Simulating selective mortality on tadpole populations in the lab yields improved estimates of effect size in nature

Many populations normally experience high levels of mortality throughout larval development, but this is generally overlooked with laboratory experimental protocols. Evidence suggests that mortality is nonrandom in natural tadpole populations, so high survivorship, typical of laboratory populations, may poorly represent populations in nature. We compared survival, growth and development, and population variance of tadpoles in natural ponds with those in the laboratory at low and high densities. In the laboratory, high-density groups were reared with no selection and with selection imposed against different size classes to identify if, and how, mortality influences natural tadpole populations and to investigate whether imposing selection against certain size classes produces responses more consistent with those observed in natural systems. Our results suggest that selective mortality removes smaller individuals in natural populations. We demonstrate that introducing selection against small individuals artificially, in the laboratory, results in individual growth and development, population variance, and statistical power that more closely resembles that observed in natural populations. This is important from an ecological perspective because it demonstrates how selection acts on natural tadpole populations. More importantly, this demonstrates that laboratory experiments can be designed to provide better qualitative estimates for responses of natural populations by considering and simulating natural rates of mortality.