The main objective of this study was to evaluate whether daily consumption of caffeine-containing beverages is related to the prevalence and severity of premenstrual syndrome apart from any effects of daily total fluid consumption. A secondary objective was to determine whether daily total fluid consumption itself is related to premenstrual syndrome. The study is based on 841 responses to a questionnaire probing menstrual and premenstrual health, and daily fluid consumption, which was mailed to female university students in Oregon. Analysis of the data revealed that consumption of caffeine-containing beverages was strongly related to the prevalence of premenstrual syndrome. Among women with more severe symptoms, the relation between consumption of caffeine-containing beverages and premenstrual syndrome was dose-dependent, with prevalence odds ratios equal to 1.3 for consumers of one cup of a caffeine-containing beverage per day and increasing steadily to 7.0 for consumers of eight to 10 cups per day. The effects were apparent among both caffeine-containing tea/coffee consumers and caffeine-containing soda consumers. The observed effects were only slightly reduced when daily total fluid consumption was controlled. Daily total fluid consumption also was related to the prevalence of premenstrual symptoms although the effects were large only for consumers of 13-19 cups of fluid per day (the largest amount studied).