Behavior and Recruitment Success in Fish Larvae: Repeatability and Covariation of Survival Skills

Survival of larvae of highly fecund fishes is often considered to be random, although some investigators have hypothesized that survivors are exceptional individuals. If behavior is important to survival, individuals that perform both well and repeatably will be favored. We conducted experiments on red drum larvae (Sciaenops ocellatus, mean length = 7.7 mm) to assess individual variability in skills related to foraging and predator evasion. We measured routine swimming speed and, for responses to two startle stimuli, response score, latency, distance, duration, and speed. We tested 100 larvae five times each. A large proportion of larvae exhibited repeatable performance for 10 of the performance variables. However, repeatability (or performance level) in one skill was not correlated with repeatability (or level) in an independent skill, except for a positive correlation between acoustic and visual response scores. This result suggested that the level of attentiveness to predators varies among individuals. The absence of individuals with consistently low performance was taken as evidence for historical selection on performance and repeatability of escape speeds and distances. Individuals that performed repeatably or well in multiple skills represented 1-2% of the sample, which is not a trivial number of individuals when extrapolated over a female's reproductive lifetime. The persistence of many poor performers may be due to limited heritability of behavioral traits, a trade-off of behavioral skills with another selected trait, or extreme variability in the nature of mortality in coastal environments. The presence of repeatable behavior and apparent selection for a small number of related traits imply that, under some circumstances, survivors of a cohort may be individuals with exceptional behavioral skills, but behavior may be more important to survival and recruitment in fish larvae from more stable environments.