8 Extraction, Reduction, and Production at a Late Paleoindian Chert Quarry in Eastern Québec
Summary
Late Paleoindian bifacial technology is best understood within the context of local geology and the constraints posed by the raw material. Geological characteristics of the material influence fracture and thus not only methods of extraction but also of artefact reduction. Theses natural constraints completely eliminate certain technological choices and oblige the flintknapper to apply certain gestures (gestes). Understanding the geological context present at the raw material source thus becomes a critical factor in understanding the organization of Late Paleoindian lithic technology.
Introduction
This paper describes how the geological characteristics of a chert member or outcrop articulate with the reduction sequence or chaîne opératoire at two prehistoric chert quarries. It demonstrates that the formal aspects of Late Paleoindian stone tool technology are best understood within the context of local geology and the constraints posed by the raw material that control most, but not all, dimensions of extraction and initial reduction. The quarry and habitation/workshop sites described are located in the small village of La Martre on the north shore of the Gaspé Peninsula, Quebec, Canada (Fig. 8.1). All of the sites date to the Late Paleoindian period, which corresponds roughly to 10,000 to 8000 BP (uncalibrated) for this part of far northeast North America. These dates are based solely on stone tool morphology and flaking styles referred to as parallel flaked or “Plano” (Chalifoux 1999a and b; Chapdelaine 1994; Dumais et al. 1996). Understanding the lithic extraction and production process at these quarries is important at a regional level because the Gaspé region contains the largest concentration of Late Paleoindian sites in Quebec, and the presence of chert outcrops at La Martre is undoubtedly one of the main reasons for this considerable Paleoindian presence. Studying the raw material economy of Late Paleoindians in Gaspé is also significant at a continental level given the central role that high quality lithic materials seem to have played in North American Paleoindian adaptations (Goodyear 1989; Meltzer 1989).
Two quarries have been discovered to date in La Martre. The Suroît quarry (Borden site code DhDn-8) is located on a high plateau above the village. This quarry overlooks the La Martre river valley to the east and the Saint Lawrence river estuary to the north. The quarry is between 270–300m (900 and 1000 feet) above sea-level and covers at least 200,000m2. The second quarry, Montagne Bleue (DhDn-9 and 10), is located below in the La Martre River valley along the west branch of the river at 105m (350 feet) altitude. The Montagne Bleue quarry has not been as extensively explored but its extent is estimated to be at least 80,000m2. Systematic surface collections were carried out at both quarries, and extensive excavations have been carried out at thirteen related Late Paleoindian habitation/workshop sites in the village of La Martre, within 3km distance of the quarries (Chalifoux 1999a and b; Chalifoux and Tremblay 1998). Tools and debitage recovered at the quarries are described here with particular attention to the bifacial reduction sequence.
Figure 8.1. Location of the village of La Martre where the quarries are located (indicated by a star), and the Gaspé Peninsula, Quebec (polygon).
Geology
The quarries are located within the Ordovician aged Cap Chat mélange (Occ). This geological unit is a chaotic mix of shale, claystone, chert and other lithologies that were combined during the Taconic orogen. The allochtonous chert within the mélange is originally from the related Ordovician Des Landes formation (Ode) immediately to the south and east of La Martre. The chert blocks that have been included into the mélange can be quite large – up to 1km in maximum dimension in some cases (Slivitzky et al 1991). While the Cap Chat chert is not strictly speaking bedded since it is no longer in situ, some larger blocks measuring several meters will exhibit the original bedding of the chert as it was deposited and formed in the Des Landes formation (Fig 8.2). This fact makes Cap Chat chert attractive to flintknappers since the chert blocks that can be extracted are not necessarily limited in size as they might otherwise be in a mélange. As we will see below, the chert is in fact extracted in a tabular form that often follows the original bedrock bedding. In addition, thin section petrography of the chert shows that it has not undergone significant deformation or recrystallization due to later metamorphism which means that the chert is relatively “fresh” and good to knap (Burke 2002; Slivitzky et al. 1991). However, not all of the Cap Chat chert is highly siliceous or “cherty” and it often has the appearance of a dull, conchoidally fracturing, siliceous mudstone. The geological context of the chert, as well as its visual characteristics, are presented in greater detail in another paper by the author (Burke 2002).
Figure 8.2. Photograph of an outcrop of chert at the Montagne Bleue quarry (DhDn-9 and 10) showing the original sedimentary bedding in the chert blocks that have been incorporated into the Cap Chat mélange.
Bifacial reduction and production at the quarries
Seventy-nine bifacial tools were analyzed from the Suroît quarry (DhDn-8) (Fig 8.3). Callahan’s classification scheme for biface production was used (Callahan 2000, vii, 36–37), which was expressly developed for the Eastern Paleoindian fluted point tradition and is an excellent analogue for the early stages of our Late Paleoindian reduction sequence. The results of the classification are presented in Table 8.1. The range of dimensions for the bifaces using only complete pieces or dimensions is as follows: length, 97mm to 226mm, width, 34mm to 148mm, and thickness, 11mm to 52mm.
Bifacial tools from the Surôit quarry are clearly dominated by Callahan’s (2000) stage 3 “primary pre-forms”. There are also many earlier stage 2 and 2–3 “rough outs”, but few of the later stages 4 and 5 “secondary pre-forms” and “final pre-forms”. Finished tools equivalent to Callahan’s stages 6+ or “flaked implements” (e.g., broken projectile points) are occasionally found at the quarry site, which suggests that the lower number of stage 4 or 5 pre-forms could be an artifact of our sampling. Stage 3 bifacial preforms should have a width-thickness ratio of between 3.0 and 4.0 according to Callahan (2000, 30–31). All of the 79 bifaces together have an average width-thickness ratio of 3.3, and the 35 Stage 3 bifaces average only 3.0. This is an indication that, in fact, many of the pre-forms from the quarry are quite early in the reduction sequence which is corroborated by the large number (22) of Stage 2 and 2–3 “rough outs”. As a comparison, the biface pre-forms recovered at the Early Paleoindian quarry of West Athens Hill in the Hudson River Valley of New York have similar width to thickness dimensions: ratio of 2.7 for stage 2 equivalents, and 3.4 for stage 3 equivalents (Funk 2004). This is noteworthy because the chert at West Athens Hill is also an Ordovician chert making up part of a mélange dominated by shales. Nodules or chunks extracted at West Athens Hill are similar in dimension as well as form to the La Martre quarries, but less often tabular (personal observation at the quarry). Edge angles on the biface pre-forms provide additional information. They range from 30° to 65° for Stage 3 pre-forms, and all but 4 fall into the 40°–60° range predicted by Callahan for stage 3 primary pre-forms.
Figure 8.3. Plan and side views of two complete Callahan Stage 2 bifacial preforms (top DhDn-8.43, bottom DhDn-8.26). Note the sinuous edge on these early stage bifaces as well as the original sedimentary layer still visible on parts of the preform faces. The longer bifacial preform, no. 43, pictured at top is 265mm long by 97mm wide, 29mm in thickness and 898.8g in weight. The shorter example, no. 26, depicted beneath is 188mm long by 76mm wide, 38mm in thickness and 532.8g in weight.
| Stage 1 (Blank) | 3 |
| Stage 2 (Rough Out) | 9 |
| Stage 2–3 | 13 |
| Stage 3 (Primary Pre-form) | 35 |
| Stage 3–4 | 7 |
| Stage 4 (Secondary Pre-form) | 9 |
| Stage 5 (Final Pre-form) | 3 |
Table 8.1. Classification of bifaces recovered from the Suroît Late Paleoindian quarry using Callahan’s (2000) biface stages.
Judging from what was left on the surface of the quarry, the majority of production appears to be dedicated to the production of stage 2–3–4 pre-forms for transport to the habitation/workshop sites in the valley below. In particular, production at the quarry is dominated by the production of “primary pre-forms” (Stage 3) defined by Callahan as having a “symmetrical handaxe-like outline with a generous lenticular cross-section and a straight and centered, bi-convex edge. Principal flakes should generally just contact or overlap in the middle zone… and be without such concavities, convexities, steps, or other irregularities as would hinder successful execution in the next stage” (Callahan 2000, vii). This description corresponds well to our Stage 3 bifacial pre-forms (Fig. 8.3). It also suggests that the majority of production failures at the quarry occurred at this stage of production.
The initial steps of the reduction process or the chaîne opératoire
In analyzing the bifacial production at the La Martre Paleoindian quarries I was intrigued both as a lithic analyst and as a flintknapper as to how the raw material, once extracted, could constrain the initial steps of the reduction process. Would the form of the raw material as it was created geologically, and then extracted and selected at the quarry, set the initial critical parameters, in part or in whole, for the chaîne opératoire? How did prehistoric flintknappers use the tabular chunks to their advantage and how did they tackle problems such as blocks with rectangular cross sections and 90° angles? Of those pre-forms that still retain enough information for analysis, the large majority is knapped parallel to the sedimentary layers in the chert: 52 parallel to the sedimentary layers, and 9 parallel to a joint set (Fig. 8.3). This means that the two faces of the bifacial pre-form and its center plane are parallel to the original sedimentary layers in the tabular chert block. On the early stage bifaces that still contain two layers of sedimentary chert ‘cortex’ the average thickness of the original tabular piece appears to have been rather thin; about 30mm (measured on 10 artifacts). Half of the earlier stage bifaces (Stages 1 to 3) still retain some sedimentary ‘cortex’ on at least one face. This is not cobble or nodular cortex but rather the coarser and less cherty sedimentary layer found between chert layers. Many bifaces (36 in total) show evidence of joint surfaces, and half of these (18) show two or more joints at obtuse angles to each other. Joint surfaces are roughly perpendicular to the sedimentary layers.
The choices that the flintknappers at La Martre have made are in part constrained by the form in which the material is extracted. At the same time, the tabular form also benefits the flintknapper in the production of the bifacial forms by producing a starting point that already has an excellent, or high, width to thickness ratio. Moreover, if the joint surfaces or sets that also help define the shape and size of the tabular chert blocks provide useful angles to begin the reduction process, then the raw material constraints can prove to be an advantage rather than a hindrance for specific reduction sequences or chaînes opératoires. We can use Waldorf’s (1993, 33) models for the initial stages of biface pre-form manufacture to look at how Late Paleoindian flintknappers exploited the combination of sedimentary layers and joint surfaces present in the chert at La Martre (Fig. 8.4).
Figure 8.4. Two approaches to reducing a tabular block with a rectangular or trapezoidal cross section as proposed by Waldorf (1993, 33). Drawings by Val Waldorf. Reproduced with kind permission of the author, D.C. Waldorf, and the publisher, Mound Builder Books.
The majority of bifacial pre-forms analyzed show alternate flaking along each of the long edges of the tabular block. Figure 8.5 presents the schéma diacritique or diacritical schema for two complete biface pre-forms recovered from the Suroît quarry (see also Fig. 8.3). Each edge removal is carried out relatively independent of the other, that is, it does not proceed in a continuous circumferential manner. Some pre-forms show a second technique referred to by Waldorf (1993, 33) as “platform reversal” (Fig. 8.4). This technique produces a characteristic beveled cross-section in which the original sedimentary layers are no longer parallel to the center plane of the biface (Fig. 8.6). This reduction strategy will be favoured when the edges of the tabular block are not at right angles, but rather are formed by joint surfaces at acute angles thus facilitating the initial flaking of the tabular block.
Figure 8.5. Schéma diacritique or diacritical schema showing the sequence of flake removals (sequential Arabic numerals) on each face of the stage 2 bifacial preforms shown in Figure 8.3. In the cases where the sequence of removals alternates between faces and can be clearly distinguished, these are indicated in order of removal by Roman numerals. Grey stippling represents original sedimentary layer ‘cortex’. Drawing and analysis by Manek Kolhatkar.
Figure 8.6. Cross section of two bifacial preform fragments broken during manufacture. Note the opposing beveled edges.
This analysis was not geared specifically towards trying to evaluate or measure the rate of manufacturing errors or failures (cf. Brumbach and Weinstein 1999). It was possible to identify a few end thinning and lateral thinning fractures. Only three overshot (outrepassé) flakes led to failure and rejection. Failure may not be an accurate descriptor since the bifaces could be salvaged, but these three biface pre-forms were abandoned. Many of the production failures show snaps and can be related to bend or perverse fractures as defined by Whittaker (1994, 212–217, see also Waldorf 1993, 50–53; Callahan 2000, 108–113). These would seem to occur on a regular basis in the manufacture of larger bifaces. Only five of the biface pre-forms had visible stacks which probably led the flintknapper to stop the thinning process and abandon the biface preform. Raw material flaws do not seem to be a major factor in the failure to complete the chaîne opératoire. I have identified only one biface which seems to have failed along a joint surface. The chert is in fact surprisingly homogeneous, and even the laminations that reflect the original sedimentary layers do not have any incidence on flaking. There are no vugs or clasts that interfere with the knapping quality either.
Debitage
A sub-sample of 50 complete flakes that were surface collected at the Suroît quarry was analyzed. These flakes are short and wide, average length to width ratio is 1.1, and they are relatively thick (13.5mm). Only 20 have cortex on the dorsal surface (covering 25% to 75% of the dorsal surface) (Fig. 8.7). Those flakes without cortex often have several prior flake scars on the dorsal surface, with an average of 5 flake scars and a range of 3 to 10. Previous flake scars almost invariably originate from opposing directions (Fig. 8.7). Platform angles range from 45° to 80° degrees, with an average of 62.5°. 24 flakes have one flake scar on the striking platform, 15 flakes have two, 3 flakes have three or more scars. Surprisingly, none of the flakes analyzed have cortex on the platform. This lack of cortex on platforms may be a sampling artifact. Conversely, it may be due to the fact that most tabular pieces will not exhibit “cortex” on their sides since that is where the chert block has fractured or separated along a joint surface.
Figure 8.7. Debitage recovered from the Suroît quarry (DhDn-8). Top row: flakes from the initial series of removals along the edge of the tabular block with remnant sedimentary layer on dorsal side, large platforms, and large platform angles. Middle row: second series of flakes removed showing multiple flake scars on the dorsal side originating from various directions, small platforms, and acute platform angles. All flake platforms are towards the top. Bottom row: biface abandoned due to an overshot flake, and overshot flake with remnant biface edge at the top.
Thirty-three flakes have evidence of platform grinding for preparation, only 4 do not. Platform preparation by grinding is also apparent on the bifacial pre-forms. A large number of flakes (18) have bulbar scars (éraillures) and one fifth (10) have strongly lipped platforms. The dimensions of the flakes and the bulb scars, combined with the fact that many biface pre-forms have deeply plunging and indented flake scars (Fig. 8.3), suggests that most early reduction work at the quarry was done by hard hammer percussion. This is confirmed by the discovery of several hammerstones made of rounded cobbles that may have originated in the river valley below or in the surface tills at the quarry itself (Fig. 8.8). The hard hammers are made of quartzite, and an arkose/quartz-arenite/sandstone rock with similar knapping characteristics to quartzite. Soft hammers (caribou antler billets?) most likely were used at the quarry as well since there is ample evidence of thinning of pre-forms to produce finished bifaces (Stage 5+). Most of this thinning would have been difficult with the hard hammers found at the quarry. A substantial proportion, if not the majority, of the finishing of bifaces took place at the nearby habitation/workshop sites. This is based on the many late stage bifaces broken during thinning and the tens of thousands of flakes recovered during excavations at these habitation/workshop sites (Fig. 8.9).
Figure 8.8. Utilized hard hammers found at the Suroît quarry (DhDn-8) showing crushing along the edges. Materials are quartzite and various types of arkose, quartz-arenite, or sandstone based on visual inspection.
Figure 8.9. Late stage bifaces broken during thinning recovered at Late Paleoindian habitation/workshop sites around La Martre village (DhDm-1). Drawings by Sophie Limoges.
Conclusion
The aim of this paper is to describe how the geological characteristics of a chert source articulate with the reduction sequence or chaîne opératoire of chipped stone tools. Two Late Paleoindian chert quarries were used to explore this concept. Our analysis demonstrates that the highly formalized aspects of Late Paleoindian bifacial technology are best understood within the context of local geology and the constraints posed by the raw material. These constraints control most aspects of the extraction of the raw material and therefore the forms of raw material made available to the flintknapper. In this case the raw material takes the form of rectangular tabular blocks bounded by sedimentary layers and joint sets. These geological raw material dimensional characteristics consequently place natural constraints on the initial reduction of the raw material. The natural constraints, in turn, completely eliminate certain technological choices and oblige the flintknapper to apply certain gestures (gestes). This can be seen in those segments of the chaîne opératoire that are visible at the La Martre quarries.
Raw material constraints therefore control most, but not necessarily all aspects of lithic reduction. The related Late Paleoindian habitation/workshop sites in the valley below have in fact produced smaller tools (e.g. scrapers) that are made from flake blanks but these are rare. Most, if not all, formal tool production at the quarries and the habitation/workshop sites follows a bifacial reduction sequence leading to the production of knives, projectile points, drills, and possibly large sidescrapers. Consequently, the geological constraints described above should not be seen exclusively as limitations. The sedimentary layers within the chert that lead to large tabular blocks being extracted provide an excellent starting point for the production of large bifaces given their high width to thickness ratio. In addition, the joint sets often present ideal angles for the initiation of bifacial thinning thus providing opportunities for technological choices rather than simply imposing constraints. As a comparison, the Témiscouata quarries of eastern Quebec also provide a medium to high quality chert that is very similar to the La Martre chert (Burke and Chalifoux 1998). Chert from Témiscouata is also sedimentary, and blocks extracted from the quarry are usually bounded by joint sets. However, in this case the joint sets are more closely spaced than at La Martre and therefore it is impossible to make the large bifaces we see at La Martre. Understanding the geological context present at the raw material source thus becomes a critical factor in understanding the organization of Late Paleoindian lithic technology. In this paper we have only addressed the first steps taken by flintknappers at the quarry, but it seems clear that without a better sense of these initial steps it will be difficult to accurately reconstruct the rest of the chaîne opératoire and the choices made by these people at their campsites in the valley below.
Acknowledgments
I wish to thank Manek Kolhatkar who helped me to do the analysis for this paper and who produced the schémas diacritiques presented in Figure 8.5. Thank you to Sophie Limoges and Éric Chalifoux for allowing me to use the drawings of bifaces in Figure 8.3. My thanks also to D. C. Waldorf for allowing me to reproduce Figure 8.4 from his book. Thank you to the two reviewers who improved this paper. Thank you especially to Éric Chalifoux for inviting me to participate in the research on Late Paleoindians in the Gaspé Peninsula; it was a lot of fun!