RESUMEN
Both sexually selected traits and mate preferences for these traits can be context dependent, yet how variation in preferred traits could select for context dependent preferences has rarely been examined. The signal reliability hypothesis predicts that mate preferences vary across contexts (e.g., environments) in relation to the reliability of the information preferred traits provide in those contexts. Extensive variation in copy number of mc4r B alleles on the Y-chromosome that associates with male size in Xiphophorus multilineatus allowed us to use a split-sibling design to determine if male size is more likely to provide information about male genotype (i.e., dam) when males were reared in a warm as compared to a cold environment. We then examined strength of preference for male size by females reared in the same two environments. We found that males were larger in the cold environment, but male size was more variable across dams in the warm environment, and therefore male size would be a more reliable indicator of dam (i.e., genetics) in the warm environment. Females reared in the warm environment had stronger mate preferences based on male size than cold reared females, with a significant influence of dam on strength of preference. Therefore, strength of female preference for male size was influenced by the temperature in which they were reared, with the direction of the difference across treatments supporting the signal reliability hypothesis. Understanding how the reliability of male traits can select for contextual variation in the strength of the female mate preferences will further our discovery of adaptive mate preferences. For example, a relationship between the strength of a female's mate preference and their growth rates was detected in the context where females had a preference based on male size, supporting a hypothesis from previous work with this species of disassortative mating in relation to growth rates to mitigate a documented growth-mortality tradeoff.