«By James R. Keron Graduate Program in Anthropology Submitted in partial fulfillment of the requirements for the degree of Master of Arts Faculty of ...»
To the east there are a number of chert bearing formations that outcrop along Lake Erie. One source that appears in low frequency on Iroquoian sites in the study area is Selkirk chert from the Dundee Formation which is middle Devonian in age. Primary outcrops of this chert occur between Port Dover and Dunnville, both inland near Selkirk and along the lake shore (Eley and Von Bitter 1989). The Onondaga Formation is also Middle Devonian and grades laterally into the Dundee Formation as one proceeds west from Dunnville. The Dundee Formation outcrops bearing Selkirk chert are slightly younger than the chert bearing portions of the Onondaga Formation. The chert bearing portion of the Onondaga Formation outcrops along Lake Erie from a point south of Selkirk to Fort Erie and on into New York State (Eley and Von Bitter 1989). These outcrops are particularly rich with high quality chert making up a large percentage of chert versus the rock formation as a whole (Parkins 1977). The third chert occurring in this same general area is Haldimand or Bois Blanc chert from the Bois Blanc Formation which is lower Devonian in age (see Parker 1986a, 1986b). These outcrops are found north of the Onondaga outcrops in the Dunnville area. This chert type is also found in small quantities in the study area. It is generally a lower grade chert.
The other major source of chert is secondary deposits of chert that occur in the glacial till throughout the study area. A study by the Department of Highways of Ontario (Ingram and Dunikowska-Koniuszy 1965) provides some quantification of the amount of this occurring in the till by measuring the quantity of chert in gravel deposits within the area and concludes that for most of the western part of the study area chert comprises 5of the aggregate. In the eastern part of the study area in North Dorchester Township, this rises to 10-20% of the aggregate. In the western part of the study area in the Caradoc sand plain, the fine sand obscures much of the chert except in the creek and river beds.
The basic problem is that, with any of this chert, it is not possible to identify the original source as there are no outcrops directly in or near the study area. Further, glacial action can move blocks of chert around in a confusing pattern and mix up chert from several different sources. At the end of the last glaciation the ice action left a series of recessional moraines where the ice movement was from south to north out of the Lake Erie basin. However, there are also large erratics of granite that were transported from the Canadian Shield far to the north. Luedke (1976) notes chert tends to break down quickly under glacial action so that it can be assumed that any sizable fragments of chert have not been moved too far from the bed rock from which they were derived.
The look and feel of this till chert can be highly variable. Some of it is quite distinctive and clearly different from any of the primary deposits. Of note is a cream coloured chert occurring through Westminster and North Dorchester Townships. There are also several other distinctive varieties as well. However, some of it is virtually identical to primary source Onondaga chert and short of trace element analysis or thin sectioning, it is impossible to tell the difference. This Onondaga-like till chert could be derived from the Dundee Formation that is Middle Devonian like the Onondaga Formation. Indeed, it is reasonable to assume that the processes that led to the significant chert deposits in the Onondaga Formation were also acting on the contemporary Dundee Formation that underlies the study area and much of southwestern Ontario. However, there are no surface outcrops in this area as the bedrock is buried by Tertiary/Quaternary deposits. Whether this included the colouration or whether the similar colour is an example of different formations producing macroscopically identical chert is a mute point. In any event, there is chert occurring in the till especially to the south of the London area which is virtually identical to primary source Onondaga thus making for great difficulty in macroscopically identifying any given piece. Unfortunately, while the local till chert is readily available it is generally of an inferior quality and frequently shatters with attempts to work it.
In the study area per se this local Onondaga-like chert is generally small and well ground by the glacial action. However, along the Lake Erie shoreline, larger cobbles can be found that are indistinguishable from primary source Onondaga except for the worn cortex. Christopher Ellis (personal communication 2003) has observed such cobbles along the shore line south of the Nettling site. Also William Fox (personal communication 2003) notes the presence of blocks of AOnondaga@ chert as far west as Pelee Island. This chert is so much like the primary source material that he hypothesized that it was moved from the primary sources to the east down Lake Erie to its current location that far west. While this is well beyond the range of chert survival in glacial movement suggested by Luedtke (1976), Onondaga chert may be exceptional given the massive amounts of chert in the beds. In effect, there is so much chert in large boulders that some of it survives the distant transportation in blocks large enough to be useful.
Alternatively, it could well be derived from Dundee Formation deposits that were subject to glacial action on the bed of Lake Erie and in the far past in the study area itself. On the south side of Lake Erie Onondaga-like chert is found in the glacial tills of Ohio (Christopher Ellis: personal communication 2003). In any event, the actual source of this is a geological problem and need not concern us here. From the archaeological perspective, the chert is identical to primary outcrops of Onondaga and could thus cause problems in identifying it as to source. In conducting the analysis, it was not unusual to find material classified as Onondaga but which has a clearly water rolled cortex suggestive of lake shore action. It is also quite probable that water rolled cobbles were being collected and utilized in the vicinity of the primary outcrops along Lake Erie as well (William Fox personal communication 2003) Chert Source Identification This section describes the methodology used in this study to assign various artefacts to a chert source. A systematic approach to this is critical in order to provide consistency to the data to be used in evaluating the behaviourial hypotheses outlined above. In conducting a macroscopic analysis, the specific attributes of the various source types are, to a large extent, entirely visual making it difficult to work from detailed textual descriptions. Chert source identification is more of an oral tradition in the archaeological community where the elders are the keepers of the accumulated knowledge and gradually train the next generation in how to recognize the various source material. The most exact approach is to have a good collection in front of you as you work through assigning source types and thus, gradually knowledge is built that can stand on its own. Therefore, while some characteristics are defined below, it is not expected that the definitions alone are sufficient to facilitate assignment of any given artefact to a particular source. Despite the presence of the oral tradition, there is still a significant variance from one analyst to another in what is identified as to which source and to this end, a comparative analysis was done comparing the typing of 200 flakes amongst six different analysts (Keron 2003). That study showed that there is reasonable agreement on identification of Kettle Point chert but the distinction between Onondaga and local till chert is highly variable from one analyst to the next. This problem arises because primary source Onondaga and local till chert in the London area occur in continuous shades of grey and brown making macroscopic distinction a very error prone activity. Indeed, it has been argued elsewhere (Keron 1986) that this continuum is such that attempting to make the distinction is not worth the effort. This same approach was adopted by Timmins (1997) and Lennox (1995a). In fact, Lennox (1995a:94) claims that local till cherts are indistinguishable from primary source Onondaga and argues the material which resembles Onondaga is derived from local till chert and that easy access to local till chert means Iroquoian groups would not have gone to the trouble of acquiring primary source Onondaga.
Despite the difficulty in macroscopically distinguishing primary source Onondaga from local till chert, the distinction between these two cherts is very important to the social components of Iroquoian chert acquisition. If primary source Onondaga is being acquired, the social component of the acquisition is much different and certainly more complex than that of local till chert. Further, the local groups did go to the trouble of acquiring Kettle Point chert from some distance and most sites also have an low percentage of Selkirk chert which from the author=s experience in not easily confused with local till chert. As the source for this is close to the source for Onondaga, there is every reason to believe that the same processes that acquired Selkirk chert could be used to acquire primary source Onondaga. When it comes to comparing material from published reports of different analysts, local till chert should be lumped with Onondaga as there is too much variation between analysts to make any comparison remotely valid.
However, given the importance of the distinction, this thesis will attempt to separate the two cherts but only material examined by the author will be used to allow some level of consistency.
The first way of drawing the distinction is if there is a portion of cortex present that was battered indicating it came from a secondary deposit (Keron 1986). However, a second way of making the distinction is to observe the macroscopic characteristics of natural chert cobbles known to be derived from the local till. In fact, the author normally collects samples of this material when doing survey. What has emerged from this is that there is a class of local till chert that is easily identified even when there is no cortex present. That leaves a third class of till chert that is certainly difficult to distinguish from primary source outcrop Onondaga. Distinguishing these two sources is a definite problem. After several false starts, it became apparent that the only way to maintain consistency in sorting was to work from a reference collection. To do this a number of artefacts from earlier time periods (e.g. Archaic) were selected that were derived from primary sources. Then, when doing the data collection, unless a particular piece matched at least one of the sample, it was not counted as Onondaga chert. This method has proved to be workable and while the assignment of a few pieces might be questionable, on the whole, the method is reliable. Finally, as colour hues are sometimes critical to making the distinction, care was taken that the same lighting system was used throughout the entire analysis.
Following is the list of the chert sources identified during data collection as well as a short description of the visual cues used to make the assignment.
Kettle Point This chert can be reddish, greenish or even grey. It tends to be very lustrous and is translucent most of the time. This translucency is a key attribute in making the assignment.
Onondaga As used here, the assignment of a particular artefact as Onondaga chert was based on it matching one of a known sample. Onondaga chert varies from grey to a grey/brown mottled appearance. It is usually fine grained and only slightly translucent or not at all. The grey brown mottling varies from light to dark.
Local Till Chert Local till chert can be highly variable in both colour and quality.
Some of it can be quite distinctive in colour from Onondaga and the grain is generally coarser. The cortex can be crushed and battered but is frequently composed of conchoidal fractures making easy distinction between natural pebbles and cores difficult except for a slight rounding of the ridges. Indeed, one site examined (intentionally not referenced) in the study lists five cores in a widely published source. On examination all of these Acores@ were determined to be natural chert cobbles without any human modification. Another site (also not referenced) analyzed had a large number of Acores@ Awedges@ Aspokeshaves@ and even one Abiface@ cataloged that were all natural chert pebbles. Colours can vary from a cream brown through to a grey that is almost identical to that found in primary source Onondaga chert. While some of the material can be finegrained most is generally coarse leading to irregular fractures if one attempts to work it.
Despite this coarse nature some of it is fine enough to permit bifacial working, most likely pieces that have not been as badly battered during the glaciation process that led to their deposition.
Other Chert This category includes all chert sources other than those named above. Selkirk chert is included here as are any other chert sources that are clearly foreign to the area (e.g. Flint Ridge, Collingwood, Bayport chert). These sources are a very small percent of the total in all sites. Included in this category are any flakes which are clearly not one of the above but where the source in not known.
Unidentified Included in this category are all pieces that could not be identified as to source due to, generally, two reasons. Frequently debitage has been burned. making identification problematic. In some cases after working with enough material a source assignment might be made, for example the coloration of burned Kettle Point chert might be consistent enough to be identified. However, in order to maintain consistency this material is placed into the unidentified category as the distinction only became evident half-way through the analysis. The other reason to classify a piece as unidentified is when the flake is too small to allow reliable identification.
Site Information With respect to gathering data for the study about the specific sites, the first step was to build an inventory of Iroquoian sites in the study area. This inventory was developed using a number of sources. The first draft was assembled using published material (Keron 1986; Pearce 1996; Williamson 1985). These data were then augmented through a request for information from the database of the Ministry of Culture, Heritage Branch. Unfortunately, due to other time constraints within the Heritage Planning Branch, it was not possible to get a complete list of all Iroquoian sites in the study area.