Selective mortality on early life-history traits of a temperate reef fish

Variability in recruitment and early life-history traits is widespread in many marine organisms. Phenotypic variation is particularly prevalent in the early life-history stages (e.g., larvae and juveniles) of reef fish, and provides the basis for selective mortality on growth and size-related traits, with important ecological and evolutionary consequences. Recruitment variability can alter the effective densities experienced by these early life stages, raising additional questions about the interplay between selection and density-dependent processes. While many examples of growth- and size-selective mortality have been documented for young reef fish (typically caused by predators), few studies consider how the strength and/or direction of selective mortality changes with ontogeny, or how these patterns may be mediated by density. I explore spatio-temporal variability in early life-history traits of the common triplefin, Forsterygion lapillum, using metrics derived from otoliths (a re-analysis of two previously collected data sets). I evaluate patterns of variation in traits with respect to early life-history stage (either larvae or post-settlement juveniles) and document shifts in the distributions of traits that are consistent with selective mortality favouring slower growing individuals. I conclude that a cohort of juveniles (sampled after settlement) was comprised of individuals that were smaller at hatch and grew slowly throughout the pelagic larval period relative to a cohort of larvae (sampled prior to settlement). I then conducted an experiment using a set of mesocosms to evaluate whether selective mortality on early life-history traits in common triplefin could be caused by a natural predator, the variable triplefin, Forsterygion varium. Specifically, I exposed groups of fish of each stage to a pair of predators and I used otoliths to reconstruct the traits of fish that survived versus fish that were consumed (i.e., I recovered otoliths from the guts of predators). Selection trials were conducted across realistic density gradients for each developmental stage. Fish size was negatively correlated with relative fitness for larvae (indicating larger fish were consumed preferentially by predators) but not for juveniles (where no size-selective mortality was observed). These patterns were consistent across the range of densities evaluated. Both larvae and juveniles experienced significant selection against fast larval growth (estimated from growth increments in otoliths), and the strength of selection was inversely related to density (i.e., strongest at lower densities, weakest at higher densities). However, juveniles also experienced selective predation for fast growth at the larval-juvenile transition. As with larval growth, selection was strongest at lower densities and weakest at higher densities. Collectively, these results suggest that predators may preferentially target larger larvae, and faster growing individuals regardless of developmental stage. However, this effect may be mediated by density, such that the strongest selection occurs during low recruitment. Density-dependent selection could explain how faster growing individuals can survive this vulnerable stage. These results provide evidence for carry-over effects of larval growth on juvenile survival, and suggest conspecific density should be considered when evaluating patterns of selective mortality.