Other studies based on multifactorial measures of aggression, such as the BDHI, suggest some variability in the heritability estimates across subscales, although the patterns are not entirely consistent across studies. For example, “indirect hostility” showed the lowest heritability (28%) in one study of adult twins (Coccaro et al., 1997), compared to modest heritability for “verbal hostility” (47%) and “assault” (40%). Yet, Cates et al. (1993) found no genetic influences for assault, but strong heritability for both verbal hostility (78%) and indirect hostility (70%). In multivariate genetic analyses, both studies found some support for genetic specificity for the various subscales, similar to what has been found for reactive and proactive aggression, in that genetic correlations (rG) among the BDHI subscales were less than unity: rG = 0.39 between indirect hostility and assault, rG = 0.60 between indirect hostility and verbal hostility, rG = 0.17 between verbal hostility and assault (Coccaro et al., 1997), rG = 0.35 between indirect hostility and assault, rG = 0.39 between indirect hostility and verbal hostility, and rG = 0.49 between verbal hostility and assault (Cates et
