Dietary intake information was collected on 2,426 rural and 5,828 urban men free of coronary heart disease in the Pureto Rico Heart Health Program. Serum cholesterol and triglyceride, body weights, and related parameters were measured during a detailed cardiovascular examination. There were statistically significant differences between rural and urban values with respect to lipids, relative weight, and intake of many nutrients. Rural subjects had lower lipids and relative weights. Serum cholesterol was associated with relative weight in both rural and urban areas. In the urban areas, serum cholesterol was associated with intake of total fat, saturated fatty acids, polyunsaturated fatty acids, total carbohydrate, and starch (all measured as percent of calories). It was also related to percent carbohydrate from starch and dietary cholesterol intake, even after relative weight was taken into account. The associations were of a low order, although statistically significant. The association of fasting triglycerides with some of the carbohydrate variables was statistically significant after relative weight was taken into account in the urban area. Overall, diet and relative weight can account for at most 6% of the variability in serum cholesterol observed, with at most 2.5% of the variability due diet alone.