An epidemiological study was carried out among a random sample of women aged 18 to 69 years to examine possible determinants of plasma high density lipoprotein and total cholesterol (HDL-C and T-C). In a multiple regression analysis consumption of alcohol, fatty fish, and parental longevity showed positive associations with HDL-C, which were statistically significant. Smoking habit, sucrose consumption, and a family history of ischaemic heart disease showed significantly negative associations. In contrast, T-C was associated significantly only with age and Quetelet's index of body mass. Five variables having significant associations with HDL-C explained only 8% of the total variance; in the case of T-C 17% of the variance was explained by age and body mass.