In a random sample of the adult population (aged 25-64 years) from Ghent and Charleroi in Belgium, dietary fat consumption was estimated using both qualitative and quantitative dietary methods. A fat consumption score (FCS) was constructed from the qualitative method, and it correlated significantly with the dietary polyunsaturated/saturated fatty acids (P/S) ratio calculated from the quantitative method (r = -0.48; p less than 0.001). Differences in FCS were then studied using univariate and multivariate analyses. In both men and women, FCS was higher in Charleroi than in Ghent. In men, FCS was inversely related with educational level and with age. In women, significant correlations were observed with marital status. No differences were found in FCS by smoking behaviour. With the coronary risk factors, low-order relationships were found, but only in men. In the middle-aged group, significant correlations were found between FCS and the fatty acid composition of the serum cholesteryl esters, in particular with the linoleate/oleate ratio.