«By James R. Keron Graduate Program in Anthropology Submitted in partial fulfillment of the requirements for the degree of Master of Arts Faculty of ...»
Out of this inventory, a number of sites were selected for analysis.
The following information was kept for each site in the site inventory. Appendix A details all of the sites used in the analysis and also contains the specific publication references for each site.
1. Site name
2. Borden Number
3. Location (UTM coordinates)
4. Cultural affiliation
5. Site Type
6. Cluster assignment
7. Comments on the nature of the sample and its representativeness.
8. References The cultural affiliation of the site was normally that assigned by the investigator.
Categories used here were Glen Meyer, Uren, Middleport and Neutral, basically following Wright=s (1966) stages. While actual dates would have been preferable, this would have meant using radio carbon dates and that would have severely limited the number of sites that could be used. Dates assigned using artefact styles are normally just derived from Wright=s (1966) stages anyway and, at the level of analysis used here, the stages are a good ordering device and have proved to be reasonable over the years.
With respect to site types, research in the area has demonstrated that there is a variety of site types throughout the Late Woodland period in southwestern Ontario.
Williamson=s work (1985) in Caradoc has demonstrated that this pattern arose early.
Similarly, Pearce (1996) has demonstrated the same pattern in the Neutral period in the vicinity of the Lawson site as has the author (Keron 1986) in the Pond Mills and Lambeth vicinity. Terms such as villages, hamlets, special function sites and cabin sites have been used but unfortunately not with any kind of consistency. For quite a while the term Ahamlet@ was applied to any small site and Avillage@ was applied to the larger sites (c.f.
Pearce 1996; Williamson 1985). More recently the term Acabin@ has been used for the small special function sites that are frequently assumed to be for horticultural purposes.
Pearce (1996) in the addendum to his thesis reviews the definitions of village, satellite, hamlet and cabin that were given in Finlayson and Pearce (1989) and modifies that of cabin to include one or two long houses. Those definitions will be adopted here. In operationalizing these definitions it should be noted that they depend to a large extent on classifying a site which has been extensively excavated exposing the patterns of longhouses. Here again, restricting the study to excavated sites would severely limit the number of sites that could be analyzed. In practice, the terms defined by Finlayson and Pearce (1989) will be used but the assignment to the terms will be based primarily on the site size and density of artefacts. While this practice will introduce an element of error into the results, in general the extremes of the large/small dichotomy are almost certainly accurately assigned. The potential errors occur with sites in the middle. These could well be satellites in the sense used by Finlayson and Pearce (1989) or possible hamlets or small villages. Nonetheless, accepting this risk greatly increases the number of sites that can be included.
For cluster, the assignment of sites to a cluster generally follows those used by Pearce (1996; see Figure 3). However, Pearce=s Dingman community is divided into two groups labeled Pond Mills and Lambeth. Clusters are assumed to be reflective of a single community occupying a territory through time where successive villages were built as the group relocates through time in the classic Iroquoian pattern as has been demonstrated for some areas (e.g. Tuck  for the Onondaga and Pearce  for the London community of the prehistoric Neutral). In reality, group movements can be quite complex so that heavy dependence on the cluster assumption without detailed inter-site comparisons can be somewhat risky. However, the concept of cluster has been retained more as a descriptive device and, to a large extent, the cluster assignment does not figure heavily in the analysis. What is difficult, as noted earlier, is to connect the Glen Meyer clusters with the succeeding Middleport sites. Pearce (1984) theorizes that the Arkona, Caradoc, and Byron clusters coalesced to form the Middleport sites on Oxbow Creek and that the Dorchester cluster was antecedent to the Lake Whitaker cluster. However, another equally valid interpretation would link the Byron cluster and its associated community with the Lambeth community while the Dorchester cluster could be antecedent to the Pond Mills cluster. Then, of course, there is the conquest theory (Finlayson 1998; Wright 1966, 1992) where the Uren Substage is a cultural replacement of Glen Meyer so any attempt to connect Glen Myer with later sites is a non-issue. In any event, the cluster construct will be retained but does not figure in the analysis and care should be used in reading too much into it.
Finally, the nature of the lithic sample from each site is of utmost importance. In order to establish the validity of the conclusions here, it is necessary that the material used be reflective of the real percentages of various artefact types and chert sources used at the site. The only sure way of doing this would be to excavate the entire site (including screening all the topsoil) and then to analyze all the material or at least a statistically representative sample of it. This procedure is not possible as few sites in the study area have been completely excavated. Thus, it is necessary to consider to what extent the sample from a given site would be representative. The best sample would be a CSP of the entire site as it would be representative, lend itself to spatial analysis, and not be too time consuming. Excavated samples run the risk of not being representative unless the excavation covered a large part of the site or covered most areas of the site. Comments as to the representativeness of a site are included in Appendix A.
Artefact Information The data from each site were entered into a Microsoft Excel spreadsheet. This spreadsheet contained the debitage data in the form of totals for all combinations of chert source and flake type in both raw counts and in percentages for the dominant five flake types. Further a summary of all the artefact types broken out by chert source was also included. These data are presented in Appendix F. In addition, various qualitative comments on the assemblage were kept in this spreadsheet.
For the sites selected to conduct intra-site spatial analysis, a second Excel spreadsheet was created that had one row per individual artefact. This included the following information.
2. Chert Source
3. Artefact Type
4. Artefact SubType
5. Provenience (internal location in Cartesian Co-ordinate form) Frequently, it was necessary to convert transit and compass readings into Cartesian Coordinates. These were all done using the spreadsheet formula from the London Chapter, OAS, web site (Keron and Prowse 2001). The data in the appendices are in Cartesian Coordinate format.
The artefact types and associated codes used in Appendix F are those defined by the London Museum of Archaeology (again available on the London Chapter OAS web site). However, some additional types were added or the meaning was used differently as follows.
Foliate Biface These bifaces are assumed to be knives. See Fox (1981e, 1982d) for a detailed description.
Preform This is a type of biface in a state that was antecedent to creating a projectile point. To be put in this type, it must be clearly intended for a projectile point.
Used Cobble Given the amount of local till chert available to the Iroquoian people, it is not surprising that occasionally an unmodified piece of till chert would be put to use. To be included in this category the use must be clearly evident and purposeful. A scratch with a plough does not qualify as use.
Additionally, subtypes were kept for projectile points. It should be noted here that projectile points from earlier sites are frequently found on Iroquoian sites. While some of these could be the result of a multiple occupations they are frequently interpreted as resulting from Iroquoian collection on earlier sites. In this analysis these earlier points are ignored.
Notched This classification indicates a notched Iroquoian projectile point. The classic type for this would be Nanticoke Notched (Fox 1981d). Another term sometimes used is Middleport Notched but this has never been formally defined and is best considered an earlier form of Nanticoke Notched.
Triangular During the late period, this category would be termed Nanticoke Triangular (Fox 1981c). There are also triangular forms during Glen Meyer times.
These tend to be larger and more variable. Some fit the description of Levanna (Ritchie 1971) while a variant called Glen Meyer Tanged Triangular has been identified by Fox (1982a).
Tip and Mid-section These are broken fragments where the base is missing prohibiting further typing.
Scrapers were broken down into four types as follows.
SCR This is a classic end scraper but lacks ventral retouch, Side SCR This is a scraper where the working edge is on a lateral margin.
SCR-VR This is a classic end scraper but possesses ventral retouch SCR-flake The line between what would normally be called a utilized flake and this category is somewhat blurred. These are flakes with retouch along one or more margins. The artefact has not been purposely shaped and the retouch might very well be through use. It is however, quite definite and not problematic as are some of the artefacts that are frequently cataloged as utilized flakes.
Ventral retouch is an interesting treatment frequently applied to end scrapers by Iroquoian people. It is found on a number of late prehistoric Neutral sites and even into the historic period (Lennox 1981:242). It varies from a few flakes purposefully removed from the ventral surface to the point where the entire ventral surface of the flake has been removed, in effect, turning it into a biface. These are referred to as bifacial snub nosed scrapers by Lennox (1981). In this analysis, an artefact is assigned to this type if it possesses ventral retouch. No attempt is made to differentiate types based on the amount of retouch.
Finally a commentary should be made regarding utilized flakes. In this analysis, this so-called type has been excluded. The initial concern was that it would be better to assign the flake to one of the defined flake types rather than to a unique type simply because it has been used, as this would obscure the reduction data. In practice, while working through the various collections cataloged by other researchers, a very healthy scepticism developed as to how consistently this category was applied. Some flakes were cataloged as Autilized@ when there was no use-wear evident. Some flakes were cataloged as Autilized@ that had damage to one edge that, at least to the author, appeared most likely to have been plough induced. And some flakes did, indeed, appear to be used. These three tend to occur in about equal frequency. Yet another kind of Autilized flake@ was a natural chert cobble with some form of plough damage. Further, in working through the debitage or Aunutilized flakes@, it was common to encounter flakes that clearly looked Aused@ but for some reason had not been recognized as such. In general, for the author at least, any future archaeological report that discusses Autilized flakes@ will be greeted with a high degree of scepticism. I agree fully with Shen (1999) who concluded that the term Autilized flake@ should be completely abandoned as a type. Use-wear analysis is still a very viable tool in our repertoire but as Shen (1999) notes, needs to be conducted with a microscope by a qualified use-wear analyst. Including a Autilized flake@ type without this level of analysis is highly problematic.
Approaches to Debitage Typologies Critical to development of this thesis is the comparison of debitage from a number of sites in the study area. While the post-processual mantra that Adata are laden with theory@ has been used to justify chaos in archaeological analysis, the criticisms inherent in this perspective have some justification so it is necessary to give some thought to the analytical procedures used to create the comparative data.
Archaeologists have approached the subject of debitage analysis in many ways depending on the particular analytical problem at hand as well as their own predilections.
One of the earliest rigorous analysis of debitage from North American sites was that done by Anta White (1963). In this work, she describes a series of flake types which result from the stages of reduction and artefact finishing. Since then, numerous variants of this type approach have been used and other alternative techniques developed. For purposes here, these are classified into three different methodological approaches: lithic reduction stage typologies, Sullivan and Rozen’s (1985) typology (SRT), and mass analysis.
The stage typology method, following White_s (1963) analysis, is used most often and involves defining a series of types deemed to be diagnostic of the reduction sequence based on clusters of flake attributes. There are a number of different typologies that have been used over the years and these vary from only a few types (e.g. Lennox 1986) up to a much larger number (e.g. Ellis 1984). White (1963) identifies two types associated with decortication, six types associated with core reduction and four with artefact production.
The types associated with artefact production are not as clearly defined, being limited to the method of removal (e.g. percussion, pressure flaking or steep marginal retouch). In general, all these typologies share the same basic assumption, that flake types can be associated with the stages of stone working varying from the initial flakes struck from a nodule through various core preparation and reduction, to bifacial and unifacial tool finishing, and finally to tool resharpening and reworking and that these types are recognizable. Thus, it is possible by looking at the frequencies of the various types, to determine some of the activities that were conducted at the site and even make inferences about the mobility patterns of the inhabitants. For instance a high degree of bifacial thinning and a low frequency of core reduction is indicative of Binford_s (1980) pattern of curation associated with the collector strategy and when used with chert source identification leads to inferences regarding group mobility. (e.g. Andrefsky 1994; Ingbar, 1994; Morrow and Jeffries 1989).