Allocation of effort in stream food-web studies: the best compromise?

The amount and allocation of effort needed to characterize stream food webs was investigated in five replicate streams. Two areas were considered: ana lysis of community composition (number of individuals sampled) and of diets (number of individuals gutted per animal taxon). Food webs were described by use of consistent methodology, then the effort was retrospectively reduc ed by considering half of the gut samples (halving dietary analysis effort) and by successively reducing the number of individuals included. Food webs with a reduced number of individuals overestimated connectance and prey: p redator ratios, and underestimated species richness, links per species and mean chain lengths. These changes were due to loss of some invertebrate pre dator species when effort was reduced. In contrast, for dietary analysis th e amount of effort expended on non-predatory invertebrates was more influen tial; halving effort in dietary analysis of non-predatory taxa reduced esti mates of connectance and links per species. This study suggests that the ef fort needed to produce a reasonable estimate is highly dependent on the foo d-web attribute in question, and that aiming for equity of effort across ta xonomic groups is as important as expending greater effort in general.