Evolution of pace-of-life
Generation time can be used to predict a species’ position along the slow–fast continuum of life-history variation (Gaillard et al.2005; Araya-Ajoy et al. 2018). Classic life‐history theory (Charlesworth 1994) suggests that variation in the mean age at reproduction evolves as consequence of compromises between age‐specific survival and reproduction, resulting in some combination of life-history traits that optimizes mean fitness under particular ecological conditions. The contemporary evolution of the generation time, and thus the pace-of-life of an organism, hinges upon there being genetic variation underpinning the traits that determine generation time. When we estimated the additive genetic variation in individual measures of generation time, we found that evolutionary potential or evolvability was within the range of values estimated for other life-history traits (Hansen et al. 2011). Individuals with longer generation times had greater lifetime reproductive success, but not necessarily a greater individual growth rate, and thus it is not clear that selection should favor the evolution of longer or shorter generation times.
The evolution of generation time is predicted to be constrained by trade-offs between reproduction and survival (Stearns 1992; Roff 1993). In this study, we found the expected negative covariance between reproductive rate and generation time, and between reproductive rate and lifespan (Figures 1B and 1F). This is partly unexpected, because variation in acquisition of resources is expected to mask life-history trade-offs in observational studies (van Noordwijk & de Jong 1986). To further explore resource acquisition versus allocation in our metapopulation, we analyzed the patterns of covariation between individual life-history traits using principal component analyses (Table S3). We found that the most important driver of the first axis of variation was lifetime reproductive success, with positive loadings for reproductive rate and lifespan, reflecting among-individual variation in resource acquisition. Whereas the strongest driver of the second principal components axis was generation time, where the opposing signs of reproductive rate and life span revealed the classic survival versus reproduction trade-off.
To further explore evidence of allocation trade-offs, we partitioned the covariance between annual reproduction and survival at the among- and within-individual levels (Table 3). If there was a trade-off between survival and reproduction, then in years where individuals invested more in survival, they cannot have invested a lot in reproduction. This should generate a negative within-individual covariance due to resource allocation trade-offs. Variation among individuals in resource acquisition should, in turn, be manifested as a positive among-individual covariance, where individuals with higher than average recruit production will also have a greater chance of surviving, because they are able to acquire more resources. We found a positive covariance between survival and reproduction at both levels, and therefore we find no evidence of an allocation trade-off at the within-individual level.
The positive covariance between survival and annual reproduction partly appears to contradict the results above concerning the trade-off between lifespan and recruitment rate. We addressed this contradiction using statistical simulations based upon the estimates of annual survival and reproduction (see Appendix S3). We explored how the demographic characteristics of this metapopulation could have resulted in a negative covariance between lifespan and reproductive rate, in the presence of a positive covariance between survival and annual reproduction (see Figure S3). We found that even when there was variation among individuals in acquisition of resources, causing a positive covariance between survival and annual reproduction at the within- and among-individual levels, the resulting distribution of individual life-histories in the population tended to generate a negative covariance between lifespan and reproductive rate. Given the observed patterns of age- and density-dependent survival and reproduction, even when individuals that had a greater survival probability also had a higher reproduction rate, a negative correlation between lifespan and reproduction rate was possible, simply as a result of the distribution of individual life-histories in the population.