COS 86-10
Development time and the stability of a host-parasitoid interaction
Theoretical models have demonstrated that predator-prey stability is strongly affected by the duration of various life stages, as well as the generation time of the prey relative to the predator. In particular, one widely recognized prediction from these published models has been that long invulnerable prey stages increase predator-prey stability. Using the cowpea weevil Callosobruchus maculatus and its parasitoid Anisopteromalus calandrae as a model predator-prey system, we provide the first rigorous experimental test of this theory. In replicate experimental microcosms, the duration of the invulnerable juvenile host stage of the weevil was experimentally lengthened by 60% (long-duration [LD] treatment) or reduced by 60% (short-duration [SD] treatment) relative to unmanipulated controls (UC). Based on theoretical models, a long invulnerable stage (e.g., the LD treatment) should induce generation cycles and favor greater predator-prey stability than a short invulnerable stage (SD treatment). Experimental microcosms consisted of large petri dishes containing cowpea weevils, parasitoids and moth beans (Vigna acontifolia) as hosts. The experiment was run for two years (> 30 host generations).
Results/Conclusions
Microcosms subjected to the different duration treatments exhibited very different population dynamics. In the SD treatment, host densities were the lowest (2 times lower than the LD), parasitoid densities were the highest (33% greater than the LD), and the variability in densities was the lowest among the three treatments. The LD treatment had the highest frequency of a host/parasitoid extinction event - 50% in the LD versus 10% in the other treatments. Parasitoids had the most variable densities in the LD treatment, and was the species that generally went extinct. Based on a wavelet analysis, UC microcosms exhibited a powerful signal for period-two oscillations within the host population, or approximately one host generation. The parasitoids exhibited relatively low-power oscillations with a period of 2-6. The LD and SD microcosms exhibited qualitatively similar cyclical behavior to the controls, but the signal for periodicity was absent for the SD treatment. Overall, our experimental results, that the SD treatment favored greater host-parasitoid stability than the LD treatment (in terms of reduced variability in population densities and elimination of periodicity in the time series), contradicts existing theory. Despite our findings, the prediction that age-structure is a key factor affecting predator-prey population dynamics is upheld.