Mate Preference in Poeciliid Fish and the Quantification of Complex Color Patterns

A long-standing paradox in evolutionary biology is the maintenance of genetic variation in traits that are strongly tied to fitness. How is it that variation can be maintained in ecologically important traits when directional selection and genetic drift are expected to erode this variation? There are several evolutionary processes that are capable of maintaining population diversity above levels expected under mutation-selection-drift balance. One of these processes is negative frequency dependent selection (NFDS), wherein rare phenotypes have an inherent fitness advantage. A well-supported example of NFDS is found in the Trinidad guppy, Poecilia reticulata. Guppies exhibit extreme within population variation in male-limited color patterns and these color patterns are highly heritable. In this system, both natural and sexual selection appear to favor males bearing rare or unfamiliar color patterns. Evidence from laboratory studies suggest that female preference may be the underlying mechanism for the reproductive advantage of these rare male color patterns. However, this preference has only been assessed in laboratory-reared animals derived from a few natural and feral populations. It is uncertain whether the preference for rare and unfamiliar color patterns is exhibited by wild females, or how variable the preference is across a range of natural populations. In my dissertation work, I tested predictions based on the hypothesis that negative frequency dependent female preference promotes male color polymorphism in guppies, and I created a novel tool to evaluate variation in color patterns. I first determined whether female preference for rare and unfamiliar color patterns was prevalent in wild females from natural populations. Additionally, I tested whether there was among-population variation in this preference and whether any variation in preference was associated with variation in predator communities. Over a two-year period, I tested for negative frequency dependent female preference in 11 natural guppy populations, ten of which were paired high- and low-predation populations. I predicted that females would exhibit a preference for males bearing rare and unfamiliar male color patterns and that this level of preference would vary among populations. I found that females in natural populations do exhibit preference for rare and unfamiliar color patterns, and this behavior was consistent rather than variable among populations. This is the first study to examine female preference for both rare and unfamiliar color patterns across many natural populations. These data support the hypothesis that negative frequency dependent female preference promotes the maintenance of the male color pattern polymorphism in the guppy. In the next chapter of my dissertation, I describe a novel method for quantifying color patterns. This method was designed to enable investigation of a wide range of questions about color patterns, including those in study systems other than guppies. Current methods used to measure color patterns typically segment digital images into areas of color categories (e.g., “orange spots”, “black spots”) and quantify patterns using both frequency and area of these color pattern elements. However, color quantification in this manner ignores the spatial location of colors within the whole-body pattern. In addition, image segmentation categorizes rather than maintains the continuous nature of color. Currently, the ability to assess variation in whole color patterns is currently lacking because there are no methods available that enable spatially-explicit color sampling among images while maintaining the continuous measure of color. Therefore, I designed a method that sampled color at individual sampling points across a whole pattern; this method of sampling allows for the total variation in color of whole patterns to be quantified. I then demonstrated some applications of the pipeline using the guppy system. With this new approach to color sampling, I tested hypotheses that color patterns differ in predictable ways across predator-community types and that different river systems exhibit parallel evolution of color patterns between high- and low-predation populations. My novel method to quantify color patterns expanded the toolbox available to researchers allowing for the examination of questions left unanswered by current methods. In a third study, I tested for negative frequency dependent female preference in a close relative of the guppy, Poecilia picta. Populations of this species have varying levels of male polymorphism. Populations from the mainland of South America (Guyana) are polymorphic, having males that exhibit three different color morphs, while populations from the island of Trinidad are typically monomorphic. I tested the hypothesis that the level of female preference for unfamiliar (novel) male color patterns would predict the level of male polymorphism. To do this, I familiarized Guyanese and Trinidadian females to Guyanese males bearing one of two possible color morphs (red or standard). Although I did find population differences in some female preference behaviors, female population did not interact with the status of the male morph (familiar or novel). Results of this experiment do not support the hypothesis that negative frequency dependent female preference promotes color polymorphism in P. picta. In conclusion, my dissertation work robustly supports the hypothesis that female preference for males bearing rare and unfamiliar male color patterns produces a reproductive advantage for these males in natural guppy populations, thereby generating negative frequency-dependent selection. In order to test other hypotheses about the evolution of male color patterns in guppies, I developed a method for sampling and quantifying whole color patterns from digital images. This method is capable of addressing questions about color patterns that could not be addressed by previous methods and its use is not restricted to the guppy system. Finally, although negative frequency dependent female preference appears to be important in the maintaining variation in male guppy color patterns, this mechanism does not appear to explain color polymorphism in a related species. Overall, my dissertation research illuminates the causes of ecologically-important genetic variation in an evolutionary model system, and expands the tools available to researchers for addressing fundamental questions about color patterns.