Main content area

Interactive Effects of Predators and Early Larval Dynamics of the Treefrog Hyla Chrysocelis

Fauth, John E.
Ecology 1990 v.71 no.4 pp. 1609-1616
Cambarus, Hyla, Notophthalmus viridescens, crayfish, food webs, freshwater, intraspecific competition, models, ponds, predators, prediction, salamanders and newts, tadpoles
Ecologists need to determine the frequency and nature of higher order interactions and to predict when such interactions will be important to understand community dynamics. I experimentally tested for higher order interactions in a freshwater food web by independently manipulating the presence and absence of two predators (the salamander Notophthalmus viridescens and the crayfish Cambarus bartonii) and their tadpole prey (Hyla chrysoscelis) in a series of artificial ponds. By destructively sampling selected ponds early in the larval period (days 5 and 6), I hoped to predict whether the predators would have an interactive effect on tadpole survival and growth on day 30. Early destructive samples adequately predicted subsequent influences of interactions on the survival, but not on the size of Hyla. Early samples predicted that the predators would interact to determine tadpole survival and size. The two predators did not act additively to determine tadpole abundance: more survived in the presence of both predators than was predicted. This higher order interaction was most likely caused by physical interference between the predators and predator avoidance by the tadpoles. Early samples correctly predicted higher order interactions would be seen on tadpole growth, but completely failed to predict their direction because the biological interactions changed over time. Early in their larval period, tadpoles behaviorally avoided predators, which caused slower growth, but later, tadpoles in predator treatments grew faster due to diminished intraspecific competition. These results suggest that higher order interactions may be important in nature. Short—term experiments may prove useful in predicting when higher order interactions will occur, but detailed models may be required to predict their outcome accurately.