Objectives: To examine whether a number of nutritional and familial factors were associated with menopausal development. Methods: A prospective postal survey amongst a random sample of 1227 women aged 47 to 51 who were premenopausal in a cross-sectional survey 2 years previously. Women were classed into three groups. premenopause (regular menstruation); irregular menstruation; postmenopausal (absence of menstrual cycle for at least 6 months). Proportional odds regression was used to identify those factors which were independently predictive of subsequent menopausal development. Results: There was an 80% (n = 983) survey response rate. After exclusion of current HRT users (n = 178), 150 (19%) women were postmenopausal, 277 (34%) had erratic menstruation and 378 (47%) were premenopause. There were significant univariate associations between menopausal status and age (P < 0.001), age of maternal menopause (P = 0.006), alcohol consumption (P = 0.005) and social class (P = 0.03). Maternal age and alcohol consumption were significantly correlated with estradiol levels (r = 0.45, P = 0.02, and r = 0.61, P = 0.02 for maternal age and alcohol consumption, respectively). In proportional odds regression analysis, age, maternal menopausal age, alcohol consumption and smoking were independently associated with menopausal status. Conclusions: These results suggest that, (1) there is a strong familial association in menopausal age, and (2) moderate consumption of alcohol is associated with delayed menopausal development. Copyright (C) 1997 Elsevier Science Ireland Ltd.