Food-web models use the effect size of trophic interactions to predict cons
umer-resource dynamics(1-3). These models anticipate that strong effects of
consumers increase spatial and temporal variability in abundance of specie
s, whereas weak effects dampen fluctuations(4-6). Empirical evidence indica
tes that opposite patterns may occur in natural assemblages(7). Here I show
that spatial variance in the distribution of resource populations is sensi
tive to changes in the variance of the trophic interaction, in addition to
the mean effect of consumers, relative to other causes of spatial variabili
ty. Simulations indicate that both strong and weak direct effects of consum
ers can promote spatial variability in abundance of resources, but only tro
phic interactions with a large mean effect size can reduce variation. Predi
ctions of the model agree with the results of repeated field experiments an
d are consistent with data from published consumer-resource interactions, p
roving to be robust across widely varying environmental conditions and spec
ies' life histories. Thus, food-web models that embody variance in trophic
interactions may have increased capacity to explain the wide range of effec
ts of consumers documented in empirical studies.