Bi-directional sex change: testing the growth-rate advantage model

Bi-directional sex change in coral-dwelling fishes (genera Gobiodon and Paragobiodon) has been attributed to a growth-rate advantage for females during the non-breeding season. This model predicts that the smallest individual in a newly formed pair should always be female. To determine if a growth-rate advantage exists for female Gobiodon histrio, I monitored the growth of males and females in natural pairs during the breeding and non-breeding season. I then used a manipulative field experiment to test four predictions of the growth-rate advantage model: (1) the larger individual should change sex to male in new pairs containing two females; (2) the smaller individual should change sex to female in new pairs containing two males; (3) neither individual should change sex in heterosexual pairs where the male is larger than the female; and (4) both individuals should change sex in heterosexual pairs where the female is larger than the male. A growth-rate advantage was detected for female G. histrio during the non-breeding season; however, only the first three of the predicted outcomes were observed in the manipulative experiment. Sex change did not occur in heterosexual pairs where the female was larger than the male. Furthermore, growth did not differ between sex-changed and non-sex-changed fish; therefore, the absence of sex change in these pairs is not due to a growth cost to sex change. I propose that the risk of moving among spatially isolated habitat patches and the low probability of finding a mate have been more important than sex-specific differences in growth rates to the evolution of bi-directional sex change in coral-dwelling gobies.