“…Several reports about the usefulness of the Li ratio as an index of prediction suggested that the Li ratio, which might or might not be related to the responder/nonresponder dichotomy, could be a marker for the biological heterogeneity that may underlie clinical subtypes of the disease [Litkens et al, 1973[Litkens et al, , 1976Mendlewicz et al, 1973;Morabito et al, 1982]. Hypotheses about different distributions of the Li ratio according to polarity of dis ease [Ribakowski et al, 1978;Knorring et al, 1976;Elizur et al, 1982;Albrecht and Miilier-Oerlinghausen, 1976;Andrews and Chiu, 1978] and to sex of probands [Litkens et al, 1973;Greil et al, 1979;Deniker et al.…”