This report expands on previously presented evidence for a trait-dependent deficiency of erythrocyte sodium pump activity in bipolar affective disorder. Several parameters of erythrocyte NaK-ATPase activity in different affective states and the effects of lithium therapy were examined. In lithium-free bipolar affective disorder patients, the mean percent differences from the individual sex- and age-matched controls for erythrocyte sodium pump activity were: manic + hypomanic group, -21.5%, n = 16, p less than 0.02; depressed group, -12.4%, n = 14, p less than 0.02; euthymic group, -6.9%, n = 18, p less than 0.10. Ouabain-sensitive potassium ion uptake was less than the controls in the manic group (-25.4%, n = 3, p less than 0.02), and in the combined affectively ill group, manic + depressed (-23.4%, n = 7, p less than 0.02). Cell ouabain binding was less than the controls in the manic group (-19.0%, n = 6, p less than 0.05). NaK-ATPase activity in washed erythrocyte membranes (ghosts) was significantly lower than the controls in the manic group (-14.5%, n = 4, p less than 0.02), and the mean value for the whole group (manic + depressed + euthymic) was lower than the controls (-11.8%, n = 18, p less than 0.05). Values for ouabain binding in ghosts from the bipolar subjects were not significantly different from the matched controls.(ABSTRACT TRUNCATED AT 250 WORDS)