Predation on lizard eggs by ants: species interactions in a variable physical environment.

Abstract

One explanation for long-term fluctuations in population density is that the intensity of interactions between species is variable. A population can experience variation in the intensity of a species interaction if (1) the density of species with which it directly interacts changes and/or (2) the strength of the interaction (i.e., per capita effects) changes. At Barro Colorado Island, Panama, the tropical lizard Anolis limifrons exhibits wide annual fluctuations in density. Previous studies have indicated that (1) the density of A. limifrons is negatively correlated with the amount of wet-season rainfall, (2) fluctuations in density are related more to variation in egg mortality than to variation in lizard mortality or to fecundity, and (3) most egg mortality is the result of predation by Solenopsis ants. We hypothesized that the amount of wet-season rainfall indirectly alters the density of A. limifrons by producing variation in the intensity of egg predation by Solenopsis. Additionally, we also wanted to determine if variation in the amount of egg mortality was influenced more by variation in the density of Solenopsis, or by variation in the rate of predation. We tested this hypothesis by manipulating litter moisture on experimental plots to simulate the wettest (HW) and driest (LW) wet seasons in the last 20 years, and then monitoring the density of Solenopsis, amount of egg mortality, and rate of predation. The amount of egg mortality was greater on the HW than on the LW treatment and all egg mortality resulted from predation by Solenopsis. ANCOVA indicated that the amount of egg mortality was significantly higher on plots with a greater density of Solenopsis. Treatment effects, however, explained more of the variation in the amount of egg mortality than did Solenopsis density. Our water manipulations did not change the density of Solenopsis, but Solenopsis found and attacked eggs faster on the HW than on the LW treatment. This suggests that moisture during the wet season modified the strength of the interaction between Solenopsis and A. limifrons, supporting the hypothesis that annual variation in the amount of wet-season rainfall indirectly produces parallel variation in annual density of lizard populations by modifying the rate of ant predation on eggs.