Plasma cortisol concentrations were studied in venous blood samples taken in the early morning from a group of volunteers. Concentrations were higher when there had been mild sleep deprivation the night before the samples were taken, than after a normal night's sleep. The use of plasma concentrations of cortisol in the assessment of the relief of pre-operative anxiety and apprehension by premedicant drugs is discussed in the light of this finding. It is suggested that any attempt to use plasma cortisol concentrations as an index of the relief of pre-operative anxiety and apprehension by premedicant drugs may, in some circumstances, have to take into account the duration and perhaps quality of the patient's sleep. If this is not done the anxiolytic and soporific effects of the drugs studied may be confused.