Plasma cortisol concentrations increased after injections of nicotine. This increase was variable and was no greater than that seen after substitution of normal saline for nicotine. Plasma cortisol concentrations did not increase after nicotine in patients with hypopituitarism and after inhibition of corticotrophin release with triamcinolone. It is concluded that the increased plasma cortisol concentration after injection of nicotine was due to non-specific pituitary stimulation associated with the experimental procedure and not due to any direct effect of the nicotine. Nicotine resulted in a similar increase in plasma cortisol in four patients with diabetes insipidus. Neither rapid infusion of hypertonic mannitol nor ingestion of ethanol had a consistent effect on the plasma cortisol concentration. It is therefore concluded that the antidiuretic hormone is not the 'corticotrophin release factor' in man.