Twenty-four middle-aged healthy men were given a low-fat high-carbohydrate (5.5 g fat; L), or a moderately-fatty, (25.7 g fat; M) breakfast of similar energy contents for 28 d. Other meals were under less control. An oral glucose tolerance test (OGTT) was given at 09.00 hours on day 1 before treatment allocation and at 13.30 hours on day 29. There were no significant treatment differences in fasting serum values, either on day 1 or at the termination of treatments on day 29. The following was observed on day 29: (1) the M breakfast led to higher OGTT C-peptide responses and higher areas under the curves (AUC) of OGTT serum glucose and insulin responses compared with the OGTT responses to the L breakfast (P < 0.05); (2) treatment M failed to prevent OGTT glycosuria, eliminated with treatment L; (3) serum non-esterified fatty acid (NEFA) AUC was 59% lower with treatment L than with treatment M, between 09.00 and 13.20 hours (P < 0.0001), and lower with treatment L than with treatment M during the OGTT (P = 0.005); (4) serum triacylglycerol (TAG) concentrations were similar for both treatments, especially during the morning, but their origins were different during the afternoon OGTT when the Svedberg flotation unit 20-400 lipid fraction was higher with treatment L than with treatment M (P = 0.016); plasma apolipoprotein B-48 level with treatment M was not significantly greater than that with treatment L (P = 0.086); (5) plasma tissue plasminogen-activator activity increased after breakfast with treatment L (P = 0.0008), but not with treatment M (P = 0.80). Waist:hip circumference was positively correlated with serum insulin and glucose AUC and with fasting LDL-cholesterol. Waist:hip circumference and serum TAG and insulin AUC were correlated with factors of thrombus formation; and the OGTT NEFA and glucose AUC were correlated. A small difference in fat intake at breakfast has a large influence on circulating diurnal NEFA concentration, which it is concluded influences adversely glucose tolerance up to 6 h later.