Elucidating the degree to which complex food webs are compartmentalized can lead to a better understanding of how species interactions, such as subsidized trophic cascades involving generalist predators, may affect rates of ecosystem-level processes such as decomposition and nutrient cycling. In the temperate forest ecosystem, the aboveground food web is composed of two potentially linked components: the grazing and decomposition subwebs, which are based on living primary production and detritus, respectively. Litter-dwelling spiders, which are generalist predators, can initiate trophic cascades in the detrital subweb that may affect decomposition rates. Energy flow from the grazing subweb could affect the strength of these trophic cascades if spiders receive extensive subsidies in the form of herbivore prey. We examined patterns of natural abundance of stable carbon isotopes (δ¹³C) to test the hypothesis that litter-dwelling spiders consume substantial numbers of herbivorous insects, thereby linking grazing and decomposition subwebs. Baseline carbon signatures were determined for herbivores inhabiting the understory vegetation or falling from the canopy, and for detritivores living in the litter layer. A single-isotope mixing model was then used to estimate the relative contributions of herbivores and detritivores to the diets of twenty-four abundant litter-dwelling spider species representing a range of foraging strategies.

Results/Conclusions

Detritivores, such as Collembola, had significantly higher δ¹³C values than herbivorous insects such as Lepidoptera larvae. This difference allowed us to partition the basal resources of the two subwebs on the basis of different carbon signatures. The mixing model revealed that the majority of litter-dwelling spiders derive most of their energy from the decomposition subweb (>75% of diet is detrital prey). Several species (e.g. the wolf spiders Arctosa virgo and Schizocosa spp. (Lycosidae), juvenile nursery web spiders, Pisaurina mira (Pisauridae), and juvenile stealthy ground spiders, Drassyllus spp. (Gnaphosidae)) rarely or never consume herbivores, whereas the jumping spider Phidippus whitmani (Salticidae) and the sac spider Elaver excepta (Clubionidae) probably consume herbivores regularly (30-50% of diet). Thus our original hypothesis is not supported. Although some spiders do forage in both subwebs, predator-mediated links between grazing and detrital subwebs are relatively weak in this system.  This supports the idea that the two subwebs are best viewed as distinct compartments of the aboveground forest food web.