Both in patients with primary (n = 15) and secondary (n = 10) adrenal insufficiency a conventional substitution with 30 mg cortisol daily (3 single doses of 15, 10 and 5 mg) leads to an unsatisfactory adjustment in most cases. Plasma maximal values after the morning dose were obtained only between 9 and 11 a.m. and were 3.5 times higher than cortisol values of healthy probands (n = 35). The half-life of cortisol in plasma (t1/2 = 0.8-3.2 hrs) varied to such an extent that some patients had subnormal values relatively quickly or that concentrations above normal were found permanently. The same also applies to free cortisol in plasma. The amount of free cortisol in urine cannot be used as a parameter of a physiological substitution of cortisol. The individual adjustment of the treatment plan can thus be only established according to cortisol concentrations in plasma.