Marine and Terrestrial Protein in Prehistoric Diets on the British Columbia Coast

by BRIAN S. CHISHOLM, D. ERLE NELSON, and HENRY P. SCHWARCZ Department of Archaeology, Simon Fraser University, Burnaby, B.C., Canada V5A 1S6 (Chisholm and Nelson)/Department of Geology, McMaster University, Hamilton, Ont., Canada L8S 4M1 (Schwarcz). 12 x 82 Ethnographic and archaeological descriptions of subsistence on the Northwest Coast of North America are generally incomplete or imprecise (Fladmark 1975). This makes it difficult to determine quantitatively the degree of dietary reliance of prehistoric Northwest Coast people upon marine species. In this study we use a newly developed analytical tool to make direct measurements of average relative amounts of terrestrial and marine protein in the diets of ancient people by determining stablecarbon isotope ratios in human bone collagen. The analytical technique is based upon the difference in carbon isotope ratios that exists between the atmospheric and marine carbon-dioxide reservoirs (Chisholm, Nelson, and Schwarcz 1982). This difference is reflected in terrestrial C3 plants' and in marine plankton, and also at each succeeding trophic level of the terrestrial and marine food chains based upon them. Because of this, it is possible to determine whether a human consumer was obtaining food from either a C3-plantbased or a marine-plankton-based food chain or a mixture of the two. (Terrestrial plants, such as maize, that use the C4 photosynthetic ycle differ in carbon isotope ratios from C3 cycle plants and could produce results that overlap those from marine food chains. Fortunately, C4 plants are very rare in the Northwest Coast study area.) In situations, such as that on the Northwest Coast, in which protein intake is high and carbohydrate intake is very low (Drucker 1965), the collagen-forming amino acids, both essential and nonessential, will be derived from the protein portion of the diet (White et al. 1978:687), as synthesis of nonessential amino acids from carbohydrates is not necessary. The collagen stable-carbon isotope ratio measurements therefore reflect he protein portion of the diet. They are also thought to give the effective lifetime average marineterrestrial protein intake ratio for an individual (Chisholm et al. 1982). The quantity measured in these studies is the 6'3C value, which relates the stable-carbon isotope ratio of an unknown sample to an international standard2 and which is defined by a13C (per mil) = (RU/RR 1) X 1,000, where RU is the ratio (13C/12C) for the unknown, and R. is that for the standard. In our earlier work (Chisholm et al. 1982) we determined that the best available 8'3C values for this region for consumers of purely C3-plant-based and purely marine-based diets were -20 ? 0.9 per mil and -13 ? 0.9 per mil respectively. If intake from the two diets is mixed, the resulting value for a consumer falls in the 7-per-mil nterval between these two end points, allowing us to interpolate and thereby determine the relative proportions derived from each source. The uncertainty of ? 20% applied to the proportion estimates (Chisholm et al. 1982) was derived from the ? 0.9-per-mil uncertainties for the two end points and for the sample being considered. One difficulty involved in the application of this technique is that the end-point values are not yet well known, and slightly different ones have been used for purely C3-plant-food consumers (van der Merwe and Vogel 1978; Bender, Baerreis, and Steventon 1981). These perhaps result from differences in instrument calibration (Mann 1982) or in sample preparation. Our feeling is that values of about -22 per mil and -12 per mil, for purely C3-plant-food consumers and purely marine-food consumers respectively, could possibly be used instead of the -20 per mil and -13 per mil used here, but that any more extreme alteration in 613C values is unlikely. Use of these alternative end-point values would result in a decrease, from our calculated results, of about 10% in the estimated proportion of marine protein in the diet of a consumer. It would also result in a slight decrease in the uncertainty for the proportion estimate. However, it would not affect he proportion estimates relative to each other and does not affect the conclusions drawn here. A total of 48 samples from prehistoric human skeletons were obtained from osteological collections at Simon Fraser University and from the National Museum of Man, Ottawa,3 for 15 sites along the British Columbia coast (fig. 1). The average B13C