workers. Urine samples have been collected from day and night shift workers in a light engineering factory over periods of 24 or 48 hours during which blood was also collected for determination of plasma 11-hydroxycorticosteroids (11-OHCS). Different subjects divided their 24-hour output into as few as 4 or as many as 14 portions. Where possible, the timing of the excretory rhythms was assessed objectively by fitting a sine curve. The potassium excretory rhythm was, in most subjects, well adapted to night work in that they excreted least during their hours of sleep; no relationship could be discerned between potassium excretory and plasma 11 -OHCS rhythms. The sodium excretory rhythm was less regular than that of potassium in both night and day workers, and in night workers both sodium excretion and urine flow were often high during the hours of sleep, so that the subjects were sometimes awakened by the need to micturate. There was, however, sufficient association between the behaviour of sodium and potassium to suggest that their excretory rhythms have a common cause, and that this is not the secretion of 11-OHCS. Phosphate excretion fell sometimes on rising, even though plasma 11-OHCS concentration was not high, and sometimes after a high level of plasma 11-OHCS, at a time other than on rising. This suggests that these are two independent contributory causes of the usual morning fall in phosphate excretion.In our previous paper (Conroy, Elliott, and Mills, 1970) we compared the rhythms in plasma corticosteroid concentration of a group of night shift workers in a light engineering works with those of a group of day workers in the same factory. The subjects produced serial urine samples on the same occasions, and their electrolyte excretory rhythms are described in this paper.Subjects and experimental procedure The designation of the groups of subjects is the same as in our previous paper.Group la Eleven night workers were provided with 'Present address: Department of Physiology, Royal College of Surgeons in Ireland, St. Stephen's Green, Dublin 2. measuring cylinders and small bottles and were asked to record the time and volume of each urination over a period of 24 hours, and to provide us with a sample for analysis.Group lb Fifteen night workers, including three from group la, were asked similarly to collect serial urine samples over a period of 48 hours.Groups 2a and b Eight and 12 day workers in the same factory were similarly asked to produce urine samples over continuous periods of 24 hours and 48 hours.The frequency of voiding varied widely between subjects, so that we received as few as 4 or as many as 14 samples to cover a 24-hour period; some subjects failed to produce samples to cover the whole period requested, and a few samples were lost, so that we had information about the volume but not the composition. The number 356