Role of Indirect and Direct Genetic Effects in Modification of Behavior and Maintenance of Color Polymorphism in Male Gambusia Holbrooki

Indirect genetic effects (IGEs) on traits are present if variation in the social environment provided by others is at least partially due to genetic variation among those individuals. IGEs can result from a variety of interactions between conspecifics and may contribute to the high phenotypic variation in and rapid evolution of behavior. "GxG" epistasis", also known as direct-by-indirect genetic interaction, is a potentially widespread type of IGE that occurs when the expression of behavior in a social context depends on the genotypic combination of the individuals(s) creating the social contexts for a behavior and the individual expressing that behavior. Despite the potential widespread importance of IGEs, few studies have evaluated IGEs on natural genetic variation in behavior occurring in its natural context. The Eastern mosquitofish, Gambusia holbrooki, is a tractable system in which to study IGEs on behavior. Male Eastern mosquitofish exhibit a discrete natural color polymorphism: a majority of males are silver (S), while a minority is mottled black-and-white (M). This color polymorphism is associated with variation in behavior, responses from conspecifics, survival rate, and selective advantages in the presence of predators. These findings suggest that variation in the social environment provided by others may affect the behavioral differences between M and S male morphs. To determine whether variation in the social environment (IGEs) is a critical influence on differences in male morph behavior, we conducted three studies. In the first study, we compared association with social partners of M and S males in both laboratory and artificially constructed social groups. We found that M and S males associated nonrandomly with different partners, suggesting that these male morphs experience different immediate social environments that could result in IGEs on behavior. We evaluated the role of IGEs on differences in M and S behavior in the second study, in which we compared differences in behavior of male morphs placed in social contexts that varied in the genotypic composition of social partners. We found that social context affected some behavioral differences between M and S males, but that the genotype of the focal individual also influenced focal male behavior. In the third study, we evaluated how M and S juvenile behavior changed during ontogeny in response to different social environments, and whether IGEs seen in adult mosquitofish have their developmental origin in social experiences during development. Results showed that IGEs strongly affected juvenile behavior and maturation: juveniles were more interactive with social partners, fed more, and reached maturity faster when reared with S male adults compared to those reared with M males. There was additional evidence that M and S juveniles behaved differently depended on their rearing context. In a fourth study, we also evaluated M and S male responses to predator and conspecific odor cues; male morphs may respond to odor cues in a way that promotes differences in their relative survival and informs association with social partners. While we found some evidence that males respond differently to odor cues, this evidence was limited to one replicate of the study. In two studies, we also compared association and interactions with social partners in both freely interacting wild social groups and constructed laboratory environments to evaluate whether IGEs influenced differences in male morph behavior similarly in natural and artificial conditions. As described above, we examined patterns of association with social partners between M and S male morphs; we found similar nonrandom association trends between both laboratory and wild social groups, though these trends were nonsignificant for field data. We also compared the behavior of freely interacting individuals in two wild populations using underwater video to behavior of male morphs in a gradient of laboratory social environments. Results showed that both the genotype of the focal individual, and the genotype of their social partner, affected behavior in both wild and laboratory settings. These results implicate that both IGEs and GxG epistasis are important influences on behavior across both settings. In conclusion, this work is some of the first to describe the role of IGEs on differences in behavior associated with a natural polymorphism in its natural context. We establish that males differing in a genetically determined color pattern associate nonrandomly with social partners, and that the behavior of these males depends on both their own genotype and their genotypes of their social partners. We draw comparisons between behavior in laboratory and natural social environments, suggesting that the Eastern mosquitofish is a tractable system in which to continue to explore the role of the social environment on behavior.