Variation in Predator Phenology Affects Predator Performance and Prey Community Composition

Abstract

Variation in predator phenology strongly influenced the performance of four species of anuran larvae in experimental temporary ponds, the composition of the assemblage of anuran metamorphs from those ponds, and the reproductive performance of the predator. The resolution of variance by a statistical model of these communities was greatly improved when the model incorporated phenological variation. These results suggest that the predictive power of more general models of species interactions could be improved by incorporating size— or stage—specific effects. The experimental communities contained larval Rana utricularia, Scaphiopus holbrooki, Bufo americanus, and Hyla chrysoscelis. Dates of introduction of these species to experimental ponds corresponded with their breeding in nature. The predators were adult broken—striped newts (Notophthalms viridescens dorsalis). The experiment was a 2 x 4 factorial design:two levels of initial anuran density (1 x and 4 x) crossed with four predation regimes (predators absent; initially present and removed after 9 d; initially present and removed after 51 d; always present). The predator—removal treatments simulated variability observed in the timing of departure of adult newts from a natural pond. At both initial anuran densities, competition and predation interacted in determining the composition of the metamorph assemblage. The contribution of competition was greater at high initial density than at low. Competition among the four anuran species did not fall into a simple hierarchy, but formed a more complex network of effects. The presence of predators during the first weeks of the larval period of any anuran species greatly reduced the number of metamorphs of that species. Predator removal after the first few weeks of species' larval periods did not affect their final abundance, indicating that they reached body sizes that reduced their vulnerability to predation. Later breeding anurans benefited when newts were present but were removed before they reproduced, reducing competition from earlier breeding species. When earlier breeding species were exposed to newt predation, they performed better when newts were present throughout the summer than when they were removed before later breeding species reproduced. The vulnerability of these early breeding species to newt predation had greatly decreased by the time of the late newt removal. Presence of newts past this time benefited them by reducing the competitive effects of late breeders. A statistical model of community composition that included the effects of predator presence or absence and anuran initial density explained 47% of the variation in composition of the metamorph assemblage. A model that explicitly considered predator phenology explained 76%, leaving less than half as much variation unexplained. The reproductive success of newts was affected by their own phenology and by the density of anuran larvae. Longer periods of adult newt residence increased the numbers of newt offspring while decreasing offspring mean body mass. There was strong correlational evidence that the body masses of newt offspring were limited by their own density. The final masses and numbers of newt offspring were also affected by both initial anuran density and the numbers of anuran metamorphs emerging from experimental ponds. Newt offspring responded more strongly to the density of anuran species with which they had greater temporal overlap.