For many year two Danish psychiatric hospitals having used different lithium treatment regimens. In one, slow-release tablets were given in two daily doses and, in the other conventional tablets were given in a single daily dose. In both hospitals many patients developed polyuria. Multiple regression analyses with sex, age, treatment duration, serum lithium concentration, and treatment regimen as predictor variables showed that the two treatment regimens did not affect the glomerular filtration rate or the proximal reabsorption differently, but that distal water reabsorption was significantly less affected and polyuria less pronounced in the patients given conventional tablets once daily than in those give slow-release tablets twice daily. The authors are divided among themselves as regards the implications of these findings.