A Meta-Analysis on Facultative Responses of Embryonic Amphibians to Predation Risk
Natural selection is predicted to favor the evolution of life history plasticity in response to differences in mortality risk between life stages, such that the duration of the more dangerous stage is reduced. I report a meta-analysis of 29 published tests of this hypothesis involving 20 amphibian species, in which exposure of embryos to predation risk was experimentally manipulated. The results confirm that amphibians are sensitive to risk during the embryonic stage, but the predictions of theory were only weakly upheld. Amphibians hatch at a less advanced developmental stage (0.24 SD units) and smaller body size (0.10 SD units) when exposed to predator species that consume eggs, but the expected decline in age at hatching (0.16 SD units) was not significant. Predators that consume larvae did not significantly delay hatching (0.19 SD units) or influence any other response.
Effects of predation risk on (A) survival, (B) age, (C) developmental stage, and (D) body size at hatching of amphibians exposed to predation risk during the embryonic stage. The figure shows Hedges' G, which measures the effect of predators in units of standard deviation of the original measurements. The vertical dashed lines indicate the situation in which Hedges' G = 0. Points and horizontal lines illustrate the mean and 95% highest posterior density (HPD) intervals. The figure shows the overall average effect size for every trait (estimated from Eqn 1 with only the intercept in the fixed part of the model) and all covariates that significantly influenced effect size. The covariates, defined in Table 1, are listed along the right-side border; levels of the covariate are along the left-side border; sample size is at the top of each panel.
Correlations between the effects of predators on age at hatching, developmental stage, and body size. Values are averages of Hedges' G calculated for unique combinations of study, amphibian species, and predator type (consumes embryos, larvae, or both). The dashed line is a major axis regression through all observations; solid lines indicate the origin. Table 2 reports the correlations, confidence intervals, and sample sizes.
Contributor Notes
Associate Editor: J. Kerby.