Coral reefs are increasingly ecologically destabilized across the globe due to climate change. Behavioural plasticity in corallivore behaviour and short-term trophic ecology in response to bleaching events may influence the extent and severity of coral bleaching and subsequent recovery potential, yet our understanding of these interactions in situ remains unclear. Here, we investigated interactions between corallivory and coral bleaching during a severe high thermal event (10.3-degree heating weeks) in Belize. We found that parrotfish changed their grazing behaviour in response to bleaching by selectively avoiding bleached Orbicella spp. colonies regardless of bleaching severity or coral size. For bleached corals, we hypothesize that this short-term respite from corallivory may temporarily buffer coral energy budgets by not redirecting energetic resources to wound healing, and may therefore enable compensatory nutrient acquisition. However, colonies that had previously been heavily grazed were also more susceptible to bleaching, which is likely to increase mortality risk. Thus, short-term respite from corallivory during bleaching may not be sufficient to functionally rescue corals during prolonged bleaching. Such pairwise interactions and behavioural shifts in response to disturbance may appear small scale and short term, but have the potential to fundamentally alter ecological outcomes, especially in already-degraded ecosystems that are vulnerable and sensitive to change.

A major goal of marine ecology is to identify the drivers of variation in larval dispersal. Larval traits are emerging as an important potential source of variation in dispersal outcomes, but little is known about how the evolution of these traits might shape dispersal patterns. Here, we consider the potential for adaptive evolution in two possibly dispersal‐related traits by quantifying the heritability of larval size and swimming speed in the clown anemonefish ( Amphiprion percula ). Using a laboratory population of wild‐caught A. percula , we measured the size and swimming speed of larvae from 24 half‐sibling families. Phenotypic variance was partitioned into genetic and environmental components using a linear mixed‐effects model. Importantly, by including half‐siblings in the breeding design, we ensured that our estimates of genetic variance do not include nonheritable effects shared by clutches of full‐siblings, which could lead to significant overestimates of heritability. We find unequivocal evidence for the heritability of larval body size (estimated between 0.21 and 0.34) and equivocal evidence for the heritability of swimming speed (between 0.05 and 0.19 depending on the choice of prior). From a methodological perspective, this work demonstrates the importance of evaluating sensitivity to prior distribution in Bayesian analysis. From a biological perspective, it advances our understanding of potential dispersal‐related larval traits by quantifying the extent to which they can be inherited and thus have the potential for adaptive evolution.