Omnivory and the stability of simple food webs

Abstract

Traditional ecological theory predicts that the stability of simple food webs will decline with an increasing number of trophic levels and increasing amounts of omnivory. These ideas have been tested using protozoans in laboratory microcosms. However, the results are equivocal, and contrary to expectation, omnivory is common in natural food webs. Two recent developments lead us to re-evaluate these predictions using food webs assembled from protists and bacteria. First, recent modelling work suggests that omnivory is actually stabilizing, providing that interactions are not too strong. Second, it is difficult to evaluate the degree of omnivory of some protozoan species without explicit experimental tests. This study used seven species of ciliated protozoa and a mixed bacterial flora to assemble four food webs with two trophic levels, and four webs with three trophic levels. Protist species were assigned a rank for their degree of omnivory using information in the literature and the results of experiments that tested whether the starvation rate of predators was influenced by the amount of bacteria on which they may have fed and whether cannibalism (a form of omnivory) occurred. Consistent with recent modelling work, both bacterivorous and predatory species with higher degrees of omnivory showed more stable dynamics, measured using time until extinction and the temporal variability of population density. Systems with two protist species were less persistent than systems with one protist species, supporting the prediction that longer food chains will be less stable dynamically.