Safran et al. (2016a) manipulated two sexual traits (ventral plumage coloration and tail streamer length) in male barn swallows (Hirundo rustica) and reported divergent effects on paternity change between two study populations, in Colorado and Israel. They concluded that geographical variation in the two phenotypic traits is maintained by divergent sexual selection. However, the response variable they used, the longitudinal change in paternity from a pre-treatment clutch to a post-treatment clutch, does not reflect an unbiased effect of the treatment. Here, I show that the magnitude of the change in paternity is influenced by variation in the initial paternity score among the treatment groups, which is presumably due to stochastic variation from low sample sizes in the treatment groups. When the bias was accounted for in re-analyses of the Israeli dataset, the statistical significance of one of two treatment effects disappeared. Similar re-analyses of the American population were not possible due to inaccessibility of raw data for individual clutches, but an assessment of the mean scores indicates that the two significant treatment effects in this population were similarly biased in their initial paternity scores. The conclusion of divergent sexual selection on male phenotypic traits between the two populations does not seem to be supported.
Keywords: longitudinal analysis; mate choice; paternity; phenotype manipulation.
© 2022 The Authors. Evolution published by Wiley Periodicals LLC on behalf of The Society for the Study of Evolution.