The outcome of an exclusion experiment depends on the method: shrimps, shredders and leaf breakdown in a tropical stream

The study of detrital food webs of streams has revealed a variety of dynamics, often implying the importance of invertebrate fauna in the breakdown of leaves and sometimes showing effects of predators and trophic cascades. Such studies generally involve the experimental manipulation of the component leaves and fauna. Leaf breakdown (or 'processing') is often studied by isolating leaves in litter bags, which in turn can have different mesh sizes used to exclude or permit the action of fauna. Such manipulation always carries the risk of introducing experimental artefacts, and care and appropriate controls must be applied. More recently and less commonly, electrical exclusion has been used for such studies, which potentially reduces the artefacts produced by bags, but care must be taken not to introduce other artefacts. Here we compared bag-exclusion and electrical-exclusion in a study of the effects of shredding fauna and predators on leaf processing in a coastal, tropical stream. The methods differentially excluded fauna of different sizes: high-intensity electrical shocks and fine-mesh (0.2 mm) bags excluded medium and large invertebrates; low-intensity shocks and medium-sized mesh (2 mm) bags excluded principally 2 species of shrimps (Macrobrachium olfersii and Potimirim brasiliana); no electric shocks and large mesh-size mesh (14 mm) bags permitted access to all fauna. Leaf packs of 2 species (Erythroxylum pulchrum and Miconia prasina) were either tethered free of bags in the electrical treatments or tied inside bags. The rate of leaf processing was higher (1.9-2.5x) for tethered leaves than for leaves inside bags indicating bags inhibited processing. The type of exclusion markedly affected differences in leaf processing between treatments. Leaf processing was 1.6 to 2.3x faster in the presence of medium-sized fauna in the treatment with the low-intensity electricity than when they were excluded by high-intensity electricity but was 1.9 to 2.0x faster in coarse-mesh bags, in which Potimirim shrimps aggregated at densities similar to 100x those observed outside bags, compared with treatments where the mesh size excluded shrimps. In the electrical exclusions, large fauna (possibly Macrobrachium shrimps) apparently inhibited medium-sized fauna (Ephemeroptera:Leptophlebiidae) in Erythroxylum but not in Miconia, producing a trophic cascade and reduced leaf processing. We conclude that the bag method produced artificially low rates of leaf processing and altered the dynamics of the fauna associated with leaf processing by attracting fauna, principally large shrimp, into the bags and masking the action of medium-sized fauna.