No abstract

One of the predictions of the 'good genes' model of sexual selection is that reproductively successful males with well-developed indicator traits should show smaller variances for non-indicator traits, that are not directly associated with mating success, when compared to non-breeding males and females. Thus sexual selection should reinforce stabilizing natural selection in reducing the variance in quantitative traits. This prediction is tested by analysing variation in eight morphological traits of breeding males, non-breeding males, and females of pupfish (Cyprinodon pecosensis). Breeding males tended to be less variable than non-breeding males for all principal component factors, and for all morphological traits except for depth, although these differences were statistically significant only for PC2, and PC5 and for pelvic fin length, number of pelvic fin rays and number of preopercular and preorbital pores. Similarly, breeding males tended to be less variable than females for all principal component factors and for all morphological traits except for number of preopercular pores. These differences were statistically significant for PC2, and for depth, pelvic fin length, number of preorbital pores and pectoral fin rays. The overall pattern of reduced variability in independent traits of breeding males revealed by principal component analysis is very consistent and highly significant (P< These results support the prediction of the 'good genes' model and show that reproductively active males are subject to more severe stabilizing selection for several quantitative traits than non-breeding males and females. Thus sexual selection, through male-male competition, female choice, or an interaction of both selective processes, results in stabilizing selection on quantitative morphological traits.