Information about drinking practices has been obtained by questionnaire from 1,984 monozygotic and dizygotic adult female twin pairs from the Australian twin register, including 1,690 pairs where both twins have used alcohol. Statistical analyses of these data show that marital status is an important modifier of genetic effects on drinking habits. In young twins, aged 30 years or less, genetic differences between individuals account for only 31% of the variance in alcohol consumption of married respondents, but for 60% of the variance of unmarried respondents. In twin pairs, aged 31 years or more, genetic differences account for 46-59% of the variance in married twins, but for 76% of the variance in unmarried twins. In our young sample (average age 35 years) there is no evidence that individuals genetically predisposed to heavy drinking are any less likely to be married than the rest of the population. Some alternative explanations of these findings are also rejected.