The predation intensity exerted by populations of the gastropod Thais lapillus at different study areas in the rocky intertidal community of New England is unrelated to predator density. Specifically, very similar intensities are exerted by populations differing in density by at least an order of magnitude. Predation intensity is, in part, a joint function of individual rates of prey consumption and various environmental characteristics. Major factors potentially affecting the individual feeding rates of Thais are (1) prey abundance and productivity, (2) other predators, (3) canopy-forming algae, (4) wave shock, (5) desiccation and (6) snail phenotype and/or history. The effects of the first two of these factors seem unimportant. The effects of the latter 4 on prey consumption rates were studied by estimating field feeding rates of snails held in cages with prey in microhabitats which were characterized by one of two alternative states of each factor. For example, microhabitats could be exposed or protected, at higher or lower levels in the mid intertidal, or under a canopy or not. In addition, exposed-phenotype or protected-phenotype snails were used in each experiment.
All of factors (3) to (6) had statistically significant effects except wave shock. The latter would probably also have had a significant effect if the experiments had been performed in the stormier part of the year as well as late summer. The results indicate that sparse populations of Thais can exert intense predation pressure on their prey if they are in protected sites covered with a dense canopy (i.e. in cool, moist habitats in calm waters). Areas with sparser canopy (i.e. greater desiccation stress) and more severe wave shock or both apparently reduce average feeding rates of snails. This appears to explain the paradoxical lack of correlation between predation intensity and snail density.
An unexpected result with potentially major implications is the nonlinear response of Thais feeding rates to combinations of factors (3) to (6). Four-way analyses of variance on experiments at exposed and protected sites indicate that 7 of 14 1st-order interactions, 2 of 8 2nd-order interactions, and even 1 of 2 3rd-order interactions are statistically significant. These results suggest that individual predators cannot be assumed to be identical, and that socalled “higher order” interactions cannot be safely ignored in models of interacting multi-species systems. Hence, it appears that to obtain a thorough understanding of the organization of natural communities, both field and theoretical ecologists alike should begin to grapple with such complexities of nature rather than ignore them.