Usefulness of lithium in the prophylaxis of bipolar disorders has been challenged for five major reasons. The authors review the empirical basis of these criticisms and come to the following conclusions. 1. Lithium efficacy is high and beyond reasonable doubt in classic manic-depressive illness. Bipolar patients presenting atypical features show a much poorer response rate to lithium. 2. There is no empirical evidence for a loss of lithium efficacy over time. 3. There is little evidence for discontinuation-induced refractoriness to lithium. 4. Lithium withdrawal phenomena are well established but seem to be rather specific to certain subgroups. Withdrawal phenomena seem to be common in atypical bipolar disorder but rare in fully stabilized classic manic-depressive illness. 5. Other factors limiting lithium efficacy in clinical practice (e. g., non-compliance) are not specific to lithium. In conclusion, prophylactic lithium does have major drawbacks and there is a clear need for more efficacious alternatives in non-classic bipolars. Compared to existing alternatives, lithium currently is to be considered the golden standard. This status might, however, be challenged by major alternative mood-stabilizers that are presently under clinical investigation.