Palaeo-eskimo novice flintknapping in the Eastern Canadian Arctic more

329 Palaeo-Esldnto Novice FlintlOlapping in the Eastern Canadian Arctic S. Brooke Milne The University of Western Ontario London, Ontario, Canada Data from two inland Palaeo-Eskimo sites on southern Baffin Island are evidencefor novice flintknapping. The interior of the island boasts a rich and reliable supply of lithic raw material) and the two sites were occupied during the Araic warm season when accessto it was unrestriaed. The use of a direct procurement strategy to acquire the material at this time of year from such a localized source suggests that acquisition aaivities were structured as part of the seasonal resourceexploitation strategies of the Palaeo-Eskimos on southern Baffin Island. The incidence of noviceflintknapping appears to be direaly linked to this procurement. By drawing on data from published replicative studies and archaeological remains) socially meaningful patterns of idiosyncratic assemblage variability can be identified) providing greater insights into the social components of lithic toolproduction and skill apprenticeship. Introduction Evidence for novice flintknapping from two Early Palaeo-Eskimo sites ca. 2250-800 B.C. (calibrated radiocarbon years) located in the interior of southern Baffin Island in the eastern Canadian Arctic (FIG. I) is presented with the aim of better understanding inter-assemblage variability. In 1990, Phillip Shelley (1990: 187) noted that no one was ever "born" an expert flintknapper. While obvious, lithic analysts should consider the need to learn flintlmapping in their attempts to account for artifact assemblage variability. Archaeologists have used replicative studies (e.g., Ahler 1989a; Bonnichsen 1977; Ferguson 2003; Flenniken 1984; Shelley 1990; Will 2000), ethnoarchaeology (e.g., Stout 2002; Weedman 2002), and archaeological theory (e.g., Finlay 1997; Lillehammer 1989) to understand the material signatures left by novice flintlmappers in the archaeological record and more broadly, differences in lmapping skill (Andrews 2003; Clark 2003). Yet of those archaeological examples of novice flintlmapping presently published (e.g., Grimm 2000; Pigeot 1990; Sackett 1999), few derive from a North American context. This does not mean that novices at New World sites are more difficult to identify than those at Old World sites, but this gap indicates that lithic analysts studying New World sites remain largely focused on techno-functional explanations for assemblage variability. While extremely important, and with some notable exceptions (e.g., Deller and Ellis 2001; Ellis 1994; McGhee 1980; Wilmsen 1974), current analyses commonly do not take steps to consider how individuals in different social and material situations act to contribute idiosyncratic variation from site to site (Wobst 2000: 44). Consequently, socially meaningful patterns of variability are often overlooked or explained away within the techno-functional paradigm. The study of the Early Palaeo-Eskimo assemblages reported here indicates that novice flintlmapping activities were structured by seasonality, site location, and the availability of local lithic raw materials. There also appears to have been flexibility in Palaeo-Eskimo social organization during the Arctic warm season (i.e., late spring, summer, early autumn), which served to facilitate novice lmapping episodes. This discussion places novice lmapping in the broader seasonal resource exploitation strategies of the Palaeo-Eskimos on southern Baffin Island. Palaeo- Eskimo Culture The Palaeo-Eskimos are the earliest lmown culture to occupy the eastern Canadian Arctic, having migrated from Alaska some 5000 years ago. Because of poor preservation conditions in many sites, few organic remains have survived in the archaeological record; thus, much of what we 330 Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic/Milne N Newfoundland 0 ~ ~ \9 I kilometers 100 o Foxe Basin Great Plain of the Koukdjuak CUlJZberland uound P Davis Strait Amadjuak Lake Frobisher Bay Hudson Strait Figure 1. Map of southern Baffin Island showing the locations and features described in the text. know about this culture derives from its lithic artifacts. The stone tools made and used by the Palaeo-Eskimos are distinctive, and typological and morphological similarities, particularly among burins, microblades, Inicrocores, and endblades (bifacially worked harpoon or lance points) have led to speculation among Arctic archaeologists that the Palaeo-Eskimos originated somewhere in the Old World, most likely Siberia (Giddings 1967). Perhaps the most striking characteristics of Palaeo-Eskimo stone tools are their small size and the fact they are complex and skillfully Journal ofFieldArchaeologyjVol. 30) 2005 331 made, suggesting that instruction by experienced lmappers was a critical component of the skill acquisition process. Among the Inuit, oral histories claim that the mal(ers of these artifacts must have been dwarfs precisely because of the small size of the tools (Maxwell 1985 : 40). Several isolated finds, however, indicate that these tiny lithics were mounted into elaborate hafts, handles, and shafts (e.g., Gr0nnow 1994: 219). In the late 1950's, William Irving (1957) coined the name ''Arctic Small Tool tradition" (ASTt), and applied it to all Palaeo-Eskimo sites from Alaska to Greenland where these tiny lithics were found since he believed, based on this trait, that they all belonged to a single culture. Material Signatures of Novice Flintknappers More than a decade ago, Shelley (1990: 188-191) described a group of attributes created by contemporary novice flintknappers when producing bifacial and blade cores. He found that regardless of the reduction strategy used and the type of lithic raw material being worked (i.e., chert or obsidian), novices typically repeated the same mistal(es resulting in the discard of the piece being worked (Shelley 1990: 188). These mistakes consistently involved unsuccessful attempts at flal(e removal from either the biface edge or the blade core platform. Moreover, novices repeatedly tried to remove flakes from the same location using too much force creating "stacked" sets of flal(e terminations on the dorsal or ventral surface of the biface, and from the front of the core directly below the striking platform (Shelley 1990: 188). These repeated attempts at flal(e removal from the same location, combined with misplaced blows, resulted in notable battering of the striking platform. Among bifaces, these stacks yielded thick cross-sections and sinuous edge shapes whereas among blade cores, excessive platform battering left a tiered appearance where all possible angles to salvage the core were effectively exhausted. Among flal(es, skill is reflected in platform preparation, platform trimming, force loads, and flake termination states (Shelley 1990: 191-192). Experts typically invest more energy in preparing striking platforms and trimming overhang from previous flake detachments than do novices. Experts also have better control over the amount of force applied to detach flakes at different stages of reduction, which affects flal(e termination states. When too much force is applied, it can result in snap and plunging terminations (outrepassi)) while not enough force can yield hinge and step terminations. Moreover, because novices have difficulty judging the amount of force required for flake detachment, which results in a lack of control over the fracture process, they generate higher frequencies of flake shatter regardless of reduction technique or lithic raw material (Shelley 1990: 191). Low skill is also associated with high frequencies of medially and longitudinally split flakes. Shelley (1990: 191) refers to these debitage types as "limited information lithic fragments (LILF S)" because they embody little information on reduction strategies since they frequently lack striking platforms and other diagnostic attributes. In contrast, precise force applications by experts result in feathered terminations on intact flal(esindicating their complete and successful detachment. It should be noted that raw material size and quality can yield higher than normal frequencies of LILF S in an assemblage even if the toolmaker is skillful. When a toolmal(er is restricted to using small cobbles and nodules, the flal(esgenerated through reduction are equally small, with thinner cross-sections and more friable edges. This makes them susceptible to breakage or collapse during detachment, which invariably contributes to higher frequencies of broken flakes and shatter (LILFS) in debitage assemblages, especially because smaller cores are harder to work by virtue of their size during early stage reduction (Milne 2003a: 314). Poor quality raw materials with vugs, voids, and other inclusions can also fail during reduction creating higher frequencies of LILFS. Evidence From Southern Baffin Island Prior to 2004, only two inland Palaeo-Eskimo sites had been recorded on southern Baffin Island (FIG. I). These sites, lmown as Sandy Point (LIDv-l0) and Mosquito Ridge (MaDv-ll), are located rougWy 10 km apart on the sw shore of Nettilling Lal(e (FIG. 2). Douglas Stenton (1989) discovered these sites in the 1980s while conducting archaeological reconnaissance to investigate the post1000 B.P. (calendar years) occupation of the region by the Thule and Inuit cultures. Sandy Point was excavated in 1986 while Mosquito Ridge was excavated in 2000 (Milne 1999, 2000, 2003a; Milne and Donnelly 2004; Stenton 1989). Sandy Point is a small single component Palaeo-Eskimo site on a narrow point of land approximately 5 km east of the West Burwash Moraine. When the site was discovered, various forces of erosion (wave action, run-off, deflation, and ice push) were causing damage. Consequently, the site was excavated before it was destroyed. Sandy Point is divided into three components: mainland, channel, and island. The mainland component lacks surface features and is defined entirely by the distribution of lithic artifacts. The island component was originally part of the mainland, but over time erosion created a shallow channel (80 cm), which now separates the two areas. A substantial peat layer underlies the island's vegetation and below that is a matrix of 332 Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic/Milne :...... ."'." ...... . . . ...........• t +-Great Plain of the Koukdjuak ~ D C\} : •• ' ::.y ",-::c Nettilling Lake cit:?tJ 0 Tikerqq.:Bay (J o J\r ..... .: .. ' Q ..... (\ ~ {S.::.:.... ~ ..... 0 I ====~_II the locations of Sandy lOkm Figure 2. Map of the Burwash Bay District on southern Baffin Island, indicating Point and Mosquito Ridge (modified from Stenton 1989). Journal ofFieldArchaeologyjVol. 30) 2005 333 Table 1. Artifact frequencies for Sandy Point and Mosquito Ridge. Site Debitage (%) Cores (%) Informal tools (%) Burin spalls (%) Burins (%) Bifaces (%) Scrapers (%) Microblades (%) Total Sandy Point Mosquito Ridge 1176 (92) 19,800 (96.9) 13 (1) 26 (0.1) 29 (2.3) 113 (0.5) 6 (0.5) 127 (0.6) 16 (1.3) 47 (0.2) 18 (1.4) 69 (0.3) 3 (0.2) 7 (0.03) 16 (1.3) 283 (1.4) 1277 (100) 20,472 (100.03) Figure 3. Novice core from Sandy Point. The image on the left is an example of a core made by a contemporary novice knapper (from Shelley 1990). cobbles and coarse sand. Most of the recovered artifacts were located between the peat and cobble layer. A sample of this peat produced a radiocarbon date of 2924 ± 65 B.P. (calibrated radiocarbon years) (Stenton 1989: 239); however, this date must be considered a minimum date for occupation since the cultural layer lies below the peat. The recovered assemblage includes 1277 lithic artifacts and of these, 1176 are debitage while 101 are formal and informal tools (TABLE r). Evidence indicates that the stone sources used here, and at the nearby Mosquito Ridge site, are local In ongIn. Mosquito Ridge is a large multi-component site that was occupied by the Inuit, Thule, and Palaeo-Eskimos. The site is located on top of a gravel esker that extends rough1y 6 km from its most \vestern point, where it intersects the West Burwash Moraine, to its most eastern point where it descends into Burwash Bay. At least 30 features have been recorded at Mosquito Ridge including Inuit and Thule tent rings and semi -subterranean winter houses. Diagnostic Palaeo-Eskimo lithic artifacts and debitage were located on the surface over the entire site. Several test pits excavated at the most eastern and western margins of the site led to the discovery of an extensive Palaeo-Eskimo component that was largely intact and undisturbed. In 2003, a caribou bone excavated from this Palaeo-Eskimo component yielded a radiocarbon date of 4290-4080 B.P., malcing this one of the oldest occupations recorded on Baffin Island, and among the earliest in the eastern Arctic (Milne and Donnelly 2004: 96-97). The Mosquito Ridge lithic assemblage comprises 20,472 lithic artifacts, of which 19,800 are debitage and 672 are formal and informal tools (TABLE r). Based on the distribution of lithic debitage and stone tools made of chert, chalcedony, and crystal quartz, which spans approximately 1 Ian in linear extent, it appears that Mosquito Ridge was an important part of the inland settlement system (Stenton 1989: 335) and that the PalaeoEskimos returned to the site many times. Most of this lithic assemblage came from the undisturbed areas of the Palaeo- Eskimo occupation of the site; it is unlikely that the small amount of debitage and tools recovered from within the Thule and Inuit features were made and used by these more recent peoples. As Stenton and Park (1998: 55) note, there are two features that distinguish the lithic technology of the Palaeo-Eskimos from the more recent occupants of the Arctic: a nearly complete abandonment of chipped stone tool production in favor of ground stone tool production and a simultaneous shift away from the use of raw materials such as chert, chalcedony, and crystal quartz to the exclusive use of slate and other slate-like materials. Because of these pronounced shifts in lithic technology, concerns of assemblage contamination are unnecessary. It is most lilcely that the Thule and Inuit camped on top of the earlier Palaeo-Eskimo deposit and unwittingly dug into it, incorporating these much older artifacts into their occupation debris. The Assemblages Novice lithic artifacts identified at the Sandy Point site are flake cores, bifaces, and debitage. There are 13 cores, of which two display evidence of low skill in the form of platform battering and stacked step terminations on ventral surfaces, indicating repeated unsuccessful attelupts to remove flakes. One of these cores is virtually identical to a novice core illustrated by Shelley (1990: 189; FIG. 3). The Sandy Point core is intact, has a tiered striking platform, and has, if the platform could be corrected, some remaining utility. Arguably, discarding a core like this prior to ex- 334 Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic/Milne Figure 4. Novice bifaces from Sandy Point. The catalogue information recorded on the center specimen obscures the stacked fractures on the artifact's dorsal surface yet flalce scars below it illustrate repeated but failed attempts to remove it. Figure 5. Novice biface from Sandy Point (right). The image on the left is a biface made by a contemporary novice lU1apper (from Shelley 1990). haustion is typical of expert behavior because experts may be able to evaluate the quality and potential of a core for further reduction (Pigeot 1990: 138); they know to quit while they are ahead. Novices lack the technical slcills to make these assessments, however, and are more lil(ely to discard a useable core in frustration if they cannot fIX the errors (Shelley 1990: 191). The Sandy Point assemblage includes four intact bifaces, three of which exhibit evidence of low technological slcill (FIGS. 4, 5). These three tools display stacked step fractures on their dorsal and ventral surfaces, and consequently, their cross-sections are thicker with bi -convex and plano-convex forms. These tools also exhibit moderately sinuous edges and lack evidence of use-wear. Approximately 23.5% of the Sandy Point debitage assemblage displays attributes associated with low slcil!. These flakes exhibit high frequencies of minimallymodified strilcing platforms (crushed, shattered, unpre- pared/cortical, single facet, and combinations of these states); platform battering; use of excessive force (eraillure scars, pronounced bulbs of percussion, heavy compression rings); and hinge, step, and outrepasse terminations (TABLES 2-4; FIG. 6). While these platform states are common during early stage reduction, their consistent co-occurrence with attributes indicating excessive force, platform battering, lack of platform trimming, and incomplete or failed flake detachment suggests, according to Shelley (1990), Grimm (2000), Finlay (1997), and Stout (2002), among others, the presence of low skill as opposed to the haphazard workings of an expert at the earliest stages of reduction. The frequency of LILFS at Sandy Point is also high with 84 pieces (7%) of shatter and 351 (29.80/0) medially and longitudinally split flakes (TABLE 5). Lithic artifacts made by novices at Mosquito Ridge include bifaces and debitage. There were only 19 flake cores, all of which are fragmentary. Of those that were recovered, 32% are made on materials that are not represented in the debitage assemblage in any notable frequency suggesting these cores were brought to the site, minimally modified, and abandoned. Interestingly, these cores are all made from high quality raw materials including a near-translucent, highly vitreous gray chert and crystal quartz. Given the intensity of use exhibited by these artifacts, it appears they were part of a heavily curated toollcit carried by the Palaeo-Eslcimos to Mosquito Ridge where they were eventually abandoned, likely in anticipation of retooling using local raw materials. Sixty-nine bifaces were found at Mosquito Ridge, but like the flake cores most were fragmentary; only six specimens were complete. Of these, two intact bifaces display evidence of low slcill (FIGS. 7, 8) with moderate to heavy edge sinuosity, ventral and dorsal stacked fractures, and no use-wear. Blade cross-sections are moderately thick with biconvex morphologies. The first specimen (FIG. 7) appears to have been rejected because it could not be thinned, while the smaller biface was discarded without being completely roughed out (FIG. 8). This specimen has considerable cortex and several raw material flaws (voids, vugs) , which suggest it was made on a poor quality flake selected from the debitage assemblage. Given these flaws, it is unlikely an expert would select it for biface production given the availability of more suitable pieces to work. Both bifaces exhibit irregular outlines suggesting a lack of foresight in the shaping process or an inability to work the material to achieve a more symmetrical end product. The frequency of debitage attributable to novice activities at Mosquito Ridge is lower (3.50/0) than that observed for Sandy Point, but attribute states reflecting minimal energy investment in flake detachment, the use of heavy force Journal of Field ArchaeologyfVol. 30y 2005 335 Table 2. Frequencies of prepared striking platforms for Sandy Point and Mosquito Ridge. Site Indeterminate (%) Minil1tally modified (%) Advanced (%) Total (%) Sandy Point 435 (37) Mosquito Ridge 6095 (39) 565 (48) 6251 (40) 176 (15) 3282 (21) 1176 (100) 15,628* (100) * A sample of 79% of the Mosquito Ridge debitage assemblagewas drawn for analysis.Thus, all attribute frequencies presented here are based on a study population of 15,628 flal<:es. Milne 2003a for a detailed discussion of this See sampling procedure. Table 3. Frequencies of force load attributes for Sandy Point (of 1176 artifacts) and Mosquito Ridge (of 15,628 artifacts). Site Eraillure scars (%) Compression rinjJs (%) Bulbs of percussion (%) Battering (%) Sandy Point Mosquito Ridge 200 (17) 682 (58) 1900 (12) 10,264 (66) 989 (84) 8055 (52) 388 (33) 1960 (21) Table 4. Distal Flake termination state frequencies for Sandy Point and Mosquito Ridge. Site Feather (%) Step (%) Hinge (%) Outre passe (%) Snap (%) Absent/ indeterminate (%) Total (%) Sandy Point Mosquito Ridge 436 (37.1) 4463 (28.5) 95 (8.1) 1332 (8.5) 48 (4.1) 712 (4.6) 11 (0.9) 74 (0.5) 502 (42.7) 7814 (50) 84 (7.1) 1233 (7.9) 1176 (100) 15,628 (100) Table 5. Debitage completeness indices for Sandy Point and Mosquito Ridge. Site Shatter (%) Medial/distal (%) Split (%) Proximal (%) Complete (%) Total (%) Sandy Point Mosquito Ridge 84 (7.1) 1231 (7.9) 348 (29.6) 4862 (31.1) 3 (0.3) 289 (1.8) 537 (45.7) 7423 (47.5) 204 (17.3) 1823 (11.7) 1176 (100) 15,628 (100) loads, and a lack of precision in the detachment process remain relatively high (TABLES 2-4). Similarly, the incidence ofLILFS is high with 1231 (6.2%) recorded as shatter and 3580 (18.1%) as medially and longitudinally split flakes (TABLE 5). Distinguishing Novice Skill from Expert Skill Archaeologists can be certain that novice lmappers contribute to the variability observed in lithic tool assemblages. Determining exactly how much of an assemblage results from novice actions presents challenges, however, particularly when trying to distinguish different levels of lmapping skill (e.g., Andrews 2003; Clark 2003). Before one can attempt to quantify flintlmapping levels, two important factors must be considered. First, the availability of lithic raw material will directly influence mistakes in lithic reduction, whether created by experts or novices. In instances where raw material is in short supply, experts are more likely to try to salvage the tool or rejuvenate it; thus, evidence of mistalces may be eliminated. In contrast, at quarries or source locations, the need to recover, repair, or rework raw material is reduced entirely because it is abun- dant and a new piece can easily be picked up to start over again. Consequently, the visibility of mistakes is potentially much higher at these sites. There are two additional factors further compounding this: tools that are successfully lmapped at an acquisition site are removed when the occupants leave resulting in a higher frequency of rejects, which can amplify or distort the actual rate of error (Clark 2003: 224); and novices are more likely to practice in locations where raw material is abundant, which means there may be two different skill levels contributing crude tools or rejects to the same assemblage. Second, experts can effectively erase novice flintknapping activities at a site by correcting novice mistalcesor salvaging their rejected pieces (Shelley 1990: 191). During the learning process, experts frequently help novices, which may involve giving advice or providing actual hands-on assistance to remove a problem area (Stout 2002: 702-703). This apprenticeship system creates a blurred boundary separating novice from expert reduction. Moreover, depending on the nature of the apprenticeship and the context in which the novice is learning, the expert may be so involved with reduction that he or she may elim- 336 Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic/Milne Figure 6. An example of a debitage flake from the Sandy Point assemblage displaying all of the attribute states characteristic of novice knapping (i.e., minimally modified striking platform, battering, pronounced hinge, distal termination). Figure 7. A novice biface from Mosquito Ridge. Figure 8. Unfinished novice biface from Mosquito Ridge. inate altogether any appreciable evidence of novice flint lmapping at the site. Of course, experts occasionally produce crude-looking tools that appear to exhibit evidence of low skill. It is also certain that experts make mistakes during tool production and that the pieces being worked sometin1es fail and are discarded as a result. Experts are less likely to repeat the same mistalces, however, since, as Clark (2003: 221) observes, they have more experience breaking rocks and become better at it the longer they do it. Clark (2003: 230) notes, "the products of an apprentice should exhibit a greater error rate and less uniformity, resulting from lack of experience, than the output of the master." In other words, the mistakes novices malce and their consistency differ qualitatively and quantitatively from the mistakes made by experts. This should make it possible to distinguish those items produced by novices from those Inade by experts who might have been having an off day. Consideration of the conditions influencing the occurrence of novice lmapping will also facilitate this distinction. These refer to the locational, material, and social contexts influencing when and where novices learn, and must be considered collectively when trying to identify the material signatures of novice lmappers in a lithic assemblage. Novice Flintknapping: Conditions of Occurrence For more than a century (e.g., Cushing 1895; Holmes 1891), archaeologists have replicated prehistoric lithic artifacts in order to understand the processes of stone tool production. Through these replications, contemporary knappers have demonstrated how difficult it is to acquire and master the skills necessary to work flint (e.g., Callahan 1979; Crabtree 1966, 1972; Whittaker 1994). As with any craft, experts consistently produce finely made objects while novices frequently malce mistakes. As noted above, novices tend to repeat the same mistakes, creating distinct material patterns, or signatures, which are discernible from those made by experts (e.g., Bonnichsen 1977; Shelley 1990; Stout 2002). With the growing database of replicative lithic experiments, archaeologists are now comparing lmown patterns of variability relating to skill acquisition Journal ofFieldArchaeologyjVol. 30y 2005 337 and archaeological assemblages to identify low skill and potential novice lmapping episodes in the past. In fact, Ferguson (2003: 115) states that the "best evidence experimental archaeology can provide is a match between archaeological indications of past activities and similar signatures in the replicated behavior." To understand these signatures it is necessary to identify those conditions that influence when and where novice lmapping episodes occur. Recent archaeological (e.g., Finlay 1997; Pigeot 1990; Sackett 1999; Walthall and IZoldehoff 1998) and ethnoarchaeological (e.g., Stout 2002; Weedman 2002; Will 2000) studies have found that novice lmapping episodes are conditioned by raw material availability, location, the age of the individual, and group social organization. Together these factors determine where and when novices will engage in tool production although some of them are easier to predict in the archaeological record than others. Because novice lmappers frequently make mistakes, their reduction episodes consume large quantities of material (Finlay 1997; Pigeot 1990; Shelley 1990; Walthall and IZoldehoff 1998; Will 2000). Consequently, archaeologists should expect to find patterns associated with novice actions in locations where raw material is abundant. These locations may include lithic quarries or other raw material source areas such as bedrock outcrops or pebble-strewn beaches (Finlay 1997; Ferguson 2003; Shelley 1990; Will 2000). Since lithic ra\v material was a critical resource for many mobile hunter-gatherers, its acquisition was a scheduled activity within the course of the seasonal round. Opportunities for novices to engage in lmapping activities would also have been scheduled along with raw material procurement and other economic tasks (Pigeot 1990: 138). If access to stone is restricted due to seasonal (i.e., snow and ice ground cover) or geological conditions, episodes of novice lmapping will be reduced. Predicting the locations where novice lmappers will learn in relation to raw material source areas is simpler than predicting who will teach them and at what age this process begins, but that does not mean this kind of information is inaccessible in the archaeological record. Technology is inherently social and lithic production tal(es place in a structured social setting where, depending on the situation and who is present, status can be negotiated through the display of individual skill (Dobres 1995; Dobres and Hoffman 1994; Sinclair 2000). For individuals who have not yet acquired the necessary skills to competently flintlmap, accessto this arena may be restricted. To gain access, individuals go through a period of apprenticeship that can last for many years (Stout 2002: 702). The acquisition of skills is a gradual operation typically occurring during an individual's childhood or adolescence (Pi- geot 1990: 136). Sometimes individuals are in their early teens when they start to learn, while in other instances they may be in their late teens or early twenties (Stout 2002: 702; Weedman 2002: 738). Ferguson notes (2003: 124) that the role of upper-body strengtl1 and the ability to precisely control the application of force may have prevented very young children from learning to flintlmap. In other words, apprenticeship may not have started until individuals acquired the necessary strength and motor skills to manipulate the tools. Finlay (1997: 207) comments further that to be an apprentice one does not have to be a child. While the age of apprenticeship undoubtedly varies among cultures, the process cannot begin until an adult agrees to instruct and guide a novice (Stout 2002: 702). Skill acquisition does not just involve learning "how to do;" it also involves learning "how to act" (Stout 2002: 694). Stout's (2002) ethnoarchaeological study of the stone adze makers of Irian Jaya, Indonesia, describes how novice lmappers are enculturated through tool-making. Older experts control each step in the learning process. They decide when a novice is ready to learn, what stones they willimap, where they can lmap, and for how long the apprenticeship will last. The interaction between novices and experts is relaxed. Novices are encouraged by experts and rewarded with praise when they successfully remove a flake. Novices receive guidance about the reduction techniques to use and the attitude to have towards this revered activity (Stout 2002: 703). This interaction not only provides support and motivation for novices while they are learning but it also fosters other important social values that are vital to this culture, including dedication, perseverance, and concentration (Stout 2002: 703). Learning to mal(e stone adzes facilitates novice enculturation by exposing them to the accepted norms that structure their technological, social, and economic environment. Among hunter-gatherers, adults who instruct novices are usually close family members (Binford 2001; Stout 2002; Weedman 2002) who ensure that children receive the necessary education that enables them to develop into functioning members of their society (Binford 2001: 467). These adults provide young people with educational experiences that include repeated visits to important places where the apprentice is guided by this lmowledgeable person (Binford 2001: 467). The time required to teach novices flintlmapping is structured so that it does not conflict with other activities performed by the group (Pigeot 1990: 138; Dobres 1995). In the course of a seasonal round, social organization is bound to change with the performance of different tasks. Some tasks pose greater risks for a group if they are unsuccessful in completing them. Therefore, to reduce po- 338 Palaeo-Eskimo Novice Flintknapping in the Eastern CanadianArctic/Milne tential risk, individuals will form different kinds of social units. These units may be based on long-term cooperative projects between families or individuals; they may also be temporary ad hoc associations that dissolve once a task is complete (Binford 2001: 465). The relationships that structure these social units are not always based on kinship. They may be determined by such factors as compatibility in age, physical abilities and skill, reproductive requirements, or individual states such as pregnancy or disability (Binford 2001: 466). This means the social organization of hunter-gatherers is potentially very flexible. Social units do not simply comprise members of a nuclear or extended family, and they do not necessarily change to accommodate seasonal subsistence pursuits. Important tasks such as skill acquisition figured prominently in these peoples' lives and undoubtedly required important changes in social organization to accommodate them. Seasonal Settlement Systems and Inland Sites land ecosystem was exploited. To understand why novice flintknapping took place at these sites we must first identify the cultural and economic conditions that governed it. Novice Flintknapping Palaeo- Eskimo Sites in Inland Raw Material Availability Access to lithic raw materials was undoubtedly of central importance to the Palaeo-Eskimos because stone tools were used in all parts of their daily lives (Maxwell 1984, 1985). On southern Baffm Island lithic raw materials are geologically restricted, and in the coastal areas there are no lmown quarry sites or outcrops where stone can be reliably procured. Those sources that are available are small, weathered chert pebbles found on the ocean floor and along the shores bordering the coastal headlands (Maxwell 1973: 10-11; Odess 1998: 422--423). These pebbles can only be obtained in the summer when the winter shore-fast ice is gone and the tides recede. In contrast to its dearth on the coast, lithic raw material is readily available in the interior of the island in the form of secondary deposits that were left by retreating glaciers. These deposits consist of chert nodules or cobbles of varying sizes, colors, quality, and texture. Amadjual( Lal(eis one of two large lakes found in the interior of the island and its name is an English corruption of the Inuktitut word ammaq or angmalik. Ammaq means chert and the reference to an ammaq lake, loosely translated, means "the place chert comes from" (Stenton and Park 1998: 25). The ubiquity of this resource in the interior would strongly favor the occurrence of novice lmapping episodes while the absence of reliable sources in the coastal regions would curtail or restrict novice lmapping altogether. This is because the costs associated with malcing inland journeys to acquire raw material and then transporting it back to the coast would have been too great to afford novices the chance to consume this stone through practice reduction episodes. Seasonality Snow and ice place severe restrictions on the seasonal availability of lithic raw material and the ease with which it can be acquired by Arctic stone-tool using cultures (Binford 1979; I(uhn 1991; Rolland 1981; Hayden 1989; Wenzel and Shelley 2001). Therefore, the best time of year to acquire this resource is during the summer and early autumn. This is not to say that raw material could not be stockpiled for use at other times of the year. Arctic archaeologists believe the Palaeo-Eskimos were higWy nomadic and may have only stayed at a campsite for one or several nights before moving on (Maxwell 1985; McGhee 1996). The Palaeo-Eskimos were higWy adapted to hunting both marine and terrestrial species. In the spring, PalaeoEskimo populations presumably traveled inland to exploit caribou, Arctic char, and nesting waterfowl, and with the onset of winter, they moved back to the coast to hunt seals on the sea ice. To accommodate this seasonal mobility, Palaeo-Eskimo social organization is thought to have been flexible, comprising 30 or more people during the winter, and dispersing into camps of one or two families for the rest of the year (Maxwell 1985: 98). Dwellings included lightly constructed skin tents during the warm season and snow-walled houses during the winter (Ramsden and Murray 1995). This inland/coastal seasonal settlement pattern operated for centuries for the Thule and Inuit, and it is commonly accepted that earlier Palaeo-Eskimo groups followed a similar pattern given the seasonal availability of resources in the Arctic (e.g., Maxwell 1985: 77-107; McGhee 1990: 49; Meyer 1977: 258-259). The vast majority of known and investigated Palaeo-Eskimo sites, however, are located in coastal regions (Bielawski 1988: 57). Consequently, the data with which to interpret Palaeo-Eskimo lifeways are biased towards the marine segment of this culture's adaptation. In fact, comparatively little is known about inland Palaeo-Eskimo occupations. Only a handful of sites have been identified in the subarctic Barrenlands (e.g., Gordon 1996, Meyer 1977), on Banks Island (Miiller-Beck 1970), and on southern Baffm Island (e.g., Stenton 1989; Milne 2000, 2003a; Milne and Donnelly 2004). Therefore, the identification of novice flintlmapping in two inland PalaeoEskimo sites represents a substantive contribution to our understanding of who these peoples were and how the in- Journal ofField ArchaeologyjVol. 30 2005 y 339 Therefore, it is unlikely that the Palaeo-Eskimos remained in one place long enough to warrant the stockpiling that more sedentary stone-tool using cultures are thought to do (Parry and IZelly1987). Moreover, the most reliable stone sources are in the interior, so if stockpiles had been assembled they would have been close to these inland locales. This means Palaeo-Eskimos would have been tethered to the interior of the island to use stockpiled material and to eliminate the added costs involved in long-distance transport of stone to the outer coastal regions. Archaeological evidence (Milne 2003a; Milne and Donnelly 2004) indicates the southern Baffin Palaeo-Eskimos did not stay in the interior for long, however. Rather, it appears they spent considerably more time (late fall, winter, and early spring) in the outer coastal regions and on the sea ice. The ethnographic record documents a similar pattern for the Inuit who remained on the coast for longer periods to hunt ringed seal, which are the most reliable food resource in the eastern Arctic (Boas 1964 [1888]). Consequently, to mitigate this patchy lithic resource environment, the Palaeo-Eskimos would have had to either move raw material long distances from an interior stockpile, or intensively reduce it to form blade cores, bifaces, preforms, and blanks to maximize the amount of useable stone, or cutting edge (IZelly1988), to be carried. Beck et al. (2002) studied the intensity of lithic field processing activities in relation to the distances stones were carried. They found that when distance exceeds 60 km from the point of acquisition (quarry) to the final destination (residential site), raw materials were more intensively reduced at the source to facilitate transport of as much useable stone as possible. This pattern is consistent with the reduction strategies identified at Sandy Point and Mosquito Ridge, and should be expected since the distance from the interior to anyone of the southern Baffin coastlines easily exceeds 60 km. Therefore, it seems unlikely raw material was stockpiled at sites in the interior by the PalaeoEskimos given the logistical challenges and costs in energy and time that would have been incurred to access it. Chert pebbles available in the coastal areas were used for tool production during warm season coastal upland occupations (Milne 2003b) but as Maxwell (1973: 11) notes, the size of these pebbles and their inferior quality would have directly impacted the size, style and shape of the tools made from them. Thus, these stones are by no means desirable sources with which to refurbish a formal toolkit, and their use appears to have been limited to expedient tool production (Milne 2003a, 2003b). Site Location Sandy Point and Mosquito Ridge are ideally located for exploiting local inland lithic source areas. To the west and sw of the sites lies the Great Plain of the IZoukdjuak (FIG. 2), which is a massive, flat, marshy tundra area that supports hundreds of thousands of nesting waterfowl during the Arctic warm season (Soper 1928; Stenton 1989, 1991a). Among the most abundant species today is the snow goose (Chen hyperboreous hyperboreous) with recorded numbers of more than one million birds (Stenton 1989: 94). The annual molt for these birds begins in early July (Soper 1928: 91-95), which makes them easy to catch since they cannot fly. A well-preserved faunal assemblage was recovered from Mosquito Ridge and indicates the occupants of the site intensively exploited snow geese (Milne and Donnelly 2004). The abundance and reliability of this food source and the fact it required relatively little time and energy to acquire, enabled the Palaeo-Eskimos to focus their attention on the acquisition of lithic raw material (Milne and Donnelly 2004: 102, 107-108). The faunal assemblage confirms the occupation of Mosquito Ridge during the Arctic warm season, because this is the only time of year when the birds are in the region. Faunal remains were not found in association with the Sandy Point Palaeo-Eskimo occupation. The site is interpreted, however, as a warm season one given the abundance of local lithic raw materials and the procurement and reduction strategies identified, which mirror those recorded at Mosquito Ridge. Activity Scheduling As with all northern populations past and present, task scheduling in accordance with the seasonal availability of resources is of strategic importance. On southern Baffin Island the resource base is today more abundant and predictable compared to many other Arctic regions (Maxwell 1985: 81-82) and, as a result, there is a certain degree of flexibility in the pursuit of seasonal subsistence activities. Ethnographically, the most important warm season activity for the southern Baffin Inuit is caribou hunting (Boas 1964 [1888]; Stenton 1989). Skins are of critical importance for making winter clothing and a specific number of hides must be secured every summer to malce new garments for the upcoming winter (Stenton 1991b: 6-7). Baffin Island supports several resident caribou herds and these animals are available throughout the year in the interior and the coastal uplands; however, their distributions vary (see Milne and Donnelly 2004: 94; Stenton 1989: 96, 1991a, 1991b). When the herds are stable, it is not necessary to travel inland to hunt since enough animals can be found in the coastal uplands (Jacobs and Stenton 1985: 62-63; Stenton 1989: 112-119, 1991a: 21-27). Other warm season species, namely nesting waterfowl and Arctic 340 Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic/Milne char, are also abundant in the coastal uplands, which, again, reduces the need to venture inland to secure adequate subsistence resources during this season. Only when the herds experience periodic fluctuations where their numbers decline is there a need to go inland to hunt because the smaller herds tend to congregate there (Stenton 1989: 119). The Palaeo-Eskimos undoubtedly exploited the same subsistence resources as the southern Baffin Inuit, which would imply a similar degree of flexibility in their land-use strategies and subsistence practices during the Arctic warm season. Because good stone sources are geologically restricted on southern Baffm Island, the Palaeo-Eskimos would have had to travel inland to procure it regardless of coastal food abundance. In effect, the journey inland for stone acquisition must have been a scheduled activity of primary importance. Both Sandy Point and Mosquito Ridge provide evidence supporting this proposition (Milne 2003a). Sandy Point was an ephemeral occupation where the principal lithic reduction activities focused on raw material testing, early stage core reduction, and the limited production of tool preforms and blanks. A high degree of homogeneity in raw material type and quality in the debitage assemblage indicates local chert sources were worked at the site. The wide variety of raw material colors and high incidence of cortex (34.6%; N =407) indicate that small cobbles or nodules were exploited. Since the chert available in the interior is of higWy variable quality, toolmal(ers lil(ely wanted to obtain the best possible stone for tool production; therefore, raw material testing at this site was expected to be present. Following AWer (1989a, 1989b), a mass analysis was conducted on the debitage assemblage. All of the flakes were size-graded using six mutually exclusive categories: ~ 1/8 inch (less than 5 mm); 1/4 inch (6-12 mm); 1/2 inch (13-19 mm); 3/4 inch (20-25 mm); 1 inch (26-30 mm); and greater than 1 1/4 inch (greater than 31 mm). This yielded a unimodal distribution where 68% of the assemblage measured between 1/4 inch and 1 1/4 inch or greater in size. Because the Palaeo-Eskimos were exploiting small parent materials to begin with, these stones would necessarily yield even smaller prepared cores, tool blanks, tools, and so forth. Therefore, the size of the debris generated from reduction at virtually every stage becomes compressed into the smallest category. So, the fact there are proportionately more large size flal(esat Sandy Point is significant even if these flakes are not as massive or numerous as those found in quarry sites documented in more southern regions (e.g., AWer 1989b; Bradbury and Franklin 2000; Larson 1994; Root 1997). High frequencies of these large flal(es at Sandy Point indicate the presence of primary reduction and/or raw material testing. Dorsal scar counts and flal(eweights were also unimodal, all of which indicate that toolmakers at the site were using a single reduction strategy focused on primary reduction and early stage tool production. Moreover, there is no evidence (e.g., small flakes with low weights, high dorsal scar counts, and elaborately prepared striking platforms) to suggest that late stage tool finishing was completed at this site. In other words, the site occupants were testing raw material and roughing it out into useable cores, blanks, and preforms that could be transported elsewhere for further reduction and use. The flal(e-to-tool ratio illustrates this process. Flal(e-to-tool ratios are simple measures used to examine the extent to which lost tool utility at a site was being replenished through tool production activities (Ricklis and Cox 1993: 450-451). Flal(e-to-tool ratios are calculated here using all complete and fragmentedburins, microblades, cores, bifaces, scrapers, retouched informal tools, and burin spall tools. Unretouched flal(etools, burin spalls, and bifacial edges were not included in the ratio calculations since they lack post-detachment modification and, therefore, do not contribute any bypro ducts to the debitage assemblage. The flal(e-to-tool ratio for the Sandy Point assemblage is 14:1 (1176 flal(es/84 tools), which is moderate. High ratios are expected in those sites where core, biface, and blank production are the principal activities (Deller and Ellis 1992: 89). Production of these tools yields significant quantities of lithic debris and, because the finished products and blanl(s are typically removed from the site, there are proportionately fewer tools in the overall assemblage. In contrast, low ratios indicate that fewer flal(es are being generated through tool production and maintenance activities while higher frequencies of tools are being discarded. When this occurs it indicates new tools are not being manufactured to replace those that fall out of use (Ricklis and Cox 1993: 450). Moderate ratios are expected when reduction activities fall between these two extremes, as appears to be the case at Sandy Point where raw material testing and limited reduction generated comparatively fewer flal(es than those expected for full tool production activities. The sheer size of the debitage assemblage at Mosquito Ridge and its distribution throughout the entire site area suggests it may have functioned as a lithic workshop where the Palaeo-Eskimos came to retool and renew their supply of lithic raw material. The flake-to-tool ratio of 40: 1 (19,800 flakes/495 tools) for this site is notably high (Milne and Donnelly 2004: 103-104), further demon- -~--~-- ----~------- Journal ofFieldArchaeologyjVol. 30y 2005 341 strating the importance of tool production. The mass analysis for Mosquito Ridge yielded unimodal distributions in flalcesize, weight, and dorsal scar counts indicating toolmalcers were focused on a single reduction strategy. The higher frequency (TABLE 2) of small flalceswith moderately higher scar counts and advanced striking platforms (multiple facets, abraded or ground, faceted or bifacial, and any combination of these states) indicates that reduction was focused more on middle stage reduction than raw material testing and early core reduction, like at Sandy Point. Toolmalcers at Mosquito Ridge were reducing local lithic sources, but the lower than expected frequency of cortex in the debitage assemblage (12.3%; N = 2430) suggests that these materials were brought to the site in a reduced or semi-reduced state. If toolmakers left Mosquito Ridge to procure these parent materials or procured them prior to their arrival, it would make sense to first test the nodules where they were found and partially reduce them to facilitate their transport to Mosquito Ridge. Sandy Point certainly provides support for this procurement strategy and these two sites are well within wallcing distance of one another along the West Burwash Moraine since they are only 10 km apart. It may be possible that carrying raw materials between these two sites was part of the learning process for novice toolmakers. As Stout (2002: 696) observed, the act of acquiring raw material was important since it taught novices where to go to find stone, how to select a good piece, and how to extract it for reduction. Acquiring stone is obviously the first and most important step in tool production. Coincidentally, Wenzel and Shelley (2001) recorded a similar pattern of raw material testing and transport at another ASTt site known as Mosquito Lake located in the interior of Alaska. The Palaeo-Eslcimosused this site as a lithic workshop. Toolmalcers obtained their raw material from a nearby quarry, which consisted of small, poor quality nodules and cobbles. Extensive raw material testing was conducted at the source locations to find good quality stone, which was transported back to the Mosquito Lalce site. This testing resulted in a lower frequency of large cortical flalcesat this workshop site, which mirrors the pattern observed at Mosquito Ridge. Conceivably, the Palaeo-Eslcimos could have traveled from Mosquito Ridge to Sandy Point where they located abundant chert nodules, tested them, partially reduced them, and brought them back to Mosquito Ridge for further reduction. Alternatively, the task group could have stopped at Sandy Point first, acquired the stone, and traveled with it further up the West Burwash Moraine to Mosquito Ridge. While the broader analysis of these sites (Milne 2003a) did not find definitive evidence (e.g., tool refits, corresponding radiocarbon dates) directly linlcing them, the similarity in raw material attributes and the stages of the lithic reduction continuum isolated in both assemblages suggests they were complementary to one another in a larger technological and land-use system. The activities at both sites focus on raw material acquisition and it appears that toolmakers were, in effect, using a direct procurement strategy. There is no evidence that lithic procurement was performed as a secondary activity to caribou hunting, a fact plainly illustrated by the lo\v frequency of caribou remains in the Mosquito Ridge faunal assemblage (Milne and Donnelly 2004: 106-107). The cooccurrence of novice bifaces, flalces, and shatter indicates novices were at Mosquito Ridge, yet the comparatively lower frequencies of these artifacts in this otherwise large assemblage seems to suggest that they were not as involved in tool production as they had been at the Sandy Point site. This could be an artifact of excavation since a large part of the Mosquito Ridge site has yet to be investigated. This difference may also be attributable to the fact that Sandy Point was an early reduction site, which might result in a higher proportion of LILF S since they are more prevalent when using heavier force loads to work smaller raw materials (see above). Other activities might well have occupied the attention of novices at the site, such as bird hunting or even socializing (Milne 2003a; Milne and Donnelly 2004). Learning to Imap stone was undoubtedly an important activity at these sites for novices but it likely did not preclude learning other slcills,such as hunting snow geese during the summer molt. This might account for the lower incidence of novice Imapping at Mosquito Ridge given the site's more direct proximity to the Great Plain of the IZoukdjuak and the extensive nesting grounds there. Novices may have been encouraged to help snare, net, or drive the birds into traps to lcill them in large numbers. Ethnographic accounts describe the Inuit (Boas 1964 [1888]: 103-105; Soper 1928: 91-92) using these approaches to hunt geese and it is possible that the Palaeo-Eslcimos did as well. Compared to the Sandy Point site, the sheer size of Mosquito Ridge indicates that this was a preferred spot for repeated occupation for the Palaeo-Eslcimos, and also the later Thule and Inuit cultures (Milne 2003a: 281; Stenton 1989: 335). One possible reason for this preference is Mosquito Ridge's location on top of a gravel esker, which resembles a compact gravel road and is easily navigable by both humans and animals. The network of moraines that characterize the West Burwash District provide the highest local relief in an otherwise extremely flat landscape and the visibility of this esker would make the Mosquito Ridge site easy to find for the Palaeo-Eslcimos year after year. In con- 342 Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic/Milne trast, Sandy Point is situated roughly 5 km away from the West Burwash Moraine and the land between the site and moraine consists of boggy, uneven tundra that is difficult to traverse. There are no discernable features on the landscape near the Sandy Point site; thus, if people were walking in the area, the site would be comparatively easy to miSS. Conclusions It is clear why novice knapping occurred in the interior region of southern Baffm Island. First, and perhaps most importandy, there was abundant raw material for novices to work, and the ease with which it could be procured, tested, and reduced (since these were warm season occupations), meant novices could consume as much of it as they wanted without posing undue demands on the immediate supply. Second, both Sandy Point and Mosquito Ridge are located in close proximity to the Great Plain of the IZoukdjuak where the Palaeo-Eskimos would have had easy access to fat-rich, flighdess geese. This ready food supply allowed more time to be spent on toolmaking. The abundance of food, combined with a warmer seasonal climate, would certainly create a relaxed social atmosphere favoring a positive teaching environment for novices. The identification of novice flindmapping at Sandy Point and Mosquito Ridge suggests that this kind of enculturation was a scheduled activity tied to raw material procurement. These activities were of central importance to the seasonal resource exploitation strategies practiced by the Palaeo-Eskimos on southern Baffm Island, and no doubt, both were most successfully practiced during the Arctic warm season. The presence of novices in the interior provides additional insights about the Palaeo-Eskimos. Those individuals who made the trip inland to learn toolmaking were older, at least in their later teenage years, and physically capable of making this long-distance journey. This suggests that stone tool apprenticeship was a structured social activity within this culture and that it may have been delayed until adolescence when basic criteria of age and strength were met. It also seems likely that the composition of task groups t~aveling to procure raw material was based on factors other than kinship. In other words, individual attributes such as age, skill level (or lack thereof), and endurance would certainly have factored into the formation of such groups. It is unlikely that these groups comprised members of a nuclear or extended family. Consequendy, we can expect a variety of different social groups to have utilized this area, depending on seasonality and the tasks at hand. Similarly, we can further expect a broader range of site types to reflect these varied tasks and the different aggregations of people performing them. Arctic archaeologists must talce this into consideration when studying Palaeo-Eskimo lithic assemblage variability; otherwise, these kinds of socially meaningful patterns will be overlooked. Replicative studies like those conducted by Shelley (1990), Ferguson (2003), and Will (2000) among others should not be underestimated for their contributions to understanding the complexities involved in lithic production. Their research provides a baseline from which archaeologists can study variability in more meaningful ways. As argued here, interpretations about past cultures can be expanded to learn about phenomena like enculturation and social organization. Interpretations of lithic assemblages are not solely in the domain of the techno-functional paradigm. This is not to say that these kinds of studies are not important, because they most certainly are; but, by building on them and expanding our analyses to look for idiosyncratic behaviors, we can learn more about the people in the past, rather than about their artifacts alone. Acknowledgments The research presented here was funded by the Social Sciences and Humanities Research Co"uncil of Canada Doctoral and Postdoctoral Fellowship Programs, the Association of Canadian Universities for Northern Studies (ACUNS) Student Fellowship Program, the Northern Scientific Training Program, and the McMaster University School of Graduate Studies. I am grateful to Douglas Stenton for his insights on the Sandy Point and Mosquito Ridge sites. His pioneering research on southern Baffin Island has helped to inspire my subsequent investigations of the Palaeo-Eskimo occupation of the region. Several people offered comments and suggestions that helped to improve this paper, including Aubrey Cannon, Chris Ellis, Mike Spence, Susan Jamieson, Douglas E. Clark, Jay IZ. Johnson, John C. Whittalcer, and John E. Clark. S. Brooke Milne (ph.D. McMaster University) 2003) is a Postdoctoral Research Fellow and Adjunct Professor in the Department of Anthropology at The University of Western Ontario. Her primary research interests focus on lithic technology) small-scale hunter-gatherer societies)subsistence/settlement strategies) social O1;ganization)and the early prehistory of the eastern Canadian Arctic. Mailing address: Department of Anthropology) Social Sciences Center, The University of Western Ontario) London) Ontario) Canada) N6A 5C2. Ahler, Stanley A. 1989a "Experimental I<napping with KRF and Midcontinental Cherts: Overview and Applications:' in Daniel S. Amick and Raymond P. Mauldin, eds., Experiments in Lithic Tech- Journal of Field ArchaeologyjVol. 30) 2005 343 nology. BA.R. 199-234. International Series 528. Oxford: B.A.R., 2001 shed of Southwestern Ontario. Memoirs of the Museum of Anthropology24. Ann Arbor: University of Michigan. "Evidence for Late Paleo-Indian Ritual from the Caradoc Site (AfHj-l04);' American Antiquity 66(2): 267-299. 1989b "Mass Analysis of Flalcing Debris: Studying the Forest Rather Than the Tree;' in Donald O. Henry and George H. Odell, eds., Alternative Approaches to Lithic Analysis: Proceedings of the University of Tulsa Conference on Lithic Analysis. Tulsa: University of Tulsa, 85-118. Andrews, Bradford 2003 "Measuring Prehistoric Craftsman Slcill: Contemplating its Application to Mesoamerican Core-Blade Research;' in K. G. Hirth, ed., Mesoamerican Lithic Technology:Experimentation and Interpretation. Salt Lake City: University of Utah Press, 208-219. Beck, Charlotte, Amanda K. Taylor, George T. Jones, Cynthia M. Fadem, Caitlyn R. Cook, and Sara. A. Millward 2002 "Rocks are Heavy: Transport Costs and Paleo archaic Quarry Behaviour in the Great Basin;' Journal ofAnthropological Archaeology 21: 481-507. Bielawski, Ellen 1988 "Palaeo-Eskimo Variability: The Early Archaic Small Tool Tradition in the Central Canadian Arctic;' American Antiquity 53: 52-74. Binford, Lewis R. 1979 "Organization and Formation Processes: Looking at Curated Technologies;' Journal ofAnthropological Research 25: 255-273. 2001 Constructing Frames of Reference: An Analytical Method for Archaeological Theory Building Using Hunter-Gatherer and Environmental Data Sets. Los Angeles: University of California Press. 1888.) Lincoln: Dobres, Marcia-Ann 1995 "Gender and Prehistoric Technology: On the Social Agency of Technical Strategies;' World Archaeology 27 (1) : 25-49. Dobres, Marcia-Ann, and Chris R. Hoffman 1994 "Social Agency and the Dynamics of Prehistoric Technology;' Journal of Archaeological Method and Theory 1(3): 211-258. Ellis, Christopher J. 1994 "Miniature Early Paleo-Indian Stone Artifacts from the Parkhill, Ontario Site;' North American Archaeologist 15(3): 253-267. Ferguson, Jeffrey R. 2003 '~Experimental Test of the Conservation of Raw Material in Flintknapping Skill Acquisition;' Lithic Technology 28(2): 113-131. Finlay, Nyree 1997 "'Kid-IZnapping': The Missing Children in Lithic Analysis;' in Jenny Moore and Eleanor Scott, eds., Invisible People and Processes:Writing Gender and Childhood into European Archaeology. London: Leicester University Press, 203-213. Flenniken, Jeffrey J. 1984 "The Past, Present, and Future of Flintknapping;' Review ofAnthropology 1984: 187-203. Giddings, J. Lewis 1967 AncientMen Annual Boas, Franz 1964 The Central Eskimo. (Original published University of Nebraska Press. of the Arctic. Alfred A. IZnopf: New York. Bonnichsen, Robson 1977 Models for Deriving Cultural Information From Stone Tools. Archaeological Survey of Canada) Mercury Series Paper 60. Ottawa: National Museums of Canada. Bradbury, Andrew P., and Jay D. Franldin 2000 "Raw Material Variability, Package Size, and Mass Analysis;' Lithic Technology25 (1): 42-58. Callahan, Errett 1979 "The Basics ofBiface IZnapping in the Eastern Fluted Point Tradition: A Manual for Flintknappers and Lithic Analysts;' Archaeology of Eastern North America 7: 1-180. Clark, John E. 2003 "Craftsmanship and Craft Specialization;' in K. G. Hirth, ed., Mesoamerican Lithic Technology: Experimentation and Interpretation. Salt Lal(e City: The University of Utah Press, 220-233. Crabtree, Donald E. 1966 '~Stoneworker's Approach to Analyzing and Replicating the Lindenmeier Folsom;' Tebiwa 9: 3-39. 1972 An Introduction to Flintworking. OccasionalPapers of the Museum of Idaho State University 28. Pocatello, ID: Museum of Idaho State University. Cushing, Frank H. 1895 "The Arrow;' American Anthropologist 8: 307-349. Deller, Brian D., and Christopher J. Ellis 1992 Thedford II: A Paleo-Indian Site in the Ausable River Water- Gordon, Bryan C. 1996 Peopleof the Sunlight; Peopleof the Starlight: Barrenland Archaeologyin the Northwest Territories of Canada. Mercury SeriesArchaeological Survey of Canada Paper 154. Hull, Quebec: Canadian Museum of Civilization. Grimm, Linda 2000 '~pprentice Flintknapping: Relating Material Culture and Social Practice in the Upper Palaeolithic;' in Joanna S. Derevenski, ed., Children and Material Culture. New York: Routledge, 53-71. Gr0llllow, Bjarne 1994 "Qeqertasussuk -The Archaeology of a Frozen Saqqaq Site in Disk:o Bugt, West Greenland;' in David Morrison and Jean-Luc Pilon, eds., Threads of Arctic Prehistory: Papers in Honour of William E. Taylor, Jr. Archaeological Survey of Canada Mercury SeriesPaper 149. Hull, Quebec: Canadian Museum of Civilization, 197-238. Hayden, Brian 1989 "From Chopper to Celt: The Evolution of Sharpening Techniques;' in R. Torrence, ed., Time) Ene1EY)and Stone Tools.Cambridge: Cambridge University Press, 7-16. Holmes, William H. 1891 "Manufacture of Stone Arrow-Points;' American Anthropologist 4: 49-58. Irving, William N. 195 7 '~ Archaeological Survey of the Susitna Valley;' Anthropological Papers of the University ofAlaska 6(1): 37-52. Jacobs, John, and Douglas Stenton 1985 "Environment, Resource, and Prehistoric Settlement in 344 Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic/Milne Upper Frobisher Bay, Baffm Island:' Arctic Anthropology 22(2): 59-76. Kelly, Robert L. 1988 "Three Sides of a Biface:' American 717-734. Antiquity 53 (4) : Kuhn, Steven L. 1991 "'Unpacking Reduction': Lithic Raw Material Economy in the Mousterian of West-Central Italy:' Journal of Anthropological Archaeology 10: 76-106. Larson, Mary-Lou 1994 "Toward a Holistic Analysis of Chipped Stone Assemblage:' in Philip Carr, ed., The 01fJanization of North American Prehistoric Chipped Stone Tool Technologies. International Monographs in Prehistory) Archaeological Series 7. Ann Arbor, Michigan: International Monographs in Prehistory, 57-69. Lillehammer, Grete 1989 ''A Child is Born: The Child's World in an Archaeological Perspective:' Norwegian Archaeological Review 22(2): 89-105. McGhee, Robert 1980 "Individual Stylistic Variability in Independence I Stone Tool Assemblages from Port Refuge N.WT.:' Arctic 33(3): 443-453. 1990 Canadian Arctic Prehistory. Hull, Quebec: Canadian Museum of Civilization. Resources to Pre-Dorset Subsistence Strategies in the Interior of Southern Baffin Island:' Arctic Anthropology 41 (1) : 90-112. Miiller- Beck, Hans 1970 Illustrations of Artifacts From Umingmak Site) Banks Island. Archaeological Survey of Canada Manuscript 613. Ottawa: National Museums of Canada. Odess, Daniel 1998 "The Archaeology of Interaction: Views from Artifact Style and Material Exchange in Dorset Society:' American Antiquity 63(3): 417-435. Parry, William J., and Robert L. Kelly 1987 "Expedient Core Technology and Sedentism:' in J. K. Johnson and C. A. Morrow, eds., The 01fJanization of Core Technology. Boulder: Westview Press, 285-304. Pigeot, Nicole 1990 "Technical and Social Actors: Flintknapping Specialists at Magdalenian Etiolles:' Archaeological Review From Cambridge 9(1): 126-14l. Ramsden, Peter, and Marribeth Murray 1995 "IdentifYing Seasonality in Pre-Dorset Structures in Back Bay, Prince of Wales Island, N.WT:' Arctic Anthropology 32(2): 106-117. Ricklis, Robert, and Kim Cox 1993 "Examining Lithic Technological Organization as a Dynamic Cultural Subsystem: The Advantages of an Explicitly Spatial Approach:' American Antiquity 58 (3): 444-461. Rolland, Nicholas 1981 "The Interpretation of Middle Paleolithic Variability:' Man 16: 15-42. Root, Matthew J. 1997 "Production for Exchange at the Knife River Flint Quarries, North Dal(ota:' Lithic Technology 22: 33-50. Sackett, James R. 1999 The Archaeology of Solvieux. Los Angeles: Institute of Archaeology, University of California at Los Angeles. Shelley, Phillip H. 1990 ''Variation in Lithic Assemblages: An Experiment:' Journal of Field Archaeology 17: 187-193. Sinclair, Anthony 2000 "Constellations of Knowledge: Human Agency and Material Affordance in Lithic Technology:' in Marcia -Ann Dobres and John Robb, eds., Agency and Archaeology. New York: Routledge, 196-212. Soper, John D. 1928 A Faunal Investigation of Southern Baffin Island. Canada Department of Mines Bulletin 53. Ottawa: National Museum of Canada. Stenton, Douglas R. 1989 "Terrestrial Adaptations of Neo-Eskimo Coastal-Marine Hunters on Southern Baffin Island, NWT:' unpublished Ph.D. dissertation, University of Alberta, Edmonton. 1991a "Caribou Population Dynamics and Thule Culture Adaptations on Southern Baffin Island, N.WT.:' Arctic Anthropology 28(2): 15-43. 1991 b "The Adaptive Significance of Caribou Winter Clothing for Arctic Hunter-Gatherers:' Etudes/Inuit/Studies 15 (1): 3-28. Stenton, Douglas R., and Robert W Park 1998 Ancient Stone Tools of Nunavut: An Illustrated Guide. Ottawa: Parks Canada. 1996 Ancient People of the Arctic. Vancouver: University of British Columbia Press. Maxwell, Moreau S. 1973 Archaeology of the Lake Harbour District) Baffin Island. N ational Museum of Man Mercury Series Archaeological Survey of Canada Paper 6. Ottawa: National Museums of Canada. 1984 "Pre-Dorset and Dorset Prehistory of Canada:' in David Damas, ed., Handbook of North American Indians 5, Arctic. Washington, D.C.: Smithsonian Institute, 359-368. Prehistory of the Eastern Arctic. Orlando: Academic Press. 1985 Meyer, David 1977 Pre-Dorset Settlements at the Seahorse Gully Site. Mercury Series Archaeological Survey of Canada Paper 57. Ottawa: National Museums of Canada. Milrte,S.Brooke 1999 "Pre-Dorset Lithic Technology: A Study of Lithic Assemblage Variability in an Inland Pre-Dorset Site:' unpublished Master's thesis, Trent University, Peterborough, Ontario. 2000 "Pre-Dorset Lithic Technology: A Study of Lithic Assemblage Variability in an Inland Pre-Dorset Site:' in Martin Appelt, Joel Berglund, and Hans C. Gullov, eds., Identities and Culture Contacts in the Arctic. Danish Polar Center Publication 8. Copenhagen: Danish Polar Center, 149-158. 2003a "Peopling the Pre-Dorset Past: A Multi-scalar study of Early Arctic Lithic Technology and Seasonal Land Use Patterns on Southern Baffin Island:' unpublished Ph.D. dissertation, McMaster University, Hamilton, Ontario. 2003b "IdentifYing Pre-Dorset Structural Features on Southern Baffin Island: Challenges and Considerations for Alternative Sampling Methods:' Etudes/Inuit/Studies 27(1-2): 67-90. . Milne, S. Brooke, and Sarah M. Donnelly 2004 "Going to the Birds: Examining the Importance of Avian Journal of Field ArchaeologyjVol. 30) 2005 345 Stout, Douglas 2002 "Skill and Cognition in Stone Tool Production: An Ethnographic Case Study from Irian Jaya;' Current Anthropology 43(5): 693-722. Walthall, John A., and Brad Koldehoff 1998 "Hunter-Gatherer Interaction and Alliance Formation: Dalton and the Cult of the Long Blade;' Plains Anthropologist 43(165): 257-273. Weedman, Kathryn J. 2002 "On the Spur of the Moment: Effects of Age and Experience on Hafted Stone Scraper Morphology;' American Antiquity 67(4): 731-744. Wenzel, Kristen E., and Phillip H. Shelley 2001 "What Put the Small in the Arctic Small Tool Tradition: Raw Material Constraints on Lithic Technology at the Mosquito Lake Site, Alaska;' in William Andrefsky, ed., Lithic Debitage: Context) Form) Meaning. Salt Lake City: University of Utah Press, 106-123. Whittaker, John C. 1994 Flintknapping: Making and Understanding Austin: University of Texas Press. Stone Tools. Will, Richard T. 2000 "A Tale of Two Flint-IZnappers: Implications for Lithic Debitage Studies in Northeastern North America;' Lithic Technology25: 101-119. Wilms en, Edwin 1974 Lindenmeier: A Pleistocene Hunting Harper and Row. Society. New York: Wobst, Martin 2000 ''Agency in (Spite of) Material Culture;' in M-A. Dobres and J. Robb, eds., Agency and Archaeology. New York: Routledge, 40-50.