In contrast to the high genetic correlations, the overlap across substances in non-shared environmental factors was negligible for both use and dependence symptoms (rE = 0.298 and 0.246, respectively), indicating that the individual-specific environmental conditions that promote use and related problems vary across the two substances. This conclusion does not appear to be in keeping with evidence in the larger literature of common risk factors for alcohol and cannabis-related behaviors. This apparent discrepancy has three likely explanations. First, the low degree of overlap in non-shared environmental factors may be attributable to substance-specific measurement error. Second, many of the well-established risk factors, including externalizing behaviors and parental substance-related problems (Elkins et al., 2007; Fergusson et al., 2008; Hayatbakhsh et al., 2008; Kuperman et al., 2001; Monshouwer et al., 2006; Slutske et al., 1998) are themselves highly heritable (Knopik et al., 2005; Krueger et al., 2002; McGue et al., 2006; Sherman et al., 1997; Slutske et al., 1997). Moreover, they share a substantial proportion of genetic variance with alcohol and cannabis-related phenotypes (Hicks et al., 2004; Miles et al., 2002; Slutske et
