Circadian phase estimates and their 95% confidence intervals can be obtained by curve-fitting with the electronic computer. Such diagrams assess objectively the external and internal timing of circadian rhythms in oral temperature, in urine volume and in the urinary excretion of sodium, potassium, 17-hydroxycorticosteroids and 17-ketosteroids. A reference to a precise time of circadian crest constitutes a risky undertaking whenever it rests on short stretches of data covering a day or two of study on a single subject rather than a group. Inferences drawn in this study are based upon the behaviour of »ensembles« of phases, rather than upon values for a single given crest, and they must be viewed with the realization that they were obtained with serially dependent sampling in the case here presented. With such qualifications, the ensembles serve to make several points. Under the specified conditions, including a diurnal-activity-nocturnalrest schedule, most circadian crests of 17-hydroxycorticosteroid excretion occurred around noon. The circadian crests of 17-ketosteroid excretion occurred, on the average, about one hour or so before noon. These two urinary functions and three others show a consistent lead in phase in relation to the oral temperature rhythm. The typical crest values, if they are provided with their confidence arc, represent objective quantitative endpoints for endocrine studies, among others.