Predation on lizard eggs by ants: interaction modifications in an unstable physical environment

The importance of abiotic influences on the strength of biotic interactions is largely unknown. To explain large annual fluctuations in the population size of the tropical lizard, Anolis limifrons, on Barro Colorado Island, Panama, I hypothesized that annual variation in lizard population size is the result of modifications in the rate of predation on lizard eggs by Solenopsis ants induced by annual variation in wet season rainfall. I tested this hypothesis by manipulating water availability on experimental plots to Simulate the wettest (HW) and driest (LW) wet seasons in the last twenty years. The mean time to find and attack eggs by Solenopsis ants was significantly shorter on HW plots (range=6.6-21.7 days) than LW plots (range=17.8-30.8 days). Exponential models that regressed time on the cumulative percent mortality indicated that 1) the rate of predation was 3-5 times faster on HW plots than LW plots and 2) the predicted mortality of lizard eggs during their 42 day incubation period was 82.2-95.7% on HW plots and 56.1-58.6% on LW plots. Thus, the amount of rainfall during the wet season affected the population size of A. limifrons by modifying the strength of the interaction between Solenopsis ants and the eggs of A. limifrons.