On the basis of our own experimental data and those from the literature, the reproducibility of day-night variations in physiological functions--their 'circadian pattern'--is examined in five sections. They concern the problem of intra- and interindividual variability, demonstrated by the rhythm of deep body temperature, the dependence of the pattern on experimental conditions, and as further examples circadian rhythms in the endocrine system, in blood constituents, in blood pressure and in psychomotor performance. The survey is considered to contribute to the question of how important the circadian temporal structure may be in the search for reference values and hence in diagnosis.