The objective of this study was to assess the association between different types of alcoholic beverages and 34-year incidence of dementia. Among a random sample of 1,462 women aged 38-60 years and living in Göteborg, Sweden, in 1968-1969, 164 cases of dementia were diagnosed by 2002. At baseline as well as in 1974-1975, 1980-1981, and 1992-1993, the frequency of alcohol intake, as well as other lifestyle and health factors, was recorded and related to dementia with Cox proportional hazard regression, by use of both baseline and updated covariates. Wine was protective for dementia (hazard ratio (HR) = 0.6, 95% confidence interval (CI): 0.4, 0.8) in the updated model, and the association was strongest among women who consumed wine only (HR = 0.3, 95% CI: 0.1, 0.8). After stratification by smoking, the protective association of wine was stronger among smokers. In contrast, consumption of spirits at baseline was associated with slightly increased risk of dementia (HR = 1.5, 95% CI: 1.0, 2.2). Results show that wine and spirits displayed opposing associations with dementia. Because a protective effect was not seen for the other beverages, at least part of the association for wine may be explained by components other than ethanol.