Twenty-five years ago the research group of the Canadian psychiatrist de Montigny reported treating antidepressant-refractory depressive patients successfully by adding lithium to their antidepressant. The report, published in 1981 as an open-label uncontrolled observation of only eight patients, falls short of today's methodological standards, but the treatment method, subsequently known as lithium augmentation, nonetheless was to change profoundly the pharmacological strategies for depressive disorders. The story of its development is remarkable, starting with a strictly theoretical idea conceived by Montigny and his colleagues after animal experiments in the 1970s had revealed that pretreatment with an antidepressant over several weeks led to sensitization of central nervous serotonin receptors. The team postulated that the proserotonergic characteristics of lithium, which had been systematically used as a psychotropic drug since 1949, could thus be used specifically to stimulate these receptors. Lithium augmentation demonstrated its effectiveness in the 1980s and 1990s, first in open-label and later in randomized and placebo-controlled studies. In the late 1990s studies aimed at optimizing its clinical application indicated that lithium augmentation must be administered for at least 2 weeks, with lithium serum levels within the range established for prophylactic treatment and assuming patient response, and that the combination of lithium and antidepressant must be continued as a maintenance therapy for 6 to 12 months. Research has yet to clarify how lithium augmentation actually works. Current results show that in addition to the idea postulated by Montigny, lithium could also have an activating effect on the cortisol axis. Thanks to the sound body of evidence which has accrued in the meantime, lithium augmentation is recommended in most guidelines and treatment algorithms as a main strategy for patients who do not respond to antidepressant monotherapy.