Testing the assumptions of mixing models in isotopic ecology

By use of experiments, this thesis examines two relatively unexamined assumptions of the mixing models often applied to dietary studies in isotopic ecology. Mixing models assume that (1) elements of ingested nutrients are incorporated into animal tissues in the proportion that they are ingested, and (2) the discrimination factors between diet and tissues are either equal among dietary sources, or if they are unequal that they are additive. Assumptions 1 and 2 can be called the "no-routing" and the "equality and/or additivity of discrimination factors" assumptions, respectively. To test assumption 1, I fed tilapias diets in which the protein content varied, and in which the delta13C and delta15N of protein differed from that of other macronutrients (carbohydrates and lipids). Except at the lowest dietary protein level, the muscle of fish had delta 13C values that differed from the values predicted by concentration-dependent mixing models. Thus, using these models would have led to overestimation of dietary protein to diet. To test assumption 2, I fed tilapias diets formulated with corn and wheat, which differ in carbon and nitrogen isotopic composition, mixed in different proportions ranging from 0 to 100% of each of these sources. I found that the discrimination factors for pure corn and pure wheat diets differed significantly, so one cannot assume that discrimination factors are equal among dietary sources. However, the additivity assumption was correct in this experiment.