The Sourcing of Chert Artifacts by INAA:
Some Examples from the Great Lakes Region

Dr. Patrick J. Julig PJULIG@LAUADMIN.LAURENTIAN.CA

Department of Sociology and Anthropology, Laurentian University, Sudbury, Canada, P3E 2C6

Submitted: Mon, 22 Aug 1994

ABSTRACT

Instrumental neutron activation analysis (INAA) is used for the chemical analysis of archaeological and geological chert samples and to assist in determining the sources of prehistoric artifacts. This paper summarizes INAA analysis of some Great Lakes chert sources and artifact assemblages that could not be reliability sourced by traditional macroscopic or microscopic visual means. Although INAA is not yet a common method for chert analysis, it has proven useful for the sourcing of some visually similar archaeological cherts, particularly in situations where the geological sources are spatially or temporally well separated. In addition, INAA with the SLOWPOKE reactor for the determination of the short half-life producing elements is an appropriate non- destructive method for the analysis of whole artifacts, which can then be returned to their curators undamaged.

INTRODUCTION

A primary classification required in the analysis of lithic artifacts is determining the type or source of the raw material from which the specimen is manufactured. Archaeologists readily gain familiarity with the common types of lithic materials used by the prehistoric inhabitants of sites within a region, and the major primary and secondary geologic sources. Normally lithic identifications are made by visual inspection, either macroscopic comparison with reference samples, or microscopic identification of thin-sections using petrographic methods. However in many assemblages reliably identified by macroscopic or microscopic visual features. Also, if specimens are small flakes or too rare for thin- sectioning for petrographic analysis (such as projectile points), other lithic experts are often consulted, who may have special knowledge of cherts from suspected source region. Some widely distributed lithic materials are visually quite unique, for example Fossil Hill Formation chert from the Georgian Bay area of Ontario, and positive visual identification can be readily made from fossil inclusions and other features (Storck and von Bitter 1989). However this is not normally the case, and archaeological site reports commonly list the small percentages of lithic materials that cannot be identified in summary tables as "other" or "unidentified". Often visual identifications are made of distant imports or "Exotics", and these (sometimes tentative) identifications may then form the basis for reconstruction of trade mechanisms and/or population movements (e.g. Gramly 1988; Clark 1984). Since many anthropological archaeology research questions such as territory or band size, trade patterns and social interaction depend on correct provenance data it is important that appropriate methods be utilized. When examining such research questions some form of chemical analysis may be necessary to test the visual identification of lithic artifacts. In addition to INAA other available methods include atomic absorption spectrophotometry (AAS) and inductively coupled plasma atomic emission spectroscopy (ICP). The question of which method to use in a provenance problem involves various considerations, but probably the most important is the need for a team approach with someone who has analytical experience with the technique and the limitations. Other important considerations include time constraints, the element or compound range, number of samples and the accuracy desired from the analysis. There are advantages and limitations to most analytical techniques and the one chosen depends on the specific needs of the research, however, INAA still remains one of the better methods for highly sensitive, cost-effective, quantitative, multi-element analysis of archaeological materials. This work will discuss select examples of the application of INAA for the analysis and sourcing of chert artifacts in the Great Lakes and Upper Midwest region. Rather than providing the geochemical data on specific chert analyses the reader will be referred to previous publications. I will also review the basic methodology, and the advantages of the SLOWPOKE analytical reactor for the analysis of whole artifacts.

A BRIEF REVIEW OF INAA ANALYSIS OF CHERT

Although samples may appear visually homogeneous, chert is a rather heterogeneous lithic material from a chemical standpoint. INAA of archaeological materials began with the chemical analysis of pottery (Perlman et al. 1972; Sayre and Dodson 1957), and the successful analysis of archaeological chert materials developed subsequently (DeBruin et al. 1972; Ives 1974, 1984; Leudtke 1978, 1979; Tite 1972). INAA can be used to determine the major (>1%), minor (>0.1%), and trace (<0.1% [1000 ppm]) elemental constituents, which can be used to chemically characterize both archaeological and geological chert. The works mentioned above demonstrated that a suite of elemental data derived from INAA can be used to solve archaeological provenancing problems, despite the complications introduced by chemical variations and the heterogeneous nature of chert.

INAA analysis to differentiate and source chert and other cryptocrystalline silicates has been conducted on assemblages in the Great Lakes and adjacent regions (Luedtke 1978, 1979; Hoard et al. 1993; Jarvis 1990; Julig et al. 1987, 1988, 1989, 1991a,b, 1992; Goatley 1994), however, the routine sourcing of chert by INAA is not yet too common. Luedtke (1978; 1979) chemically characterized a large suite of Great Lakes region cherts, and demonstrated that chemical differences were generally greater between chert bearing geological formations then within them. The more recent work of Julig et al (cited above) and Hoard et al. (1993) have supported this general principle. We have had reasonable success in chemically "fingerprinting" and sourcing of visually similar materials, from widely separate sources, rather than the differentiation of sources close in space and/or time (outcrops from the same geological formation), which are often chemically more similar. However a study of Onondaga chert by Jarvis (1990) found strong chemical separation between chert from different outcrops of the same formation (Devonian), and attributed the differential elemental concentrations to facies shift. However, in that study the discriminate analysis classification for the geological samples did improve with distance between sources.

With further testing and characterization by INAA of chert proceeds, we now have the beginnings of a chemical data base to assist in the resolution of provenance problems of certain visually similar materials (Goatley 1994, Julig et.al. 1992); as will be discussed later.

METHODOLOGY AND SAMPLING FOR INAA

This section will discuss the methodology used in the Julig et al research referenced previously. The basic research strategy employed in INAA sourcing of lithic artifacts involved: 1) chemically characterizing the archaeological unknowns, 2) likewise chemically characterizing prospective geological sources, and 3) attempting to match the unknowns to the tested geological sources. In this manner potential sources are eliminated if they are shown to be chemically very different from the artifacts. However even when close chemical "matches" are found, one can never be certain that all of the potential sources have been tested.

The characterization of the geological sources normally involves testing a minimum of 30 samples ( except in some small pilot studies), these being selected from various parts of the formation. Ideally they should include the range of variability within the formation, with respect to the knappable chert. This part of the research is obviously critical, yet very difficult to conduct unless there is good access to the geological exposures and the primary and secondary deposits are well mapped. In the Great Lakes region the chert deposits are often covered by glacial deposits as well as being distributed by Pleistocene glaciation. Some sources, such as Hudson Bay Lowland chert (HBL) have primary sources in remote regions, and secondary sources widely dispersed by the Pleistocene glaciation (Julig et al. 1992); this makes the geological sampling and chemical characterization difficult.

Geological samples in the range of 200 to 600 mg were used for INAA, while archaeological samples varied in size from 50 to 10,000 mg (in the case of some whole artifacts). The geological samples are obtained by flaking chert cobbles with a moose antler billet and then snapping off appropriate sized flake fragments, washing them in distilled water, air drying and then sealing in polyethylene vials for irradiation after being weighed. It is important to avoid contamination by flaking with metal implements in the sub-sampling procedure.

The samples were analyzed at the SLOWPOKE Reactor Facility of the University of Toronto. The individual samples in sealed vials were irradiated for five minutes at a neutron flux of 1.0 x 1011n.cm-2.s-1. After a delay time of 15 to 20 minutes to allow the 28Al to decay to analytically acceptable levels, the samples were assayed for 5 minutes using Ge detector gamma ray spectrometers. This allowed the measurement of fifteen elements including U, Dy, Ba, Ti, Sr, Br, Mg, Si, Na, V, K, Al, Mn, Cl, and Ca, the majority of which were present in measurable concentrations in most samples. The analytical precision ranged from <1% to detection limits.

Some clean chert and quartz samples that were analyzed (as part of other sourcing studies, e.g. Julig 1987) were irradiated as much as five times as much as normal to produce enough radioactivity to allow the measurement of their lower concentrations of trace elements. Elemental concentrations were calculated using the comparator method (Hancock 1978). This analytical procedure produced chemical data that allowed for the discrimination of different chert sources.

The research discussed here used only the short-lived radioisotopes noted above, and although valuable chemical information may be obtained in the long-lived radioisotopes, this information was not required for this research. An additional advantage of only analyzing for the short-lived radioisotopes is that small and valuable whole artifacts may be analyzed and then returned to the collections from where they were borrowed. Larger size vials were used in this procedure, as reported in Julig et.al.(1992).

DISCUSSION

Several examples of chert discrimination and sourcing of artifacts will follow. For details of the precise locations of sites, numbers and characteristics of the samples and the elemental data, the reader is referred to the previous references. It should be noted that this research initially developed out of rather modest pilot projects with moderate size suites of samples and was mainly problem oriented, that is, determining the source(s) of specific artifacts. If chemical discrimination was obtained between visually similar cherts in the initial set of analyses, samples sizes were expanded, additional artifacts were tested, and the chemical data base on the geological chert sources expanded. Not all pilot projects proved successful, particularly those involving source discrimination from within the same formations.

The initial INAA analysis of chert artifacts involved a problem of sourcing dark to medium brown chert and/or agate artifacts from the Cummins Palaeoindian site near Thunder Bay Ontario. These visually similar artifacts consisted of small biface thinning and scraper resharping flakes, and small scrapers (Julig 1988, Julig et.al. 1988, 1989, 1991a). Three possible geological sources that were visually very similar were tested by INAA including Knife River Flint (KRF) from western North Dakota, Hudson Bay Lowland chert (HBL) from the Hudson Bay Lowlands of northern Ontario, and agates from the Osler Formation in the northwestern Lake Superior region of Ontario. The three prospective geological sources were tested by INAA for 15 short half-life producing elements; and based on the elemental concentrations of Al, Si, U, Dy, and Cl (plotted on a ternary diagram) it was shown that geological KRF was separate from the HBL and agates. The KRF was chemically different from the other sources based on higher normalized and standardized U and Dy values, and generally lower Al/Si ratios. It was not possible to chemically separate the HBL and agate sources with the short half-life suite of elements. In testing the artifacts KRF was found to be only sparsely represented at Cummins site and that most of the brown artifacts were HBL and/or agates.

Other researchers (Clark 1984, Wright 1972) had previously reported KRF in western Great Lakes regional Archaic and Woodland assemblages. These reports had been based primarily on visual identifications, and Clark (1984) proposed several trade mechanisms to account for the reported distribution of KRF in both utilitarian and ritual (mortuary) contexts. INAA was conducted on 60 brown chert artifacts from five regional sites ranging in age from Archaic to late Woodland (Julig et.al. 1991b, 1992). Both KRF and HBL were represented in the utilitarian and ritual (mortuary) assemblages, but the number of KRF artifacts (most distant source) was less than 10%. Our research confirmed that some KRF was present, but questioned the previous visual identification of large numbers of KRF artifacts in regional assemblages (Clark 1984), and the trade mechanisms proposed for their (observed) distribution. However, this is clearly an area that requires further investigation.

Another related provenance problem addressed concerning the identification and distribution of HBL chert in secondary deposits in southern Lake Superior and northern Lake Michigan (Julig et.al. 1992). HBL is widely distributed on the Canadian Shield and in the upper Great Lakes as a result of Pleistocene glaciation. Brown, grey and tan coloured chert pebble are common in upper Great Lakes beach deposits and were used by the prehistoric native peoples (Binford and Quimby 1972). The question concerns the archaeological identification of some of these cherts as HBL, and if they were chemically similar to HBL from the source region. INAA data from a total of six secondary sources (35 samples) indicated that there was close chemical similarity between the HBL from the geological source region and the samples identified as HBL from secondary sources. Furthermore, this chert is chemically distinct from other regional (e.g. Gunflint chert from the Gunflint Formation north of Lake Superior) and distant chert look-alikes (e.g. KRF).

In south central Michigan dark and medium brown cherts of unknown provenance have been recovered in small quantities in Late Woodland assemblages (Goatley 1994). These recoveries of small brown flakes, normally less than 1% of the total lithic assemblage, compare visually with brown HBL or KRF samples. This material is reported to increase in frequency from south to north, and is not one of the southern Great Lakes brown cherts such as Flint Ridge or Plum Run (Goatley 1994, B. Luedtke pers. comm.). INAA analysis of eight samples of this unknown material indicated chemical similarity with HBL, based on the elements U, Dy, Al, Si, Cl, and that it was chemically distinct from KRF (Julig, in Goatley 1994:118).

In addition, INAA of geological samples of other Upper Lake Michigan and Lake Huron basin cherts (namely Fossil Hill Formation chert, Cordell Formation chert, and Norwood chert of the Middle Devonian Petosky Formation, and three samples of till chert) indicated these lithic materials are also chemically distinct from HBL, based on the elements Ba, Br and Cl (Julig et.al. 1992). The reasons for testing these materials is the close visual similarity of the unknown brown chert and some samples of Cordell chert. The unknown brown chert from Late Woodland sites in southwestern Michigan thus compares most closely with HBL and not with any of the other local or distant visual look-alikes. It can now be stated with confidence that they are northern imports, probably from secondary deposits of HBL on the Canadian Shield.

In summary, INAA has proven to be a useful analytical technique for the sourcing of chert artifacts as to their geological sources. This method with the SLOWPOKE reactor permits INAA of the short-lived isotopes on a large range of sample sizes and to retain the artifact in its original form, so it can be returned to the curator undamaged. This research has also demonstrated the potential for the separation of various regional chert sources based on the short-lived isotopes, and this has proven too be a very cost effective and rapid analytical method to solve some chert provenance problems.

ACKNOWLEDGEMENTS

This ongoing research has been supported by grants from the Social Science and Humanities Research Council of Canada to P. Julig. This manuscript is based on a presentation made at the "Workshops in Archaeometry", Department of Anthropology, State University of New York at Buffalo in February 1994. Any errors or omissions are the responsibility of P. Julig, and inquiries can be made to Pjulig@admin.Laurentian.Ca

REFERENCES CITED

 
Binford, L.R. and G.I.Quimby
1972 Indian Sites and Chipped Stone Materials in the            
Northern Lake Michigan Area. In L.R.Binford, _An            
Archaeological Perspective_. New York: Seminar Press. pp.         
346-372.
 
Clark, F.
1984 Knife River Flint and Interregional Exchange.            
Midcontinental Journal of Archaeology 9(2): 173-198             
DeBruin, M., P. Korthoven, C. Bakels, and F. Groen             
1972 The Use of Non-Destructive Activation Analysis and           
Pattern Recognition in the Study of Flint Artefacts.            
Archaeometry 14:55-63.
 
Goatley, D.
1994 Lithic Raw Material Utilization and Social Interaction       
in the Lower Kalamazoo River Valley. Unpublished Masters          
Thesis. Anthropology Department. Kalamazoo: Western Michigan      
University.
 
Gramly, R.M.
1988 Palaeo-Indian Sites South of Lake Ontario, Western and       
Central New York State. In R.S. Lamb, N.G. Miller, and D.W.       
Steadman (eds) _Late Pleistocene and Early Holocene            
Paleoecology and Archaeology of the Eastern Great Lakes           
Region_. Bulletin of the Buffalo Society of Natural            
Sciences, Vol. 33, pp. 265-280.
 
Hancock, R.G.V.
1978 Some aspects of the analysis of ancient artefacts by         
neutron activation. _Journal of the International Institute       
of Conservation, Canadian Group_ 3(2):21-27.
 
Hoard, R.J., J.R. Bozell, S.R. Holen, M.D. Glascock, H.           
Neff, and J.M. Elam.
1993 Source Determination of White River Group Silicates          
from Two Archaeological Sites in the Great Plains.  American      
Antiquity 58(4):698-710.
 
Ives, D.J.
1974 The Crescent Hills prehistoric quarrying area.  _Museum      
Brief No. 22_. Museum of Anthropology, University of            
Missouri-Columbia.
 
1984 Neutron Activation Analysis Characterization of            
Selected Prehistoric Chert Quarrying Areas, Unpublished           
Ph.D. dissertation, Department of Anthropology, University        
of Missouri-Columbia.
 
Jarvis, Hugh W.
1990 Instrumental Neutron Activation Analysis of Onondaga         
Chert in the Niagara Frontier. Ontario Archaeology 51: 3-15       
 
Julig, P.J.
1988 The Cummins Site Complex and Paleoindian Occupations in      
the Northwestern Lake Superior Region. Unpublished Doctoral       
Thesis. Anthropology Dept. Toronto: University of Toronto.        
 
Julig, P.J., L.A. Pavlish and R.G.V. Hancock.
1987 INAA of Archaeological Quartzite from Cummins Site,          
Thunder Bay, Ontario: Determination of Source.  Current           
Research in the Pleistocene, Vol. 4, edit. J.I. Mead, Center      
for the Study of Early Man, University of Maine, Orono, pp.       
59-61.
 
1988 INAA of lithic materials from the Western Lake Superior      
Region of North America.  Proceedings of 26th International       
Symposium of Archaeometry, University of Toronto, pp. 233-        
238.
 
1989 Late Paleoindian lithic technological organization in        
the Northwestern Lake Superior Region, Canada. In J. Lothrop      
and C.F. Ellis (eds.) _Eastern Paleoindian Lithic Resource        
Use_. Westview Press, Boulder, Colorado, pp. 293-322.             
 
1991a INAA of Lithic Materials from the Cummins site Thunder      
Bay: Determination of Raw Material Sources. In H.M. Hughes,       
M.R. Cowell, D.R. Hook (eds.) _Neutron Activation and Plasma      
Emission Spectrometric Analysis in Archaeology: Techniques        
and Applications_. British Museum Occasional Paper 82, pp.        
155-167.
 
1991b INAA provenance studies of lithic materials from the        
western Great Lakes region of North America. In E. Pernicka       
and A.W. Gunther (eds) _Archaeometry '90_. Proceedings of         
the 27th International Symposium on Archaeometry,            
Birkhauser Verlag Basel, Heidelberg, Germany, pp. 435-444.        
 
Julig, P.J., L.A. Pavlish, C. Clark and R.G.V. Hancock            
1992 Chemical Characterization and Sourcing of Upper Great        
Lakes Cherts by INAA.  Ontario Archaeology, no. 54, pp. 37-       
50.
 
Julig, P.J., L.A. Pavlish and R.G.V. Hancock
1992 Source Determination of Chert and Copper Artifacts by        
Neutron Activation Analysis: Some Examples from the Great         
Lakes Region. Paper presented at Geological Society of            
America Annual meetings, Cinninati, Ohio
 
Luedtke, B.E.
1978 Chert Sources and Trace Element analysis.  American          
Antiquity 43:413-423.
 
1979 The Identification of Sources of Chert Artifacts.            
American Antiquity 44:744-756.
 
Perlman, I., F. Asaro and H. Michel
1972 Nuclear Application in Art and Archaeology.  Annual          
Review of Nuclear Science, 22:383-426.
 
Sayre, E.V., R.W. Dodson
1957 Neutron Activation Study of Mediterranean Potsherds.         
American Journal of Archaeology, 61:35-41.
 
Storck, P.L. and P.H. VonBitter
1989 The Geological Age and Occurence of Fossil Hill            
Formation Chert: Implications for Early Paleoindian            
Settlement Patterns. In C.J. Ellis and J.C. Lothrop (eds.)        
_Eastern Paleoindian Lithic Resource Use_. Westview Press,        
Boulder, pp. 165-189.
 
Tite, M.S.
1972 _Methods of Physical Examination in Archaeology_,            
Seminar Press, London.
 
Wright, J. V.
1972  _Ontario Prehistory: an eleven-thousand-year            
archaeological outline_. National Museums of Canada: Ottawa