Diet

Although paleobotanic and paleofaunal residues are the most accurate indicators of resources available to prehistoric peoples, estimates of ancient diets are best derived from the actual physical remains of the consumers.

More sensitive to dietary differences, however, are the chemical constituents of the skeleton. The compo­sition of both the mineral and the organic phases of bone is to some degree influenced by diet. Trace elements, such as stable strontium and barium, are carried from the soil to an herbivore’s bones through plant consumption. Because both strontium and bar­ium tend to concentrate in osseous rather than soft tissue, primary and secondary carnivores receive relatively little dietary strontium. Thus, within a region, the relative herbivory of an omnivorous spe­cies such as Homo sapiens can be estimated. Zinc tends to concentrate in flesh and is thus a dietary marker for animal protein consumption.

The strontium content of bones has been used by M. Schoeninger (1979) in her investigation of status clusters of graves from the Chalcatzingo site, Morelos, Mexico. Her careful argument concludes that diet varied by status group at this prehistoric Meso-American site dating from 2900 to 3100 B.P.

Zinc and strontium have been used to investigate dietary differences within and between Middle and Late Woodland groups in west-central Illinois (Lam­bert, Spzunar, and Buikstra 1979; Buikstra et al. 1989). Although other elements such as vanadium co-vary with diet and would thus be desirable di­etary markers, most are subject to postdepositional enrichment or depletion (diagenesis) and are thus contextual signatures rather than dietary signals. Focused on intrasite variation, the earliest Illinois studies examined patterning across status groups in Middle Woodland mounds and sought gender- related differences within Middle and Late Wood­land samples. The nonrandom distribution of trace elements across burial groups in Hopewell mounds suggests the presence of status-related dietary hab­its within Middle Woodland communities. Although no gender differences were observable in Middle Woodland data, bones of Late Woodland men con­tained significantly more zinc and less strontium than those of women. This may reflect either a true gender-based dietary difference or the increased metabolic demands that Late Woodland women faced owing to closer birth spacing (Buikstra et al. 1987).

Diachronic study of Illinois Middle and Late Wood­land samples indicates that, although zinc values do not vary chronologically, strontium levels do de­crease during Late Woodland times. This decrease may seem enigmatic, given the development of maize agriculture during this period. Maize, however, is relatively poor in elements, and thus the observed pattern is expected. The values for zinc suggest that the proportion of animal protein in the diet did not vary significantly across Woodland groups. Similar patterning has been reported for a temporally se­quential series of skeletal samples from Ontario (Katzenberg and Schwarcz 1986; Katzenberg 1988).

A second form of chemical study has focused on the collagen fraction of bone. The results of studies based on stable carbon isotope ratios derived from bone collagen make it possible to predict the pres­ence of maize in diets with a precision far beyond estimates based on the paleobotanic record or dental health.

The association of maize agriculture with Missis­sippian cultures is well known. Debate has centered, however, on the degree to which the timing of maize intensification was directly associated with the de­velopment OfMississippian lifeways and the relative importance of maize to Mississippian peoples. The carbon isotope technique has provided a means of resolving these and related issues concerning agri­cultural intensification in the eastern United States. This significant methodological advance, developed only within the past 10 years, has facilitated the resolution of century-old arguments about the role of com in North American prehistory.

Partial differential carbon-13 values from across eastern North America clearly document the fact that maize consumption predates the Mississippian period (Lynott et al. 1986; Buikstra et al. 1988). In the central Mississippi valley, these values suggest a rather gradual increase during the terminal phases of Late Woodland, including a highly variable period of experimentation. Values stabilize during Missis­sippian times, indicating that corn was a significant component of the diet for the late prehistoric peo­ples. Of special interest is the fact that corn consump­tion was apparently higher in the “farmstead” com­munities of the lower and central Illinois valley than it was at the major ceremonial center of Cahokia during the early, expansive phase of the Missis­sippian period (Buikstra and Milner 1988). Also sig­nificant is the fact that the partial differential carbon-13 values from Illinois never reach the ex­tremely positive figures (—7.8 to —8.0 o∕oo) that char­acterize the Mississippian peoples of the Ohio Valley and the Nashville Basin of Tennessee.

Ratios of nitrogen-15 to nitrogen-14 have been used to identify the presence of legumes in prehis­toric diets, as well as to distinguish between marine and terrestrial resource utilization (DeNiro, Schoen- inger, and Tauber 1983; Schoeninger and DeNiro 1984; DeNiro 1987). A bivariate plot of partial differ­ential carbon-13 and nitrogen-15 values enables the researcher to distinguish the effects of marine en­richment from those of maize utilization. Diets com­posed of marine resources derived from coral-reef habitats appear to present an anomalous pattern (Schoeninger et al. 1983; Keegan and DeNiro 1988).

At present, most studies of coastal series using both carbon and nitrogen ratios are limited to small sam­ples. Expanded applications of these promising tech­niques are anticipated.