Top-down control in a detritus-based food web: fish, shredders, and leaf breakdown.

We tested the hypothesis that fish decrease shredder abundance in leaf packs, thereby reducing leaf breakdown rates. Our goal was to test for the occurrence of a trophic cascade in a detritus-based food web. Willow leaves (Salix spp.) were fastened into leaf packs and placed into cages (13×13×13 cm) in Valley Creek, Minnesota, USA. Fish were excluded from leaf packs that were placed in cages with mesh on all sides, whereas open control cages allowed fish access to leaf packs. We collected leaf packs from two replicate cages 0, 14, 31, 55, and 112 days after placement in each of three riffles (n=6 per collection). Total abundance of invertebrates and shredders inhabiting leaf packs was significantly higher in exclosures than controls (P<0.01) and increased with exposure time in the stream (P<0.01). Three of the four common shredder taxa had significantly higher biomass in exclosures than controls (P<0.015). Biomass of Hesperophylax (Trichoptera) larvae was significantly higher in controls during the final collections (P<0.03), probably because these large, case-building larvae were less vulnerable to fish predation. Leaf breakdown rates differed significantly between exclosures and controls (P=0.003), but the direction of effects varied among riffles. When shredder density was analyzed separately for each riffle, we found that shredder density may explain differences in leaf breakdown rates between exclosures and controls. The differential responses of shredder taxa to predators may explain variability in fish effects on leaf breakdown. In conclusion, leaf packs did not provide invertebrates refuge from fish predation and fish reduced the densities of most shredders. Fish can indirectly affect leaf breakdown rates, but different responses to predation among taxa within the shredder guild can cause interactions that contradict trophic cascade predictions.