Interactive effects of spatial variance and mean intensity of grazing on algal cover in rock pools

An increasing number of studies suggest that the variance around the mean of predictor variables is an important determinant of the structure of assemblages, but few have separated effects of mean from those of variance of predictor variables experimentally. We addressed these issues by examining the effects of limpets on algal recovery in littoral rock pools. An initial experiment indicated that algal cover was a negative accelerating function of density of limpets, although this effect was weak. Superimposed to this effect there was a clear trend of variability in algal cover among pools that decreased with increasing density of limpets. These results suggested that mean intensity and spatial variance of grazing could affect algal colonization both within and among pools. This proposition was tested with a novel experimental design in which the mean intensity and the spatial variance of grazing were manipulated in factorial combinations. The experiment was repeated twice to examine possible generalities of outcomes. In both runs spatial variance of grazing enhanced spatial variance in algal cover, although this effect occurred under weak intensity of grazing in the first run and in combination with strong intensity of grazing in the second run. In contrast, intensity and variability in grazing operated independently on mean algal cover, with strong intensity depressing colonization of algae and large variance increasing colonization consistently in time. By manipulating the mean intensity and the spatial variance of grazing simultaneously, we were able to explain spatial patterns in algal cover that would have been classified as unexplained variation otherwise. We suggest that this novel experimental design may be fruitful to understand spatially variable processes in a wide range of ecological systems.