Various environmental variables are hypothesized to operate differentially within identical and fraternal twin pairs. To the extent that these factors are correlated with behavioral outcomes, such as alcohol or drug abuse, traditional twin studies of concordance may be biased. Self-ratings of within-pair emotional closeness, assessed in 169 same-sex twin pairs ascertained through alcohol and drug treatment centers, were used to determine the impact of the twin relationship on concordance for alcohol dependence ( N=130 twin pairs) and other drug abuse and/or dependence ( N=85 twin pairs). In general, identical twin pairs reported significantly closer relationships than fraternal twin pairs, and female twin pairs reported significantly closer relationships than male twin pairs. The data did not indicate an overall effect of closeness on co-twin risk for alcohol dependence. In contrast, closeness was significantly related to co-twin risk for other drug abuse and/or dependence. However, the MZ/DZ concordance difference for other drug abuse and/or dependence remained significant when the effects of within-pair closeness were controlled. Thus, the initial zygosity and sex differences in concordance for substance use disorders cannot be explained solely by differences in twin relationship due to closeness as assessed in this study. © 1997 Elsevier Science Ireland Ltd.