Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 13

Chert Raw Material Utilization at the Bark Site (BbGp-12), Peterborough County, Southern

Katie M. Biittner and Susan M. Jamieson

To understand how specific cherts were used in stone tool manufacture at the fifteenth-century Bark site (BbGp-12), Peterborough County, Ontario, 164 artifacts from the chert assemblage are analysed for their raw material attributes. A combination of macroscopic, petrographic, and palynological analyses are used to identify the provenance of the most abundant chert types used by site inhabitants, assigning them to the Onondaga, Fossil Hill, and Upper Gull River Formations. We conclude, contra Eley and von Bitter (1989), that acritarch genus identification is not required to discriminate among these chert types. Instead, relative abundance of different palynomorph subgroups (based on morphology), and relative abundance of acritarchs versus other microfossils (chitinozoan, graptolite, scolecodont, and trilete spore fragments) are more significant identifying characteristics. The attribution of cherts to distant sources allows us to explore ideas about local kin-based access to high quality, exotic materials.

It has long been acknowledged that macroscopic least eight separate chert types are present in the identification is an inadequate means for chert site’s assemblage. Thus, it follows that macro- sourcing (e.g., Luedtke 1978, 1979:745, 1992). scopic analysis is inadequate for a complete char- Nonetheless, this technique is commonly used acterization of the chert artifacts and, conse- by Ontario archaeologists as it is generally held quently, inadequate for the determination of to afford a reasonable degree of reliability while toolstone sources. being both expedient and cost effective. For example, Susan Jamieson (1998, 1999b, 2000, 2002, 2004) routinely used such macroscopic Background to the Bark Site identifications in both her licence and annual published reports on her field work at the fif- The Bark site (BbGp-12) is situated in the mid- teenth-century Bark site. dle Valley, a few kilometers west of In 2005, Katie (Miles) Biittner evaluated the the City of Peterborough (Figure 1). It was dis- validity of Jamieson’s macroscopic identifications covered in 1983, when deep ploughing disturbed of the Bark site cherts for her master’s thesis human remains and revealed several settlement research (Miles 2005). She re-investigated the features. With the approval of Curve Lake, site’s chert artifact assemblage, examining macro- Hiawatha, and First Nations, the site was scopic attributes and undertaking microscopic immediately mapped, surface-collected and test- analysis for the first time. This latter effort con- ed by Mima Kapches of the Royal Ontario sisted of both petrographic (mineralogical) and Museum. In 1986 and 1987, Rick Sutton, then palynological (microfossil) analyses. This re- a master’s student in the Department of examination of the chert assemblage illustrated Anthropology, McMaster University, conducted that Jamieson’s original macroscopic identifica- intensive surface collection and testing at the tions are comparable to the petrographic and Bark site to obtain a more representative artifact palynological identifications. As only five poten- sample and to identify the nature and extent of tial chert sources were identified, Jamieson’s its settlement data. Sutton (1990:25,27) con- analysis was adequate only in terms of establish- cluded that the Bark site was a single component ing the general character of that assemblage. Raw Huron village dating to the mid-fifteenth centu- material analysis outlined here suggests that at ry, one that was not situated for defence. On the 14 Ontario Archaeology No. 81/82, 2006

Figure 1. Places in central southern Ontario mentioned in the text as well as modern collecting localities (natural outcrops, road- cuts, quarries) of cherts identified in the Bark site assemblage (this study); localities are those identified by Eley and von Bitter (1989:4). basis of the seasonality of recovered faunal and century. Ceramic diagnostics and overlapped floral remains, he concluded that it had probably pits, hearths and post moulds indicate intensive, been occupied year round. Lastly, in 1997, 1999, prolonged occupation of the site locus at this and 2000, Susan Jamieson conducted a program time (e.g., Jamieson 2004). of research at the site in conjunction with Trent Relatively little is known about the indigenous University’s field school in Ontario archaeology. peoples who occupied the middle Trent Valley She determined that the site had a late Pickering (Figure 1). Sutton (1990:1) states that “virtually (Early Ontario Iroquoian) period component in no previous Iroquoian archaeological research addition to fifteenth-, early-sixteenth,- and late- has been conducted” there. The Wilson site was, nineteenth-or early-twentieth-century compo- however, excavated by Walter Kenyon (Royal nents. The main occupation dated to the fifteenth Ontario Museum) in the 1950s and surveys of Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 15 the Trent Valley were carried out by Trent The Bark site is located in a frontier area University personnel in the 1960s and early 1970s. between Iroquoian-speaking and Algonquian- Readily accessible published archaeological infor- speaking peoples. Hence, there is a possibility that mation for the middle Trent Valley is limited to a both peoples may have occupied the site detailed analysis of the Bark site settlement data (Jamieson 1998, 2004:71). Farther west, village (Sutton 1990) and summary papers on subsequent site distributions have been used to infer increased annual excavations at this site (Jamieson 1998, trade—beginning at about A.D.1450—between 2000, 2004), a description of the Brock Street bur- representatives of these two language groups ial (Kenyon and Cameron 1961) in the City of (Warrick 2000:451). Such trade is well docu- Peterborough, and discussions of the Peterborough mented into the seventeenth century (Jamieson Petroglyphs (Dewdney and Kidd 1967; Sweetman 1981:20, 1999a:185). Interaction between the 1955; Vastokas and Vastokas 1973) near Nephton. Trent Valley Iroquoians and other Iroquoian- We know that during the fifteenth and sixteenth speaking groups in Ontario has been inferred centuries, the archaeological record of the upper from distributions of material culture (Ramsden and lower Trent Valley indicates transformation 1990; Warrick 2000). We think it likely that lith- and reorganization of society, including population ic raw materials were being exchanged among the migration and integration, restructuring and coa- different Iroquoian-speaking groups as well. lescence of smaller settlements into larger local vil- The chert assemblage analysed in this paper, lages (e.g., Damkjar 1990; Pendergast 1985; which is composed of 1401 specimens, derives Ramsden 1977:70, 1988, 1989, 1996; Sutton from Jamieson’s 1997, 1999 and 2000 investiga- 1990). This reorganization probably also occurred tions at the Bark site. Chert is the most abundant in the middle Trent Valley area. Although trade net- lithic raw material (Table 1), a pattern consistent works of earlier periods were maintained into the with that found at other sites in southern fifteenth century, there is, for this period, relatively Ontario. Preliminary macroscopic analysis by little evidence for exotic lithic raw materials—“exot- Jamieson suggested that the Bark site assemblage ica”—in southern Ontario (Jamieson 1999a:184). is composed of cherts from five formations: Data from this period thus suggest increasing Onondaga, Lower-Middle , Upper sedentism and regionalism combined with intensi- Bobcaygeon, Upper Gull River, and Lower Gull fied use of local resources and pervasive local or River. Lower and Middle Bobcaygeon Formation short-distance alliances and exchanges initiated and cherts have been combined into one type because maintained through ethnic and kin-group linkages they are indistinguishable (Eley and von Bitter (Jamieson 1992:77, 1999a:175, 184). We infer, 1989:10). Jamieson concluded that the majority therefore, that, during this period, emphasis was of the Bark site artifacts were knapped from placed on toolstone that could be locally acquired Upper Bobcaygeon Formation chert, with the but still had the prerequisite or desired flaking qual- next most abundant chert types coming from the ities. We follow Tankersley’s (1989:265) definition Onondaga, Lower-Middle Bobcaygeon, Upper that non-local materials come from a distance of Gull River, and Lower Gull River Formations, more than 30 km whereas local materials are trans- respectively (Table 2). Some bedrock outcrop ported over a shorter distance. locations for these cherts are shown in Figure 1.

Table 1. Lithic raw material distribution at the Bark site by maximal artifact category. Raw Material Type Debitage Flakes Cores Edge Retouch Tools Edge Utilised Tools n% n% n % n % n % Chert 832 80.5 478 75.4 68 94.4 15 93.8 1 50 Chloride Schist 152 14.7 135 21.3 0 0 0 0 0 0 Quartzite 18 1.7 10 1.6 2 2.8 1 6.2 1 50 Quartz 29 2.8 9 1.4 2 2.8 0 0 0 0 Unidentified 2 0.2 2 0.3 0 0 0 0 0 0 Total 1033 100 634 100 72 100 16 100 2 100 16 Ontario Archaeology No. 81/82, 2006

Table 2. Frequency of Bark site cherts by maximal artifact categories as determined by Jamieson. Chert Type Tool Core Debitage Flake Total Onondaga 4 6 253 198 461 Upper Bobcaygeon 7 28 331 184 550 Lower/Middle Bobcaygeon 3 13 148 45 209 Upper Gull River 2 17 71 51 141 Lower Gull River 0 4 29 4 37 Total 16 68 832 485 1401

Raw Material Analyses cleavage and visible twinning, uniaxial positive features, undulose to sharp extinction, low relief We know that to account for the natural varia- and weak birefringence, or polarization at 90 tion in chert, both macroscopic and microscopic degree angles (Deer et al. 1992; Moorehouse analyses are necessary for its characterization. 1959). Cloudiness can be caused by fluid inclu- According to Herz (2001:464), “the first step in sions, ranging from less than a micron to several an analysis of lithic material should be petro- millimeters in maximal dimension. Such inclu- graphic, preferably with thin sections.” sions are common and can be abundant in quartz Petrographic analysis, the identification of the (Deer et al. 1992). Cleavage is the tendency of a mineral composition of a rock, can be used to crystallized substance to split along definite crys- determine the composition and texture of the talline planes, yielding smooth surfaces. chert quickly and relatively inexpensively (Eley A number of different attributes can be exam- and von Bitter 1989:3). ined within each thin-sectioned sample. Attributes such as texture, groundmass, silica Petrographic Analysis fabric, inclusions, and the angularity, roundness The usual technique for preparation of thin-sec- and shape or sphericity of grains are used to dif- tions involves the use of a thin-bladed diamond- ferentiate between different chert types (Clough tipped saw to make two parallel cuts close togeth- and Woolley 1985; Eley and von Bitter 1989; er at a carefully selected place in the side of the Parkins 1974; Wahlstrom 1955; Winchell specimen to a depth of about 5-10 mm, leaving a 1965:1; Writt 1975). Unfortunately, studies of slice as thin as 0.5 mm (Clough and Woolley chert are usually hampered by the lack of consis- 1985:93). This rock slice is ground down until tency among publications (Writt 1975:3). perfectly flat and has an even thickness of around Textural variation is typically the most distinc- 0.03 mm. It is then cemented onto a glass slide tive feature allowing for the identification of dif- (Clough and Woolley 1985:93). The thickness of ferent chert types and/or sources. Textural varia- the slide is important because the “inference tion is used to describe the size and shape of min- colours displayed by minerals under a polarizing eral grains (Clough and Woolley 1985:94; Writt microscope vary according to their thickness” 1975:11). Under a polarizing microscope, quartz (Clough and Woolley 1985:93), as polarized light crystals composing chert are measured in will refract differently at varying thicknesses. A microns (µm) and classified as cryptocrystalline, standard thickness, such as 0.03 mm, ensures that microcrystalline, or chalcedonic quartz. Crystals specific minerals always display a similar range of less than 3 µm and too small to show optical colours, thus aiding in identification and com- properties are classified as cryptocrystalline. parison of samples. Crystals less than 10 µm showing waxy extinc- As chert is composed of quartz and its variants tion are classified as microcrystalline. Crystals (chalcedony and opal), a general optical descrip- that are 5 to 100 µm with submicroscopic water- tion of quartz will aid in textural analysis. Quartz filled pores of about 0.1 µm in diameter are clas- composition is the most distinctive feature of sified as chalcedonic quartz (Eley and von Bitter chert sources. Quartz is most readily distin- 1989:8; Folk and Weaver 1952:501). In chert, guished by its lack of colour, the nature of its the quartz crystals are associated in recognisable Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 17 patterns termed silica fabrics, produced either by $15,000. It is also relatively easy to learn (so say the dissolution of original minerals and their Prothero and Lavin!), involves simple compara- replacement by quartz, or by the growth of tive identification procedures, and yields a quartz crystals in void spaces (Eley and von Bitter “wealth of information” about texture, mineralo- 1989:8). The texture of chert depends on the gy, mode of origin, and chert diagenesis that is type and combination of silica fabrics making up not available from other methodologies its groundmass—cherts with a uniform micro- (Prothero and Lavin 1990:577). Although the crystalline groundmass are fine-textured, whereas preparation of thin sections can be time-con- chalcedonic and drusy quartz mosaics are coarse suming, the time spent “in the mechanical oper- to very coarse-textured (Eley and von Bitter ation of grinding rocks is not only remunerated 1989:8; Writt 1975:14). These varying degrees by interesting and instructive preparations, but of coarseness are determined by the relative pro- adds also to a complete knowledge of their phys- portions of the quartz grain sizes, which can be ical structure, gained by close observation of the assessed by plotting the percentages of the fine- various features they present during the process sized grains against the medium and coarse size of variation” (Fritz-Gaertner 1878:219). grains (Writt 1975:11, 15). Physical measure- As with any analytical technique, petrography ments, including the overall coarseness of the has some limitations. First, thin-sectioning is a texture, the percentage of chalcedony, and the destructive technique. This may be fine for deb- size and amounts of void fillings help to delineate itage and source samples, but the majority of and delimit the specific texture type (Writt archaeologists and curators may be reluctant to 1975:15). Grain angularity, roundness and shape cause even minimal damage to type or rare arti- are used to describe the quartz crystals and to facts. Second, petrography is dependent on vari- refine the texture types. ation, and chert can be homogeneous among Scholars have demonstrated the utility of pet- sources and heterogeneous within a source rographic analysis in chert provenance determi- (Luedtke 1992). Finally, petrographic analysis nation (Boxt and Reedy 1985; Clough and does not allow for the discernment of the chem- Woolley 1985; Eley and von Bitter 1989; Hess ical composition of the chert. This is severely 1996; Janusas 1984; Lavin and Prothero 1992; limiting when there are no distinguishing macro- Pollock et al. 1999; Prothero and Lavin 1990; scopic and microscopic properties visible, even in Shelley 1993; von Bitter and Eley 1984, 1997). thin section. Chert is predominantly composed Prothero and Lavin (1990) and Lavin and of quartz with the main cause of variation being Prothero (1992) employed petrographic analysis the presence of impurities. If these impurities in their study of northeastern and middle- occur, and they can be present in amounts of Atlantic region cherts. Very few of these cherts parts per million (ppm), chemical analysis is the were distinctive macroscopically. Prothero and only means of detection. Lavin (1990:562) were easily able to distinguish them petrographically. Weathering in the form of Palynological (Microfossil) Analysis post-depositional oxidized rings, oxidized Whereas petrographic analysis can be used to dis- opaques, and leached carbonates allowed Lavin tinguish cherts from different formations, micro- and Prothero (1992:110) to distinguish cherts fossil analysis allows for separation of cherts from acquired from secondary versus primary sources. distinct members within a formation. Customarily, Petrographic analysis has a few important microfossils are used in petrographic and strati- advantages over other means of sourcing analysis. graphic investigations to aid in characterizing First, it is relatively inexpensive. Thin sections and determining the age of cherts (Eley and von can usually be done for as little as five dollars per Bitter 1989:9; Martin 1993:476). Cherts are specimen and the method requires only a petro- generally replacements of marine sediments, so graphic microscope. Entry level petrographic microfossils incorporated into them during their microscopes can, however, cost upwards of formation can include many distinct marine 18 Ontario Archaeology No. 81/82, 2006 plants and animals (Eley and von Bitter 1989:9; Onondaga Formation chert. Parkins’s (1974:28) Luedtke 1992:26). The importance of microfossils conclusion that too little information is available in chert stems from the fact that they are fre- on acritarch assemblages in the Onondaga quently the only fossils present in the rock and are Formation (and other southern Ontario) cherts also extremely resistant to destructive geological to make acritarch assemblage data useful in processes, such as weathering, dissolution and re- determining the source of flaking debris remains crystalisation (Eley and von Bitter 1989:9). true, despite the more recent work of Eley and Microfossil analysis is conducted by examining von Bitter (1989). Only when used in conjunc- pollen and spores. The palynological processing tion with macroscopic and microscopic (petro- required for this analysis involves destroying all of graphic) analysis can acritarch assemblage data be the rock except the acid resistant organic materi- successfully applied to archaeological raw materi- al: the rock sample is first digested in hydrochlo- al provenance investigations (Parkins 1974). ric and hydrofluoric acid; the organic material Conversely, macroscopic and thin section (kerogen) remaining is then separated and con- analyses alone are insufficient for chert character- centrated on a micro slide. This organic material ization in the majority of cases (Janusas is then oxidized and sieved to make the organic- 1984:27). In her examination of Kettle Point walled microfossils (pollen, spores, dinoflagel- Chert, Janusas (1984:82) found that the combi- lates, etc.) easier to identify. Using a polarizing nation of thin section analysis and macroscopic microscope equipped with a micrometre, the analysis is worthwhile and that this combined slides can then be examined and analysed. approach to analysis cannot be “discarded in Microfossils common to southern Ontario favour of some other technique” as observations cherts include scolecodonts, chitinozoans, and may be made and recorded which “might other- trilete spores. These microfossils are counted and wise be overlooked” if only one or the other noted for presence/absence only. Specific genera method is employed. are not identified, as this identification is unnec- essary for chert type differentiation. On the other hand, acritarchs, as the most abundant microfos- Sampling Strategy sil present in southern Ontario cherts, are investi- gated more thoroughly. Following the methodol- Owing to the large population size of the Bark ogy established by Eley and von Bitter (1989:9) site assemblage and the limited time and for measuring the composition of acritarch resources of the investigator, the Bark site chert assemblages, the number of acritarchs from each assemblage was sampled using a stratified ran- subgroup is noted and expressed as a percentage dom approach. Debitage and flakes recovered of the total. When genera can be identified, from excavation units were place in correspon- generic diversity is recorded for each subgroup. ding squares representing the original grid on a Eley and von Bitter (1989), Parkins (1974), gridded tabletop. All of the squares were sam- Writt (1975), and von Bitter and Eley (1984, pled. Using a table of random numbers, 10 per- 1997) have each undertaken investigations of cent of each of the debitage and flake sub-popu- southern Ontario chert sources (see Figure 1 for lations was randomly selected to provide the sub- bedrock source locations). These studies have sample from that unit. All artifacts recovered demonstrated the utility of both petrographic from features (except middens) were studied and palynological techniques of analysis in dis- because features are believed to represent a nar- tinguishing among chert formations, as well as, rower time frame of use than the rest of the site. among members within formations. While Eley Likewise, 100 percent of cores and 100 percent and von Bitter’s (1989) work offers the definitive of informal tools were studied. In all, a total of analysis of chert formations in southern Ontario, 164 specimens were analysed using a combina- Parkins (1974) provides an excellent examination tion of macroscopic and microscopic techniques of the Edgecliff and Clarence Members of (Table 3). Of the 164 specimens selected, only Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 19

135 specimens were large enough to be thin-sec- Properties assessed microscopically are used to tioned for examination of petrographic struc- infer raw material quality. Fine-grained cherts are tures and only 39 were large enough to processes considered to be of high quality; coarse-grained palynologically. cherts, low quality. Following Brantingham et al. (2000), cherts in this study can be ordered on a Table 3. Population and sample size. scale or index of quality (Figure 2). Maximal Artifact Category Population Sample Size (n) Chert texture is greatly affected by inclusions. Debitage 832 43 Their presence or absence is one of the most use- Flakes 485 80 ful and significant petrographic attribute meas- Cores/Core fragments 68 33 Informal Tools 16 8 urements. This study determined that the most Total 1401 164 useful inclusions for determining texture are peloids, pellets, quartz eyes, glauconite, and fer- Results runginous minerals.

Table 4 presents attributes that were assessed for Palynological Attributes each of the three methods of analysis. For macro- All of the processed samples had microfossils scopic analysis, these attributes include colour, present. Few of them were, however, identifiable patina, lustre, mottling, banding, and speckling. owing to a high degree of fragmentation. For petrographic analysis, they include texture and Acritarchs from seven of the eleven subgroups inclusions. The palynological analysis was helpful were visible in most of the samples: sphaeromor- in confirming or otherwise refining the results of phitae, acanthomorphitae, herkomorphitae, the macroscopic and petrographic analyses. netromorphitae, polygonomorphitae, pteromor- phitae, and prismatomorphitae. In other words, Macroscopic Attributes all of the acritarch subgroups found by Eley and Artifact size proved to be a limitation in that some von Bitter (1989:9) in southern Ontario cherts attributes (e.g., patina) were impossible to assess occur in this sample, including the following for smaller pieces . As cores (mean size grade, 13- acritarch genera: Veryhachium, Multiplicis- 18 mm) are generally larger than flakes and deb- phaeridium, Micrhystridium, Cymatiosphaera, itage (mean size grade, 7-12 mm), cores are easier Diexallophasis, Polyedryxium, Pterospermella, to characterize on the basis of macroscopic attrib- Baiomeniscus and Baltisphaeridium. The samples utes. Patina and colour together are the most diag- could not be characterized by their acritarch nostic raw material attributes. Patina cannot assemblage composition as these assemblages are always be reliably assessed as it can be influenced not specific to any chert source and are common by features of the rock (chemical composition, to numerous chert types. Other microfossils that mineralogy, structure), features of the deposition- were present were too fragmentary for identifica- al, weathering or formation environment (soil pH, tion, though they most likely represented chiti- rainfall, temperature, bacterial activity, drainage), nozoans, scolecodonts and conodonts. Some and position of the rock in, or on, the soil definitive scolecodonts were present. However, (Sheppard and Pavlish 1992:41). When colour these, too, are common to numerous chert types. and patina suggest the presence of more than one Whereas the samples could not be character- type of raw material, lustre, as well as the presence ized by acritarch assemblage composition, almost or absence of mottling, speckling and banding can all samples contained other microfossils and aid in distinguishing those types. microlites. Following Lentifer and Boyd’s (1999) morphology-based subgroups for phytolith clas- Petrographic Attributes sification, three generalised chert types could be Texture is the most useful petrographic attribute. identified using these other microfossil and It is measured along an ordinal scale as deter- microlite assemblages. Group 1 cherts contain mined by the nature of the carbonate fabric. large (15–30 µm) microlites with fewer numbers 20 Ontario Archaeology No. 81/82, 2006

Table 4. Distinguishing macroscopic and microscopic attributes for Bark site cherts.

ettle Point ssil Hill undee nondaga pper Bobcaygeon pper Gull River K Fo Lockport D O U Lower/MidddleU Bobcaygeon Macroscopic Features Black to dark grey colour X X X X Dark to light grey colour X X Blue-grey colour X X Light grey colour X “Rusty” patination X Light (yellow or buff) patination X X X X Staining X Speckling X Mottling X Vitreous lustre X Waxy lustre XX Dull lustre X X Peloidal texture X Visible inclusions X Petrographic Features Carbonate groundmass X “Murky” groundmass X X X Organic matter X X Chalcedony fabric X X Drusy quartz mosaic fabric X Void-infilling quartz mosaic fabric X X Quartz “eyes” X Carbonate rhombs X Ferruginous rhombs X X Peloids X Pellets X Palynological Features Dominant acritarch: Leiosphaerdium X Veryhachium XXX Cymatiosphaera XX Polyedryxium X Pterospermella XX Multiplicisphaeridium XX Micrhystridium XX Other palynomorphs dominant X X Rare microfossils X X X of other palynomorph fragments. Group 2 cherts analyses, which suggests that microlites do not contain small (1–10 µm) microlites with equal accurately differentiate chert types. numbers of other palynomorph fragments. Often more than one means of analysis is Finally, Group 3 cherts have palynomorph frag- required before an artifact can be assigned a sin- ments but do not contain any microlites. Instead, gle chert type. Furthermore, certain chert types these cherts can be characterized by their acritarch are more reliably identified using one method of assemblages. These microlite/microfossil groupings characterization over another. An example of this do not, however, correspond with the chert types as is Upper Gull River Formation chert, which can determined by macroscopic and petrographic be easily identified using petrographic analysis, Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 21

Problems with Different Approaches to Analysis

Macroscopic Analysis Macroscopic analysis is rarely straightforward. Julig et al. (1988:233) have noted how for “almost a cen- tury these problems associated with the visual iden- tification of chert types and source areas have evoked comment in the archaeological literature.” Luedtke (1978, 1979:745, 1992) states “visual identification is inadequate for any serious study of chert material Figure 2. Southern Ontario cherts ordered by knapping quality. types.” Although Luedtke (1979:745) recognizes, among other researchers, that cherts from different owing to the presence of quartz “eyes” unique to sources are often visually distinctive and archaeolo- this material. gists “working in a particular geographic region Table 5 shows the distribution of items in the quickly become familiar with the most common sample of 167 artifacts examined by chert type local chert types,” several problems are associated using all methods of analysis. A single chert type with the visual identification of cherts. could not be assigned to 37 artifacts (22 percent First, this kind of expertise on the part of the of the sample). Chert type distribution is rela- archaeologist is the result of years of experience tively consistent between and within maximal and frequent handling of the material. Second, artifact categories. Onondaga Formation, Fossil this knowledge can be “arcane” and difficult to Hill Formation, and Upper Gull River communicate (Luedtke 1979:745). These prob- Formation cherts are the most common raw lems become “particularly acute when analysts material types. attempt to deal with widespread social interaction

Table 5. Chert type distribu- Chert Type Tool Core Debitage Flake Total tion for maximal artifact cat- Assignable to Single Types egories determined by agree- Fossil Hill 1 3 6 25 35 ment of results of macroscop- Kettle Point 1 1 3 1 6 ic, petrographic and palyno- Onondaga 3 19 10 9 41 logical analyses. Lockport 0 0 1 2 3 Dundee 0 1 3 0 4 Upper Bobcaygeon 0 1 1 2 4 Lower/Middle Bobcaygeon 0 1 2 5 8 Upper Gull River 2 5 9 10 26 Assignable to Multiple Types Onondaga/Fossil Hill/Dundee 0 1 0 0 1 Onondaga/Lockport 1 0 0 0 1 Onondaga/Kettle Point 0 0 1 0 1 Onondaga/Fossil Hill 0 0 0 1 1 Onondaga/Dundee 0 0 0 2 2 Onondaga/Lower Bobcaygeon 0 0 1 0 1 Onondaga/Bois Blanc 0 0 2 3 5 Onondaga/Bois Blanc/Dundee 0 0 0 3 3 Fossil Hill/Bois Blanc 0 0 2 2 4 Dundee/Bois Blanc 0 0 0 1 1 Upper Gull River/Lower Bobcaygeon 0 0 1 0 1 Lower/Middle/Upper Bobcaygeon 0 0 0 10 10 unknown 0 0 2 4 6 Totals 8 32 44 80 164 22 Ontario Archaeology No. 81/82, 2006 spheres and the concomitant movement of lithic requires the examination of larger pieces. This materials over long distances between widely conclusion is supported by the smaller number spaced source areas” (Julig et al. 1988:233). assignable to more than one type of chert (e.g., These problems are a possibility considering the Onondaga/Fossil Hill) identified among the larg- scope of this research, inasmuch as the raw mate- er pieces. Small implements “may obviously be rial sources potentially available to the site resi- made as required from chance finds such as flint dents were widely spaced in a large geographical pebbles from a stream bed or exposures of region. A third problem is the frequent disagree- ancient gravel, and in both cases these are likely ment among experts about the correct classifica- to be at some distance from the parent rock” tion of any one sample. Although characteriza- (Sieveking et al. 1972:151). Research on lithic tion is often undertaken with the intention to materials from the Munsungun Lake Formation “sort and characterize the recovered lithics by suggests that hand specimen (macroscopic) char- source and feature so as to produce macroscopic acteristics combined with polarised light categories that (are) distinguishable with reason- microscopy are “viable techniques for assigning able certitude” (Katalin 1998:3), the problem is lithic provenance” Pollock (et al. 1999:269). The in achieving that very goal. How does one create provenance of rhyolite, quartzes, metamorphic macroscopic categories that are “reasonable” and and igneous rocks was successfully identified mutually exclusive? In their attempts to “objecti- using these techniques. Essentially, macroscopic fy the process of chert identification through visu- analysis used independently of other means of al characteristics” (Luedtke 1979:745), raw material characterization is accurate only for researchers assign more significance to some larger artifacts. attributes, less to others, and change the criteria for definition from one type to another. By com- Palynological Analysis bining macroscopic, petrographic, and palyno- The biggest problems with the palynological logical approaches to analysis, the limitations of analysis presented here derive from the small expertise and the chance of error in chert type dis- sample size and the small size of the artifacts. crimination have been reduced. Only 35 artifacts were deemed to be large Another problem with macroscopic analysis is enough by the laboratory to warrant palynologi- the large number of attribute-states that are not cal processing. When analysed for acritarchs and unique to a single chert type. Macroscopic fea- other palynomorphs, these samples did not pro- tures are “not always diagnostic of a single source, duce representative microfossil assemblages. and are difficult to assess objectively” (Prothero Although the work done by Eley and von and Lavin 1990:562). To clarify, colours are Bitter (1989) suggests otherwise, acritarch genera grouped together by Eley and von Bitter identification is not required to differentiate (1989:14) (e.g., “white, grey to dark grey” as one between chert types from different formations. colour type in their chart). Rankin (1998:208) Instead, relative abundance of different paly- cites the difficulty in distinguishing between nomorph subgroups (based on morphology), Lockport Formation and Onondaga Formation and relative abundance of acritarchs versus other cherts by only assessing visual characteristics. She microfossils (chitinozoan, graptolite, scole- was unable to distinguish from which host for- codont, and trilete spore fragments), appears to mation the cherts in the Nodwell collection came. be more significant for identification. The largest Onondaga Formation chert is also difficult to dis- difficulty encountered was in distinguishing tinguish from Dundee Formation, Lockport between acritarch genera using only a binocular Formation, Fossil Hill Formation, and Bois Blanc microscope. Electron microscopy is required for Formation cherts. the identification of distinguishing generic In this study, the only artifacts classified as attributes, such as process termination type and “unknown” raw material type are either flakes or surface structures. The Bark site data suggest that debitage, which implies that macroscopic analysis focusing on the entire palynomorph assemblage Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 23 is a better means of characterization. Other paly- mobility. The difference in the way chert tools of nomorphs (chitinozoans, scolecodonts and non-local and local origin were manufactured and graphtolites) were, however, often too fragmen- discarded is not evident in the recovered lithic tary to identify. Non-acritarch palynomorph assemblage, which suggests that non-local cherts assemblages appear to be more useful for an were procured without additional energy, time accurate characterization of cherts. and labour costs to the group. However, down- the-line trade presents another, more likely means Petrographic Analysis of procurement. Support for down-the-line trade, Petrographic attributes, notably texture, are used as opposed to embedded procurement, is found to determine the relative quality of cherts present in the large number of wasted or exhausted cores at the Bark site, enabling study of the relation- in the Bark site assemblage, smaller cores being ship among lithic raw material abundance, raw expected at increasing distances from the source, material quality and the design of lithic tool kits. or towards the “end” of the line. Both Onondaga and Fossil Hill cherts are high The numerous waterway systems in southern quality and, relative to the location of the Bark Ontario would have supported a down-the-line site, non-local (exotic). We anticipated exhaus- trade network, allowing lithic raw materials to be tive use of these materials. On the contrary, how- transported from the north shore of Lake Erie ever, we found they were used in an expedient northwards, around the west end of Lake fashion. By way of explanation, we suggest that Ontario, up the Humber River to Lake Simcoe these cherts were being selected for their per- and into the Trent River–Severn River system ceived aesthetic or social characteristics, or both. (Figure 1). Alternatively, chert could have been carried northwards to Lake Huron and Georgian Bay, or from Lake Erie east through Lake Raw Material Use and Implications at the Ontario, then northwest up the Trent River Bark Site towards Lake Simcoe. During the fifteenth centu- ry, the Trent-Severn waterway and its New York The results of our analysis of the Bark site assem- linkages were a significant route of transportation blage suggest that if curated artifacts were pro- (Kapches 1994:13) as were other routes through duced at the Bark site they were carried away and the Niagara Peninsula, southwestern Ontario and retouched elsewhere. In earlier excavations (1986- to the north (Jamieson 1999a:184). 87) at this site, however, Sutton recovered seven Some of the exotic raw materials could have complete projectile points—four bases and four been acquired from secondary sources, such as mid-sections. On the basis of macroscopic char- till. Bobcaygeon cherts, for example, occur in till acteristics, he determined that they were pro- south of Lake Simcoe and across Durham region duced from Onondaga chert. We did not exam- to the east (Eley and von Bitter 1989:25; von ine these artifacts. If the tools are in fact made of Bitter and Eley 1984:135,143). Some raw mate- Onondaga chert, then it is likely that both curat- rial may have come from fluvial deposits where ed and expedient reduction strategies were used at cherts, at least smaller pebbles, have been trans- the site and that different raw material types were ported within river channels. This explanation, being used in distinct ways. Some members and however, has drawbacks: localities of Onondaga chert at bedrock source (Figure 1) are of better quality than others. The in theory, a piece of chert could be trans- expedient and curated utilization of Onondaga ported hundreds of kilometers by a gla- chert could, therefore, represent a difference in cier and then somewhat farther by the raw material quality. rivers that drain it; in reality, transport is Following the ideas of Morrow and Jeffries a destructive process and a number of (1989), the Bark site raw materials may have studies have shown that rocks tend to been acquired within an embedded pattern of decrease in size exponentially as they are 24 Ontario Archaeology No. 81/82, 2006

carried away from their sources. Thus, the that both Iroquoian and Algonquian peoples probability of finding chert in usable quan- occupied the site. Iroquoian populations are tities and sizes decreases as one moves away matrilineal whereas Algonquian populations are from the primary source.... [And] the defi- patrilineal. Difference in kin organisation will nition of a usable size fragment varies by affect how and by whom raw materials are culture. In Late Woodland Michigan, frag- moved by human agents. ments as small as 3 cm in diameter were The effect of kin organization on raw material worked using bipolar percussion... movement ties into trade rights. Trade or [though] for most stone-using peoples, exchange interactions often are limited to a few, usable fragments were considerably larger select individuals. Typically, these individuals are [Luedtke 1992:102]. men, who pass on their rights to trade (and asso- ciated prestige) to their next of kin—from uncle Parkins (1974:25) has noted that chert is too to nephew in matrilineal groups and from father brittle to resist glacial erosion, and Jamieson to son in patrilineal groups. Gould and Sagger’s (1984:67) states that “it is relatively uncommon (1985:120) work with Australian Aborigines for cherts to be moved great distances from their found that although stone tool making was not sources, either by glacial or riverine transport, in an exclusive male activity, only men made special any quantity.” We conclude, then, that although purpose trips to obtain lithic materials. This small pebbles of certain cherts from exotic strict division of labour by sex can be attributed, sources may have been available in a local con- in part, to the patriarchal social organization of text, these raw materials are not significant in the the groups involved in long-range social net- lithic assemblage at the Bark site. works. Nine out of thirteen lithic source locali- We can use petrographic attributes, notably ties visited in their study had sacred associations texture, to examine the relationship between lith- and only men with “specific affiliations to those ic raw material abundance, raw material quality sites could approach” these localities (Gould and and the design of lithic tool kits (Andrefsky Saggers 1985:120). At the Bark site, the analysis 1994:30; Brantingham et al. 2000). Brantingham of technological organization does not help in et al. (2000:261) find that the impact of raw determining cultural identity: assemblages with material quality is a constant constraint through mixed technological strategies are typical of both all stages of reduction. However, Akridge and Algonquian and Iroquoian peoples. Concerning Benoit (2001:143) note that “cherts were also the possibility of identifying differences in tech- praised for their aesthetic value.” For example, nological strategies used within domestic and Late Woodland peoples in Michigan preferred the non-domestic settings at the Bark site, no defin- “strikingly colourful” Upper Mercer chert in ritu- itive statement can be made owing to the high al contexts (Luedtke 1978). For the Anishnabe, degree of site disturbance. Onondaga cherts are prized for their association Renfrew and Bahn (1996:336) state that there with Niagara Falls, one of the stopping places on is a “real equivalence between the interaction seen their migration westward. They may have been as a communications system and the interaction secondarily prized by the Ontario and League as a system for the exchange of material goods.” Iroquois for their associations with supernatural Around the time the Bark site was occupied, beings (Barbeau 1915:8-9, 306; Smith 1983:9). notions of prestige, authority, and status associat- The Bark site posits an interesting problem ed with exotic materials were diminishing. when one examines socio-technical organization. Instead, there was an increasing dependency on Although material culture recovered from the site close-range ties and an intensified use of local has been used to suggest it is a Huron site resources in this part of Ontario (Jamieson (Sutton 1990), the settlement data do not con- 1999a:184, 1992:73). This suggests that main- form with the supposed Iroquoian site model. taining long distance interactions during a period Instead, Jamieson (1998, 2004:71) has suggested of sedentism, regionalism, and consolidation Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 25

(Jamieson 1992:77) was a way to facilitate close ent reduction and exhaustive use of high quality, kin relationships rather than reflecting a desire exotic cherts at the Bark site indicate more than for exotic raw materials and their associated ide- the satisfaction of purely technological needs. ological concepts. These ideological concepts— These particular toolstones not only had aesthet- prestige, status and authority—could have been ic appeal: they also conveyed social prestige. transferred from an earlier association with raw material to the participation in trade (trading rights) instead. Prestige, status and authority, Conclusion: Future Research resulting in power, could be ascribed both to those individuals involved in trade and to the The raw material analysis outlined here suggests trade goods themselves. that at least eight chert types are present. The McGuire (1992:132) states that power becomes original characterization conducted by Jamieson significant when access to raw material sources is (1999b, 2000, and 2002), which identified five influenced or determined by lineage and territorial potential chert raw material sources, is adequate rights. Control over the rights to raw material only insofar as it established the general character sources can belong to certain lineages or groups, of the Bark site assemblage. Macroscopic analysis and not others. Likewise, some lineages or groups alone is not an adequate means for chert type dif- may have rights to trade for materials from these ferentiation. sources, while others do not. This differential con- Fifteen years have passed since Eley and von trol could affect the composition of a chert assem- Bitter’s definitive chert typology was published. blage at a site if its occupants did not include a lin- Since then, a number of new chert sources have eage with trading rights, or kin relations with been discovered that have not yet been character- groups with access to, or control over, raw materi- ized, and the characterization of older, known al sources, or both. Warrick (1984) proposes that sources has come into question. All chert sources Iroquoian households were composed of corporate in southern Ontario and the Upper Great Lakes kin groups that did not compete with each other region need to be re-examined. for wealth but, rather, formed economic and polit- As Eley and von Bitter have established the base- ical coalitions. Nonetheless, these coalitions were line data for southern Ontario chert sources using themselves competitive with each other. This com- petrography and palynology, we propose that petition emphasizes the role that the structure or chemical sourcing of these same sources is the next organization of kinship relations played in the necessary step. Compositional (chemical) character- dynamics of trade and exchange, status and pres- ization usually provides reliable evidence for trade tige. Among the seventeenth-century Ontario and and exchange in determining whether the material League Iroquois, for example, only certain kin was acquired locally or more distantly. Luedtke groups were entitled to name chiefs. (1979, 1992) found, however, that chemical sourc- To summarize, status, prestige, and authority, ing is not always successful with lower Great Lakes which once were associated with the use of the cherts. There is often too much overlap in the raw material itself, later came to be associated chemical fingerprints of the different sources. Cost with control over (or rights to) the raw material and time required for analysis is another reason why source. As control or trade rights were acquired petrographic and palynological analyses were select- by (or maintained by) kin organization, the ed over chemical analysis in this study. maintenance of kin relations would have facili- The identification and description of paly- tated long distance trade. For example, pilgrim- nomorph assemblages has not been undertaken for ages to acquire chert, or status to enable access most chert raw materials recovered from archaeo- to, and use of, materials associated with a sacred logical assemblages. There will have to be a signifi- place (such as Onondaga chert from the vicinity cant amount of data collection before palynologi- of Niagara Falls), create considerable prestige for cal analysis can be used as an effective and efficient the material itself. An implication is that expedi- method of determining chert provenance. 26 Ontario Archaeology No. 81/82, 2006

Finally, the analysis of chert must occur with- References Cited in an archaeological context for each and every Akridge, D.G., and P.H. Benoit chert source separately, allowing us to address the 2001 Luminescence Properties of Chert and some questions of how and when raw material types Archaeological Applications. Journal of were used and sources exploited. For example, Archaeological Science 28:143-151. how was Onondaga chert, specifically, used? Andrefsky, W. Jr. What are the temporal and spatial distributions 1994 Raw Material Availability and the Organization of sites with Onondaga chert? Was this raw mate- of Technology. American Antiquity 59:21-34. rial reduced in an expedient manner, a curated Barbeau, C.M. 1915 Huron and Wyandot Mythology, with an manner, or both, at each site? Once this infor- Appendix Containing Earlier Published Records. mation is in hand, then the correlation between Memoir 80, No.11, Anthropological Series, sites, as well as the correlation between chert Geological Survey, Department of Mines, types, can be examined. To follow our previous . Government Printing Bureau, example, if Onondaga chert is found at a site, Ottawa. what other cherts are commonly found in the Boxt, M.A., and C.L. Reedy same assemblage? By answering these questions, 1985 Preliminary Thin Section Studies of Chert we will finally be able to clearly elucidate larger Artifacts from Northern Belize. Journal of regional patterns of ancient trade and exchange. New World Archaeology 5:13-23. Brantingham, P.J., J. Wolsen, J.A Rech, and A.I. The only limitation of these forms of analysis Krivoshapkin is that they are destructive. Consequently, until 2000 Raw Material Quality and Prepared Core curators and archaeologists alike become willing Technologies in Northeast Asia. Journal of to “sacrifice” their finds and collections, petro- Archaeological Science 27:255-271. graphic and palynological analysis of chert Clough, T.H.M., and A.R. Woolley assemblages will continue to be underused as a 1985 Petrography and Stone Implements. World technique for acquiring information about tech- Archaeology 17:90-100. nological, behavioural and functional activities at Damkjar, E. a site. We must be willing to move beyond these 1990 The Coulter Site and Late Iroquoian Coalescence in the Upper Trent Valley. Occasional Papers in ideas of “destruction” in order to recognise the Northeastern Archaeology 2. Copetown Press, vast amount of information to be obtained Dundas, Ontario. through these investigations. Post-processual Deer, W.A., R.A. Howie, and J. Zussman lithicists now have the means for moving away 1992 An Introduction to the Rock Forming Minerals. from morphology and function-based typologies Longman, New York. towards understanding social and ideological Dewdney, S., and K.E. Kidd structures from lithic assemblages. 1967 Indian Rock Paintings of the Great Lakes. 2nd On the basis of this research we suggest the ed. University of Toronto Press, Toronto. future of raw material sourcing in the Great Lakes Eley, B., and P.H. von Bitter 1989 Cherts of Southern Ontario. Royal Ontario region, and southern Ontario specifically, lies in: Museum, Toronto. the investigation of local and regional chert sources Folk, R.L., and C.E. Weaver and the compilation of these sources into a com- 1952 A Study of the Texture and Composition of prehensive catalogue; the application of chemical Chert. American Journal of Science 250:498- sourcing (e.g., ICP-MS, INAA, XRF) to chert 501. investigations; detailed palynomorph and phy- Fritz-Gaertner, R. tolith analyses for all chert sources; and the appli- 1878 The Preparation of Rocks and Fossils for cation of archaeological context to chert sources. Microscopical Examination. The American Naturalist Illustrated 12:219-225. Gould, R.A., and S. Saggers Acknowledgements. Figure 1 was created by 1985 Lithic Procurement in Central Australia: a Andrew Stewart using Atlas of Canada digital vec- Closer Look at Binford’s Idea of tor data (Natural Resources Canada 2003) and Embeddedness in Archaeology. American chert outcrop locations Eley and von Bitter (1989). Antiquity 50:117-136. Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 27

Herz, N. and Its Spatial and Temporal Distribution in 2001 Sourcing Lithic Artifacts by Instrumental Regional Prehistory. Mercury Series Paper Analyses. In Earth Sciences and Archaeology, 128. Archaeological Survey of Canada, edited by P. Goldberg, V.T. Holliday, and National Museums of Canada, Ottawa. C.R. Ferring, pp. 449-466. Kluwer Julig, P.J., Pavlish, L.A., and R.G.V. Hancock Academic/Plenum Publishers, New York. 1988 INAA of Lithic Materials from the Western Hess, S.C. Lake Superior Region, North America: 1996 Chert Provenance Analysis at the Mack Determination of Raw Material Sources. In Canyon Site, Sherman County, Oregon: an Proceedings of the 26th International Archaeometry Evaluative Study. Geoarchaeology 11:51-81. Symposium, edited by R.M. Farquhar, R.G.V. Kenyon, W., and N.S. Cameron Hancock, and L.A. Pavlish, pp. 233-238. 1961 The Brock Street Burial. Royal Ontario Archaeometry Laboratory, Department of Museum, Art and Archaeology Division, Physics, University of Toronto, Toronto. Occasional Paper 3, pp. 39-54. Royal Kapches, M. Ontario Museum, Toronto. 1994 Hydrogeographics and Transportation Jamieson, S.M. Network Modelling. Paper presented at the 1981 Economics and Ontario Iroquoian Social Chacmool Conference, Calgary, Alberta. Ms Organization. Canadian Journal of Archaeology on file, Department of Anthropology, Royal 5:19-30. Ontario Museum, Toronto. 1984 Neutral Iroquois Lithics: Technological Process Katalin, T. B. and Its Implications. Unpublished PhD 1998 Stones, Numbers-History? The Utilization Dissertation, Department of Anthropology, of Lithic Raw Materials in the Middle and Washington State University, Pullman. Late Neolithic of Hungary. Journal of 1992 Regional Interaction and Ontario Iroquois Anthropological Archaeology 17: 1-18. Evolution. Canadian Journal of Archaeology Lavin, L., and D.R. Prothero 16:70-84. 1992 Prehistoric Procurement of Secondary 1998 The 1997 Trent University Fieldschool at Sources: the Case for Characterization. the Bark Site (BbGp-12), Peterborough North American Archaeologist 13:97-113. County. Annual Archaeological Reports Lentifer C.J., and W.E. Boyd Ontario (New Series) 9:74-77. Ontario 1999 An Assessment of Techniques for the Heritage Foundation, Toronto. Deflocculation and Removal of Clays from 1999a A Brief History of Aboriginal Social Sediments Used in Phytolith Analysis. Interactions in Southern Ontario and Their Journal of Archaeological Science 26:31-44. Taxonomic Implications. In Taming the Luedtke, B.E. Taxonomy: Toward a New Understanding of 1978 Chert Sources and Trace-Element Analysis. Great Lakes Archaeology, edited by R.F. American Antiquity 43:413-423. Williamson and C.M. Watts, pp. 175-192. 1979 The Identification of Sources of Chert Eastendbooks, Toronto. Artifacts. American Antiquity 44:744-757. 1999b Report on the 1997 Trent University 1992 An Archaeologist’s Guide to Chert and Flint. Fieldschool at the Bark Site (BbGp-12). Ms. UCLA Institute of Archaeology, Los on file, Ministry of Culture, Toronto. Angeles. 2000 Report on the 1999 Trent University Martin, F. Fieldschool at the Bark Site (BbGp-12). Ms. 1993 Acritarchs: a Review. Biological Review of the on file, Ministry of Culture, Toronto. Cambridge Philosophical Society 68:475-538. 2002 Report on the 2000 Trent University McGuire, R.H. Fieldschool at the Bark Site (BbGp-12). Ms. 1992 A Marxist Archaeology. Academic Press Inc., on file, Ministry of Culture, Toronto. San Diego. 2004 Trent University 2000 Field School. Annual Miles, K.M. Archaeological Report Ontario (New Series) 2005 Chert Lithic Assemblage Utilization and 12:64-74. Ontario Heritage Foundation, Availability at the Bark Site, BdGp-12, Toronto. Ontario. Unpublished MA thesis, Department Janusas, S. of Anthropology, Trent University, 1984 A Petrological Analysis of Kettle Point Chert Peterborough. 28 Ontario Archaeology No. 81/82, 2006

Moorhouse, W.W. Rankin, L.K. 1959 The Study of Rocks in Thin Section. Harper 1998 Historical Context and the Forager/Farmer and Row, New York. Frontier: Re-Interpreting the Nodwell Site. Morrow, C.A., and R.W. Jefferies Unpublished PhD dissertation, Department 1989 Trade or Embedded Procurement? A Test of Anthropology, McMaster University. Case from Southern Illinois. In Time, Renfrew, C., and P. Bahn Energy and Stone Tools, edited by R.Torrence, 1996 Archaeology: Theories, Methods and Practice. pp. 27-33. Cambridge University Press, 2nd ed. Thames and Hudson, London. Cambridge. Shelley, P.H. Natural Resources Canada 1990 Variation in Lithic Assemblages: an 2003 National Scale Frameworks Hydrology- Experiment. Journal of Field Archaeology Drainage Network, Canada. Digital vector 17:187-193. data. Scale 1:1,000,000. Ottawa. Sheppard, P.J., and L.A. Pavlish http://geogratis.cgdi.gc.ca 1992 Weathering of Archaeological Cherts: a Case Parkins, W.G. Study from the Solomon Islands. 1974 Source of Chert from Welland River Geoarchaeology 7:41-53. Archaeological Sites. Unpublished BSc thesis, Sieveking G. De G., P. Bush, J. Ferguson, P.T. Department of Geological Sciences, Brock Craddock, M.J. Hughes, and M.R. Cowell University, St. Catharines. 1972 Prehistoric Flint Mines and their Pendergast, J.F. Identification as Sources of Raw Material. 1985 Huron-St. Lawrence Iroquois Relations in Archaeometry 14:151-176. the Terminal Prehistoric Period. Ontario Smith, E.A. Archaeology 44:23-40. 1983 Myths of the Iroquois. Reprinted. Iroqrafts Pollock, S.G., N.D. Hamilton, and R. Bonnichsen Ltd. Ohsweken, Ontario. Originally pub- 1999 Chert from the Munsungun Lake Formation lished 1883, Smithsonian Institution Bureau (Maine) in Palaeoamerican Archaeological Sites of Ethnology Second Annual Report 1880- in Northeastern North America: Recognition 81:5-83. Washington, D.C. of its Occurrence and Distribution. Journal of Sutton, R.E. Archaeological Science 26:269-293. 1990 Hidden Amidst the Hills. Middle and Late Prothero, D.R., and L. Lavin Iroquoian Occupations in the Middle Trent 1990 Chert Petrography and its Potential as an Valley. Occasional Papers in Northeastern Analytical Tool in Archaeology. In Archaeology 3. Copetown Press, Dundas, Archaeological Geology of North America, edited Ontario. by N.P. Lasca and J. Donahue, pp. 561-584. Sweetman, P.W. Geological Society of America, Boulder. 1955 A Preliminary Report on the Peterborough Ramsden, P.G. Petroglyphs. Ontario History 47:101-121. 1977 A Refinement of Some Aspects of Huron Tankersley, K.B. Ceramic Analysis. Mercury Series Paper 63. 1989 A Close Look at the Big Picture: Early Archaeological Survey of Canada, National Paleoindian Lithic Resource Procurement in Museum of Man , Ottawa. the Midwestern United States. In Eastern 1988 Palisade Extension, Village Expansion and Paleoindian Lithic Resource Use, edited by Immigration in Trent Valley Huron Villages. C.J. Ellis and J.C. Lothrop, pp. 259-292. Canadian Journal of Archaeology 12:177- Westview Press, Boulder. 183. Vastokas, J., and R. Vastokas 1990 Saint Lawrence Iroquoians in the Upper 1973 Sacred Art of the Algonkians: a Study of the Trent River Valley. Man in the Northeast Peterborough Petroglyphs. Mansard Press, 39:87-95 Peterborough, Ontario. 1996 The Current State of Huron Archaeology. von Bitter, P.H., and B. Eley Northeast Anthropology 51:101-112. 1984 Chert From the Upper Member Ramsden, C.N. of the Middle Ordovician Bobcaygeon 1989 The Kirche Site. Occasional Papers in Formation of Southern Ontario. Canadian Northeastern Archaeology 1. Copetown Journal of Archaeology 8:135-148. Press, Dundas, Ontario. 1997 Fossil Hill Formation Chert at the Fisher Biittner and Jamieson Chert Raw Material Utilization at the Bark Site 29

Site: Geological Source and Significance. In 2000 The Precontact Iroquoian Occupation of The Fisher Site: Archaeological, Geological Southern Ontario. Journal of World and Paleobotanical Studies at an Early Paleo- Prehistory 14:415-458. Indian Site in Southern Ontario, Canada, Winchell, A.N. edited by P.L. Storck, pp. 223-235. Memoir 1965 Elements of Optical Mineralogy: An 30. Museum of Anthropology, University of Introduction to Microscopic Petrography. 3 Michigan, Ann Arbor. vols. John Wiley and Sons, Inc., New York. Wahlstrom, E.E. Writt, R. 1955 Petrographic Mineralogy. John Wiley and 1975 The Descriptive Classification of Silurian Sons, Inc., New York. Cherts of the Niagara Peninsula. Warrick, G. Unpublished BSc. thesis, Department of 1984 Reconstructing Ontario Iroquoian Village Geological Sciences, Brock University, St. Organisation. Mercury Series Paper 124. Catharines. Archaeological Survey of Canada, National Museum of Man, Ottawa.

Afin de comprendre la façon dont certaine variété de silex était utilisé dans la fabrication d’outils de pierre au site Bark (BpGp-12), Comté de Peterborough, Ontario, les attributs des matières premières de 164 artefacts de l’ensemble d’objets de silex sont analysés. Une combinaison d’analyse macro- scopique, pétrographique et palynologique est utilisé afin d’identifier la provenance des types de silex les plus utilisés par les habitants du site. Les formations Onondaga, Fossil Hill et Upper Gull River sont identifiées. Nous concluons, contra Eley et von Bitter (1989), que l’identification du genre acritarche n’est pas nécessaire pour différencier entre ces types de silex. Au contraire, des caractéris- tiques plus significatives sont l’abondance relative de différents sous-groupes palynomorphes (basé sur la morphologie), et l’abondance relative d’acritarches comparativement à d’autres microfossiles (spore de chitinozoan, de graptolite, de scolécodonte, et de trilète). L’attribution du silex à des sources loin- taines nous permet d’explorer des thèmes comme l’accès aux matériaux exotiques de haute qualité par des groupes parentés locaux.

Katie M. Biittner Department of Anthropology 13-15 Tory Building University of Alberta Edmonton, Alberta, Canada T6G 2H4 [email protected]

Susan M. Jamieson Department of Anthropology Trent University Peterborough, Ontario, Canada K9J 7B8 [email protected]