Relationships between diurnal plasma cortisol peaks and meals were evaluated for 30 male subjects divided into 5 groups. At 1300 h, at the time of a slow increase of plasma cortisol in fasting subjects, a reproducible rapidly increasing meal-related peak appeared in all subjects studied. An identical meal at 2000 h led to a lower mean response, with larger interindividual variations. This attenuated evening response does not seem attributable to any daily rhythm in responsiveness nor to changing basal levels, since only slight nonsignificant rises in cortisol appeared after an identical meal at 1000 h. The usual mean cortisol pattern with a midday peak was observed in subjects accustomed to different activity and meal-time schedules, which excludes the role of dietary habits. Satiety did not seen to play a determining role in the response to the 1000 h meal, as was shown by comparing subjects who did or did not have breakfast after overnight fasting. Meal intake was not necessary to provoke peaking in cortisol levels, and it has been established that neural and behavioral factors associated with meal presentation play a predominant role in some subjects. The results given clear evidence of the influence of meal timing on the daily plasma cortisol pattern, but no clue was found as to why eating affects the pituitary-adrenal axis differently according to the time of day. The noon meal may at least have a synchronizing role on normally existing plasma cortisol fluctuations.