Sexually selected traits are limited by selection against those traits in other fitness components, such as survival. Thus, sexual selection favouring large size in males should be balanced by higher mortality of larger males. However, evidence from red-winged blackbirds (Agelaius phoeniceus) indicates that large males survive better than small males. A survival advantage to large size could result from males migrating north in early spring, when harsh weather favours large size for energetic reasons. From this hypothesis we predicted that, among species, sex differences in body size should be correlated with sex differences in timing of spring migration. The earlier males migrate relative to females, the larger they should be relative to females. We tested this prediction using a comparative analysis of data collected from 30 species of passerine birds captured on migration. After controlling for social mating system, we found that sexual size dimorphism and difference in arrival dates of males and females were significantly positively correlated. This result is consistent with the hypothesis that selection for survival ability promotes sexual size dimorphism (SSD), rather than opposes SSD as is the conventional view. If both natural selection and sexual selection favour large adult males, then limits to male size must be imposed before males become adults.