LIFE-HISTORY EVOLUTION IN GUPPIES (POECILIA RETICULATA): 1. PHENOTYPIC AND GENETIC CHANGES IN AN INTRODUCTION EXPERIMENT.

Previous investigations (Reznick and Endler, 1982; Reznick, 1982a, 1982b) demonstrated that genetic differences in guppy life histories were associated with differences in predation. Guppies from localities with the pike cichlid Crenicichla alta and associated predators matured earlier, had greater reproductive efforts, and produced more and smaller offspring than did guppies from localities with only Rivulus harti as a potential predator. Crenicichla preys primarily on large, sexually mature size-classes of guppies, while Rivulus preys primarily on small, immature size-classes. These patterns of predation are hypothesized to alter mean age-specific survival. Theoretical treatments of such differences in survival predict the observed trends in age at maturity and reproductive effort. We are using introduction experiments to evaluate the role of predators in selecting for these life-history patterns. The experiment whose results are presented here was conducted in a tributary to the El Cedro River (Trinidad), where a waterfall was the upstream limit to the distribution of all fish except Rivulus. Guppies collected from the Crenicichla locality immediately below the waterfall (the downstream control) were introduced over the waterfall in 1981. This introduction released the guppies from Crenicichla predation, exposed them instead to Rivulus predation only, and also introduced them to a different environment, since the introduction site has greater canopy cover than the site of origin. Changes in guppy life-history patterns can be attributed to predation and/or the environment. Evidence from fish collected and preserved in the field demonstrated that, by mid-1983, guppies from the introduction site above the waterfall matured at larger sizes and produced fewer, larger offspring. There were no consistent differences in reproductive allotment (weight of offspring/total weight). With the exception of reproductive allotment, these patterns are identical to previous comparisons between Rivulus and Crenicichla localities. A laboratory genetics experiment demonstrated that males from the introduction site matured at a later age and at a larger size than did males from the control site downstream, as predicted from the "age-specific predation" hypothesis. No differences between localities were observed for female age and size at maturity or for reproductive effort. The trends for fecundity and offspring size were the reverse of those observed in the field. Because only the males changed in the predicted fashion, it is not possible either to reject or to accept the hypothesis of age-specific predation at this time. We discuss the possible causes for these patterns and the high degree of plasticity in the life history, as evidenced by the differences in fecundity and offspring size between the field and laboratory results.