SUMMARY1. Thirty-six subjects in an isolation unit were subjected to time shifts of 12 hr, or of 8 hr in either direction.2. The rhythms of body temperature and excretion of eight urinary constituents were studied before and after the shift, both on a usual nychthemeral routine and during 24 hr when they remained under constant conditions, awake, engaged in light, mainly sedentary activity, and consuming identical food and fluid every hour.3. The rhythms on nychthemeral routine were defined by fitting cosine curves. On constant routine the rhythm after the shift was cross-correlated with the original rhythm, either with variable delay (or advance) or with an additive mixture between this variably shifted rhythm and the unshifted or a fully shifted rhythm. The process yielding the highest correlation coefficient was accepted as the best descriptor of the nature of adaptation.4. A combination of two rhythms was observed more often for urinary sodium, chloride and phosphate than for other variables.5. Adaptation appeared to have proceeded further after westward than eastward shifts, and this difference was particularly noticeable for urinary potassium, sodium and chloride.6. Partial adaptation usually involved a phase delay, even after an eastward shift when a cumulative dAlay of 16 hr would be needed to achieve full adaptation and re-entrainment. 7. Observations under nychthemeral conditions often gave a false idea of the degree of adaptation. In particular, after an eastward shift the phase of the rhythms appeared to shift in the appropriate direction when studied under nychthemeral conditions whereas the endogenous oscillator either showed no consistent behaviour or, in the control of urate excretion, a shift in the wrong direction.8. The implications for people undergoing time shifts, in the course of shift work or transmeridional flights, are indicated.