The circadian and ultradian variations of blood glucose and plasma insulin have been characterized individually and as a group phenomenon in five healthy young adults studied while adhering as closely as possible to their usual routine of sleep, activity, meal content and timing. Three complementary methods were used to analyze the data: displaying raw data as a function of time; cosinor method according to Nelson and Halberg; and time series analyses as proposed by De Prins and Malbecq. The subjects were studied in the laboratory and their life routine were controlled, but very close to that of their habitual routine. They had mainly ultradian rhythms of blood glucose (mainly about 6 hr) and circadian rhythms of immunoreactive insulin (I.R.I.). Blood glucose ultradian rhythms seem to be mainly but not exclusively mealtime dependent, while I.R.I. circadian rhythms appear to be primarily endogenous in origin. Therefore, the role played by insulin in the control of blood glucose levels seems to be programmed on a circadian basis rather than by a time independent feedback phenomenon as postulated by the conventional homeostatic hypothesis. The advantage of this chronophysiologic approach is to consider circadian rhythms of both I.R.I. and insulin effectiveness as an adaptive phenomenon able to maintain blood sugar changes in the ultradian domain of rhythms.