Multi-tissue stable carbon and nitrogen isotope models for dietary reconstruction: Evaluation using a southern...

Multi-tissue stable isotope models to reconstruct past diets (Froehle, Kellner, & Schoeninger, 2012; Kellner & Schoeninger, 2007) have lacked data from a heavily C₄ -dependent population. Using new data from southern African agriculturalists, published models are evaluated for accuracy in dietary reconstruction and applicability to isotopically diverse diets. Additionally, isotopic variation between tooth enamel and bone apatite, which are often treated as isotopically equivalent, is investigated.

δ¹³ C_{bone apatite} , δ¹³ C_{tooth enamel} , δ¹³ C_collagen , and δ¹⁵ N_collagen values for 51 adult southern African agriculturalists are presented. Bivariate (linear) and multivariate (cluster analysis, discriminant function analysis) models are recreated including these data, and the resulting dietary reconstructions evaluated against what we know of archaeological diets.

Δ¹³ C_{collagen-enamel} (5.67 ± 1.66‰) is significantly larger than Δ¹³ C_{collagen-bone apatite} (4.77 ± 1.42‰) and are significantly different from each other (Mann Whitney U-Test, p = 0.0). δ¹³ C_{bone apatite} and δ¹³ C_{tooth enamel} were uncorrelated (R² = 0.24). The agriculturalists consumed highly variable and heterogeneous diets, (mean δ¹³ C_{bone apatite} = -6.25 ± 2.49‰, δ¹³ C_{tooth enamel} = -2.88 ± 2.48‰, δ¹³ C_collagen = -8.65 ± 2.16‰, δ¹⁵ N_collagen = 10.05 ± 1.9‰). Multi- and bi-variate models under-estimate the probable contribution of C₃ energy sources, and recreation of cluster analysis results in a significant reduction in the parsimony of the dietary clusters derived in Froehle et al., 2012.

Bone apatite and tooth enamel are distinct biominerals, and their δ¹³ C values should not be treated as equivalent. Multiple tissue isotopes provide valuable insight into diet that cannot be achieved with single tissues, but current models are limited by the lack of isotopic diversity in the data on which they are based.