Parameters of cortisol metabolism were studied in pregnancy and early post partum and in non pregnant women. Total serum cortisol was measured by a specific radioimmunoassay and its unbound- and CBG-bound fractions by equilibrium dialysis at 37 degrees C. CBG was measured directly, by an immunodiffusion method (true CBG) and indirectly, by the Scatchard analysis of its interaction with labelled cortisol (PT). During pregnancy the increases in true CBG and total cortisol concentrations were grossly parallel but the PT/true CBG ratio fell from almost 1 in non-pregnant and delivered women to 0.6 during the second half of pregnancy. At 6 p.m., unbound cortisol was significantly higher in outpatients during the 2d half of pregnancy than in non pregnant controls (0.017 vs 0.009 10(-6) mol/l). In hospitalized resting patients, unbound cortisol concentrations were the same in pregnant and non pregnant women except for a significantly higher level of unbound cortisol at midnight during the second half of pregnancy. Circadian variations of cortisol secretion persisted throughout pregnancy. It is concluded that in pregnancy (1) hypothalamic control of maternal adrenal secretion still prevails, (2) the increase in total cortisol is essentially due to that of CBG, (3) the unbound cortisol is normal or, at times, slightly increased and (4) variations in the PT/true CBG ratios may be explained by the presence of steroids (progesterone and 17-OH-progesterone) with a high affinity for CBG. The physiological significance of the mild hypercortisolism of pregnancy is discussed.