Boys and girls may display different styles of aggression. The aim of this study was to identify subtypes of aggression within the Child Behavior Checklist (CBCL) aggression scale, and determine their characteristics for both sexes. Maternal CBCL ratings of 7449 7-year-old twin pairs were analyzed using principal components analyses to identify sub- types of aggression, and structural equation modeling to carry out genetic analyses. Two aggression subtypes were identified: relational and direct aggression. The correlation between these subtypes was .58 for boys and .47 for girls. Boys had higher mean scores for both subtypes of aggression, but sex differences were largest for direct aggression. For relational aggression, 66% of the variance was due to additive genetic influences, 16% to shared environment and 18% to nonshared environment. For direct aggression, additive genetic effects accounted for 53% of the variance in males and 60% in females, shared environment explained 23% of the variance in males and 13% in females, and nonshared environmental effects explained 24% of the variance in males and 27% in females. Covariance between the aggression subtypes was mostly accounted for by additive genetic (55% for boys, 58% for girls) and shared environmental influences (33% for boys, 30% for girls). Direct and relational aggression were both influenced by one underlying set of shared environmental factors, but only partly by the same genes (the genetic correlation was .54 for boys and .43 for girls). These findings may have implications for how aggressive behavior should be assessed in boys and girls.