IN the three decades since Cade's discovery,1 lithium has moved from a position of therapeutic promise to an extensively studied treatment of choice in bipolar, and certain subtypes of unipolar, affective disorders. Marketing firms and pharmaceutical companies estimate that more than 120,000 Americans currently are taking lithium (G. Graham, MD, Smith Kline & French Laboratories, personal communication), figures consistent with Jenner and Eastwood's2 estimate that one person in 2,000 in the United Kingdom is now receiving lithium. However, despite the growing number of patients receiving lithium, its well-established usefulness in the treatment and prophylaxis of affective disorders, and the common clinical impression of frequent problems with compliance, it is not known how many patients actually stop taking lithium against medical advice, for what reasons and for how long, at what point in their therapeutic regime or mood cycles, and whether or not there are sex and age