UBC Theses and Dissertations
Analysis of geographic variation in the antipredator adaptations of the guppy : Poecilia reticulata Seghers, Benoni Hendrik
The main objective of this study was to describe and explain several features of geographic variation among isolated and semi-isolated populations (Trinidad, West Indies) of the guppy, Poecilia reticulata. Three main aspects of geographic variation were considered: (i) sex ratio, (ii) body size, and (iii) antipredator behavior. (i). Extreme deviations (favouring females) from a theoretical Mendelian sex ratio were correlated with the presence of dense populations of a small cyprinodontid predator, Rivulus hartii. Laboratory experiments revealed that this variation was not caused by genetic differences in the sex determination system. In addition, sex ratios were not correlated with sexual dimorphism in colour. Predation experiments with Rivulus demonstrated that male guppies were not selectively attacked but were less adept at avoiding capture. Size-selective predation by Rivulus also placed males at a selective disadvantage. Whether conspicuous coloration increases the liability of males to predation has yet to be demonstrated unequivocally. (ii) Populations of guppies sampled in 1967 and 1969 showed a stable pattern of variation in body size; differences of over 41% in body length and 200% in weight were discovered. In compliance with Bergmann's Rule there was a significant negative correlation between body size and stream temperature. Though a substantial portion of the size variation can be explained as a direct phenotypic response to environmental differences, there is also good evidence for microevolutionary differences. Of a multitude of potential selective factors that might be responsible for this genetic diversity, only one, size-selective predation, was investigated. Field and laboratory evidence supported the hypothesis that large guppies enjoy an advantage with respect to Rivulus predation but are more vulnerable to large characid and cichlid predators such as Hoplias malabaricus and Crenicichla alta. In the laboratory, size-selective predation appeared to be caused by differences in the handling efficiency of the predators, however, in nature the interaction of several other factors must be considered. (iii) Field observations revealed that where characid and cichlid predators were present (and Rivulus absent) guppies, (a) were more restricted to the stream shore, (b) showed a greater tendency to school, (c) avoided a potential predator at a greater distance, and (d) had a lower alarm threshold. To assess the functional and evolutionary significance of this behavioral variation, predation experiments were conducted with samples of wild-caught and predator-naive (laboratory-bred) guppies originating from 5 natural populations. These tests demonstrated that fish either taken, or descended, from populations exposed to characids and cichlids were relatively less vulnerable than those exposed to Rivulus. To determine why the predator-naive guppies were selected non-randomly, a comparison was made of their habitat preferences, schooling behavior, reaction distance, and escape motor patterns. Significant interpopulation differences were found for several of the measures; generally, these were consistent with the field observations. In addition, it was hypothesized that certain populations may learn to avoid predators more rapidly. It was concluded that much of the geographic variation in antipredator behavior is caused by genetic differences attributable to differential predation pressures. In some cases, these microevolutionary differences are apparently maintained without a major barrier to gene flow.