debert style points (see Peopling Places). Crossing the unfrozen Hudson posed no problem for later Paleoindians with Group III points, when Hudson Valley cherts were moved east in significant amounts.
What effect did the Paleoindians have upon the environment? Within the Northeast there is little evidence that humans contributed to the extinctions of the Pleistocene megafauna that were already well under way by the time hunters appeared (Meltzer and Mead 1983, 1985). However, it is unlike humans to leave any area on the face of the earth unmodified by their presence. Nicholas (1999) reviews the kinds and degrees of effects that low density, highly mobile human populations can have on patchy environments, and concludes for a "light footprint." Site data for Paleoindians so far confirms a low visibility of impact at most (McWeeney 1994). At the Hedden site in central Maine, artifacts lie under dunes on a buried soil containing evidence of burning (Spiess and Mosher 1994; Spiess et al. 1995). Anthropogenic burning is difficult to distinguish from natural fires on droughty sites, which Hedden clearly was (Patterson and Sassaman 1988), particularly when only one site is at issue. The rapid environmental perturbations attendant on the Younger Dryas period create a probably impenetrable ‘noise' complicating recognition of anthropogenic factors.
Modification of vegetation succession by both intentional burning and escaped campfires is to be expected wherever human groups gather. Activities such as fishing, hunting and trapping, collecting and transporting and quarrying, change landscapes from their prehuman states. Such activities, with burning, affect the distributions and densities of keystone species, both faunal and floral. Farming, of course, has a stronger imprint. At the time of European contact, the North American forests and prairies were cultural landscapes subjected for millennia to human burning and harvesting. Evidence for Paleoindian influence is still elusive because recovery of any organic remains from such sites is rare. The animal species identified so far in Paleoindian sites on both sides of the US-Canadian border include no extinct Pleistocene species; even one example would change the prevailing evidence that people arrived after their demise. The worked mammoth tusk at Hiscock is still only circumstantial evidence, and not in an archeological context (Laub et al. 1996).
Awareness of the Younger Dryas climate changes and their now excellent dating has not filtered into the Paleoindian literature significantly. Issues still to be addressed include (1) human responses to Younger Dryas change at different latitudes, (2) efforts to refine radiocarbon calibration by linking sites to climate episodes themselves better dated, (3) the suitability of now submerged coastal areas for Paleoindian habitation (were they as xeric as coasts to the south appear to have been?), (4) Paleoindian use of the resources of the Champlain Sea coast, and (5) the apparent abandonment during part of the Younger Dryas of the easternmost international border area.
The Early Holocene was perhaps equally dynamic in terms of landscape change, perhaps even more challenging for human adaptations, if the scattered, small sites of late Paleoindian and Early Archaic times are reliable witnesses. The period is little explored in the paleoenvironmental literature, leaving archeologists to the inadequacies of the scant archeological record.
Changing Role of the United States in the World Community
In recognition that the existence of the United States has little relevance in Younger Dryas and Early Holocene times, this section addresses the question "How do Earliest American cultural resources contribute to development of broad intercontinental comparative perspectives?" The largest issue at stake is probably the complexities around adaptation by humans to the pressures of living in newly opened lands during the climate changes of Late Glacial and Early Holocene times. Vastly more detailed information is needed before social issues can be addressed at intercontinental scales.
Colonization of the northwest European plains occurred within the time spans of the colonization of North America. Even the environments in northwest Europe were comparable to those in Late Glacial eastern North America, with differences defined by specific resources and population size. The fauna in both areas were nearly modern in composition, with dramatic range adjustments underway. Similarly organized human groups, with remarkably similar equipment, moved into both areas. The largest differences at first were that the Europeans had fairly large, settled populations at their backs, no great distances to the south (e.g., Gamble 1983 and later surveys), while in the Americas, explorer pioneers were expanding the frontiers of human occupation.
The Americas were the last major continents to be inhabited by people. After the inundation of Beringia, the inhabitants lived in relative biological and linguistic isolation from people on the other major continents until contacts at very high latitudes resumed by the middle Holocene. Late Paleoindians were contemporaries with similarly organized Late Paleolithic and Mesolithic people in Europe and elsewhere. The people and their languages derived, evidence strongly indicates, from Arctic and subarctic peoples of Eurasia. How did it happen that northeastern North Americans changed so little in succeeding millennia, retaining flexible economies and small-scale negotiated social structures for several thousands of years? Native American leaders today insist that they did not ‘migrate' to the Americas, and that their culture has changed little because it was ordained. Whatever one's viewpoint, the issues are important for understanding human beings-in-the-world. Paleoindian studies matter.
Clive Gamble (1993:316) reported on "a straw poll conducted by electronic mail" that asked where, in the Americas, was the heartland of Paleoindian studies, "comparable to the Russian Plain and Franco-Cantabria…? There was no consensus except that the Eastern Paleo-Indian distribution was generally thought to be less important than the Western." In the complex, diverse, and fascinating record of the earliest people in the East—whether North or South, I find no justification for this widely held belief.
National Historic Landmark Criterion 6 and National Register Criterion D —