Most freshwater food web models assume that fish occupy the top trophic level. Yet many diet studies and a few caging and artificial stream experiments suggest that birds may be top predators in many freshwater systems. We conducted a large-scale field experiment to test whether avian predators affect the size distribution and abundance of fish in two midwestern streams. We used a combination of netting and perches to manipulate predation by Great Blue Herons (Ardea herodias) and Belted Kingfishers (Ceryle alcyon), and measured the response in the fish assemblage. Bird exclusions caused significant increases in medium size classes of two common prey, striped shiners (Luxilus chrysocephalus) and central stonerollers (Campostoma anomalum). We show that these species of piscivorous birds can alter the abundance of common prey and thus need to be considered more fully when attempting to explain the structure of aquatic food webs.

The fitness of non-feeding adult insects depends on energy accumulated during the larval stage. Larvae of the caddisfly Asynarchus nigriculus primarily feed on plant detritus, but supplement their diet with animal material obtained through cannibalism. Habitat drying constrains development in many populations of this species, and we hypothesized that cannibalism should accelerate development to facilitate timely metamorphosis. We manipulated larval diets in a field experiment by supplementing detritus with animal material, and in a laboratory experiment by varying animal material and detritus quality (conditioned vs unconditioned). We measured the effects of dietary manipulation on larval and pupal growth and development, the timing of metamorphosis, and adult fitness correlates. The results of the laboratory experiment suggest that this species can metamorphose with a detritus-only diet, but development is extremely protracted. In the field experiment, individuals with animal material in their diet had higher larval survival, shorter larval and pupal development times, and earlier emergence dates (7-10 days), than those without a supplement. This delay in emergence should have important effects on survival in natural populations where the difference between desiccation and successful emergence can be only a few days. Dietary supplementation also affected adult body mass (30-40% increase), female fecundity (30% more eggs), and proportional allocation to different adult body parts. Our results are consistent with recent growth-development models that predict coupled (earlier emergence and larger adults) rather than tradeoff responses (earlier emergence and smaller adults) to pre-threshold manipulation of larval diets. Many detritivorous aquatic insects supplement their diets with animal material, and our data provide evidence that this supplementation can have strong effects on fitness. This type of dietary supplementation should be especially important for taxa that do not feed as adults, and in temporary habitats that impose time constraints on larval development.