Stochastic models of gregarious larval settlement

Abstract

Many marine invertebrate larvae exhibit gregarious settlement behavior. Gregariousness may interact with larval preferences for different substrata in producing settlement patterns. Consequently, the results of laboratory studies of substrate selection that use more than one larva per replicate may be confounded by gregariousness. I present results of computer simulations of larval settlement that illustrate the joint effects of larval selectivity and gregariousness. Each consecutive larva is offered two surfaces for settlement. Individuals settle randomly, and the probability of settlement is determined by both the underlying preference for that surface (selectivity) and the distribution of individuals that have already settled (gregariousness). Gregariousness leads to a consistent over-estimate of preference at intermediate or high values of selectivity. At low values of selectivity, gregarious settlement increases the chances of spuriously rejecting the null hypothesis and of misclassifying the direction of the larval response. These effects may be enhanced or diminished by small sample sizes, depending on whether larvae respond qualitatively or quantitatively to the distribution of settled individuals. These computer simulations suggest that gregariousness cannot simply be ignored or “averaged out” of settlement experiments. Laboratory settlement experiments should explicitly incorporate gregariousness as an additional factor to reveal whether substrate choices are modified by the presence of previously settled individuals. Such experiments should provide additional insight into patterns of recruitment in the field.