To test the hypothesis that prolonging absorption of breakfast might improve the glucose tolerance of the subsequent meal served at lunch, normal male volunteers were administered the same carbohydrate in either a rapidly absorbed (sucrose, S) or slowly absorbed (sucrose with guar, S + G) form for breakfast (0800) and lunch (1145). Area under the curve (AUC) for glucose did not differ for S at breakfast vs. S + G at breakfast, although AUCinsulin for S at breakfast was greater than that for S + G at breakfast (3389 ± 608 vs. 1523 ± 246 μU · min · ml−1, P < .002). Plasma glucose and insulin profiles for the two breakfast meals differed markedly. Once S was ingested, plasma glucose and insulin returned to baseline after 120 and 160 min, respectively. However, once S + G was ingested, plasma glucose and insulin were still significantly above baseline values after 180 min. When S was eaten for breakfast, AUCgiucose for lunch was similar to that for breakfast, regardless of whether lunch consisted of S or S + G. However, if S + G was eaten for breakfast, AUCglucose for S + G or S at lunch was 44% (P < .005) and 75% of that for breakfast, respectively. Only one of five subjects who ingested S + G for breakfast failed to exhibit a fall in AUCglucose when S was eaten for lunch. The beneficial effect of prolonged absorption of breakfast on the glucose tolerance of lunch was not observed if the timing of lunch was delayed by 2 h (i.e., served at 1345). These observations suggest that in normal humans, prolonged absorption of breakfast results in prolonged elevation of plasma insulin levels and enhanced insulin sensitivity at lunch, thereby improving carbohydrate tolerance for lunch.