The direct and indirect effects of guppies: Implications for life-history evolution in Rivulus hartii

Abstract

1.Ecological factors that alter mortality rates, such as predation, can cause evolutionary change. However, in addition to killing prey, predators can reduce prey abundance and increase food to survivors. Such indirect effects may also cause evolution. Predictions from theory that models how life histories evolve in response to increased mortality rates often change when they include indirect effects. Thus, indirect effects need to be evaluated to couple theory with natural systems. 2.Trinidadian killifish, Rivulus hartii, are found in communities with and without guppies Poecilia reticulata. Rivulus densities decline when guppies are present, which may be due to competitive or predatory interactions with guppies that increase Rivulus mortality rates. We previously showed that Rivulus from sites with guppies begin reproduction earlier and have increased reproductive allotment compared to Rivulus from sites with just Rivulus. Such divergence is inconsistent with theory that considers changes in juvenile mortality alone, but is consistent with theory that incorporates indirect effects. Here, we explored the mechanism of divergence with mark-recapture studies that compared the population biology of Rivulus between communities that are and are not sympatric with guppies. 3.Rivulus were 50% less abundant when guppies were present but guppies were not associated with increased adult mortality rates. Related experiments show that the declines in density are likely due to guppy predation on young Rivulus. Guppies do not appear to negatively impact Rivulus growth via competition. Rivulus with guppies grow >3× faster than Rivulus from sites upstream, above waterfalls that exclude guppies. If guppies competed with Rivulus for resources, then we would instead expect to see their presence be associated with a decline in Rivulus growth rates. When Rivulus were transplanted from above to below this barrier, their growth accelerated to match the residents. This response instead argues that the differences in growth are mediated by an environmental factor, likely lower population densities, which allow Rivulus from sites with guppies to grow faster. 4.These results imply that an indirect effect of guppy predation on young Rivulus, which is the presumed agent of selection, improves the fit of empirical findings with theoretical predictions. © 2010 The Authors. Functional Ecology © 2010 British Ecological Society.