In 1986-1987, the authors assessed the fatty acid intake of 118 Boston-area men, aged 40-75 years, by a semiquantitative food frequency questionnaire administered twice, by two 7-day diet records, and by capillary gas chromatography of subcutaneous fat samples obtained by needle aspirate from the lateral buttock. Spearman correlation coefficients between diet record estimates of fatty acid intake (as a percentage of total fat) and fat aspirate measures (as a percentage of total peak area) were as follows: saturated fat, r = 0.16 (p = 0.09); monounsaturated fat, r = 0.22 (p = 0.01); and polyunsaturated fat, r = 0.49 (p = 0.0001). Spearman correlation coefficients between estimates derived from the food frequency questionnaire were as follows: saturated fat, r = 0.18 ( p = 0.05); monounsaturated fat, r = 0.14 (p = 0.14); polyunsaturated fat, r = 0.50 (p = 0.0001); and eicosapentaenoic acid, r = 0.47 (p = 0.0001). These data confirm that the polyunsaturated and eicosapentaenoic fatty acid content of subcutaneous fat is a measure of dietary intake of these fats. Although diet records are commonly thought to be the "gold standard" method of dietary assessment, the similar correlations observed between the fatty acid composition of adipose tissue and estimates of intake from the food frequency questionnaire and from diet records suggest that these two dietary assessment methods have similar validity in the measurement of polyunsaturated fatty acid intake.