A geographically based sample of 1069 Hispanic and non-Hispanic white persons aged 20-74 years, living in southern Colorado and who tested normal on an oral glucose tolerance test (World Health Organization criteria) were evaluated to determine associations of dietary factors with fasting serum insulin concentrations. Subjects were seen for up to three visits from 1984 to 1992. A 24-h diet recall and fasting insulin concentrations were collected at all visits. In longitudinal data analysis, lower age, female gender, Hispanic ethnicity, higher body mass index, higher waist circumference, and no vigorous activity were significantly related to higher fasting insulin concentrations. High total and saturated fat intake were associated with higher fasting insulin concentrations after adjusting for age, sex, ethnicity, body mass index, waist circumference, total energy intake and physical activity. Dietary fibre and starch intake were inversely associated with fasting insulin concentrations. No associations with fasting insulin concentrations were observed for monounsaturated fat, polyunsaturated fat, sucrose, glucose and fructose intake. Associations were similar in men and women and for active and inactive subjects, though associations of fibre and starch intake with insulin concentrations were strongest in lean subjects. These findings support animal studies and a limited number of human population studies which have suggested that increased saturated and total fat intake and decreased fibre and starch intake increase fasting insulin concentrations and may also increase insulin resistance. These findings, which relate habitual macronutrient consumption to hyperinsulinaemia in a large population, may have implications for studies attempting primary prevention of non-insulin-dependent diabetes mellitus.