Using the first five waves of the US Health and Retirement Study, a nationally representative survey of middle-aged persons in the USA conducted between 1992 and 2000, we assessed the association between alcohol consumption and separation and divorce (combined as divorced in the analysis) for 4589 married couples during up to four repeated 2-yr follow-up periods. We found that drinking status was positively correlated between spouses. The correlations did not increase over the follow-up period. Discrepancies in alcohol consumption between spouses were more closely related to the probability of subsequent divorce than consumption levels per se. Couples with two abstainers and couples with two heavy drinkers had the lowest rates of divorce. Couples with one heavy drinker were most likely to divorce. Controlling for current consumption levels, a history of problem drinking by either spouse was not significantly associated with an increased probability of divorce. Our findings on alcohol use and marital dissolution were highly robust in alternative specifications.