Although studies quantifying evolutionary change in response to the selective pressures that organisms face in the wild have demonstrated that organisms can evolve rapidly, we lack a systematic assessment of the frequency, magnitude and direction of rapid evolutionary change across taxa. To address this gap, we conducted a meta-analysis of 58 studies that document the effects of warming, predation or competition on the evolution of body size, development rate or fecundity in natural or experimental animal populations. We tested whether there was a consistent effect of any selective agent on any trait, whether the direction of these effects align with theoretical predictions, and whether the three agents select in opposing directions on any trait. Overall, we found weak effects of all three selective agents on trait evolution: none of our nine trait by selective agent combinations had an overall effect that differed from zero, only 31% of studies had a significant within-study effect, and attributes of the included studies generally did not account for between-study variation in results. One notable exception was that predation targeting adults consistently resulted in the evolutoin of smaller prey body size. We discuss potential causes of these generally weak responses and consider how our results inform the ongoing development of eco-evolutionary research.

In August 2020, we searched the Web of Science for papers that included (“evolution” OR “adaptation”) AND (“temperature” OR “warming” OR “thermal” OR “predators” OR “predation” or “competition” or “competitors”) in the title. This returned 595 results, all of which were assessed for inclusion (detailed search methods in Supporting Information). We then searched 20 previous reviews and meta-analyses on related topics for additional studies that met our inclusion criteria (see list in Supporting Information). Finally, we searched all of the studies that met our inclusion criteria for further studies to include.

We used the following criteria to determine inclusion in our meta-analysis. Studies had to: (1) quantify evolutionary change in response to one or more of our three focal selective agents (temperature, predation, competition) (2) compare at least one group that we could use as a control group (e.g. lower temperature, no predation, no competition) to at least one group that we could use as a treatment group (e.g. higher temperature, predator present, competitor present) (3) measure at least one of our three focal life history traits (body size, development rate, fecundity) in both the control and treatment group, and (4) seperate evolutionary change from maternal effects and plasticity by growing individuals under common conditions for at least one generation (i.e., a common garden).

We included both “experimental” studies in which the selective agent was imposed by the researcher either in the lab or in the field (e.g., experimentally imposed presence or absence of a predator) and “natural” studies in which the selective agent varied naturally across space or time (e.g., comparison of streams with predators to streams without). We excluded studies that focused on evidence from the fossil record, but placed no other restrictions on study duration and instead included the number of generations of selection in our models to determine whether this affected the outcome of evolution (see Data Analysis). We restricted competition studies to those that focused on resource competition (i.e., no mate competition). We also excluded studies for which it was not clear that the selective agent of interest was the principal factor driving trait change (e.g., studies that assessed trait change across latitudinal or urbanization gradients as a proxy for temperature change). Our inclusion criteria resulted in the exclusion of rapid evolution studies that focused on the effects of other environmental variables (e.g., drought, CO₂), other biotic interactions (e.g., disease, mutualisms), or other anthropogenic changes (e.g., fish harvesting, habitat fragmentation). Our criteria also excluded studies that used population growth rate rather than life history traits as the response, that did not separate plastic from heritable responses, that used selective mortality or selection gradients as the response, or that imposed artificial selection. Details of our study selection criteria are provided in the Supporting Information.

Our selection criteria resulted in a dataset of 58 studies focused on animals. The 58 remaining studies encompassed 33 focal species (14 aquatic and 19 terrestrial) spanning 15 orders (Fig. S1). All but two of the included studies were conducted with ectothermic organisms (the exceptions were studies with rabbits (Williams & Moore 1989) and mice (Barnett & Dickson 1984)). Many of the 58 studies measured multiple life history traits, included multiple focal species, or assessed the effect of multiple selective agents, and thus our final dataset included 124 unique effect sizes (data points).

See "Explanations" tab in excel file.