Variability effects by consumers exceed their average effects across an environmental gradient of mussel recruitment.

The implicit assumption that properties of natural systems deduced from the average statistics from random samples suffice for understanding them focuses the attention of ecologists on the average effects of processes and responses, and often, to view their variability as noise. Yet, both kinds of effects can drive dynamics of ecological systems and their covariation may confound interpretation. Predation by crabs and snails on competitively dominant mussels has long been recognized as an important process structuring communities on rocky shores of the Northwest Atlantic Ocean. We experimentally manipulated the average intensity of predation in plots across a gradient of mussel recruitment to separately estimate the average and variability of responses of mussel recruitment and community composition. Predation did not affect the average number of mussels recruited to plots, nor the average multivariate composition of the community. Plots from which predators were excluded showed a ~ 30% increase in spatial variability of mussel recruitment. After 1 year, the spatial variability in community composition was greater than that observed among plots that predators could access. An important, but less recognized, aspect of predation is its dampening effect on variability of community structure. As accelerating rates of environmental change disrupt species interactions, variability effects of ecological processes and corresponding responses are likely to be increasingly important determinants of community dynamics.