Chronic affective disorder is now well recognized as a cause of serious and prolonged psychiatric morbidity. As the paper of Malizia and Bridges points out, a high proportion of these cases could be the result of inadequate or inappropriate treatment. The classical history of such patients is often, however, that of a biological illness, perhaps following pregnancy which, in the first instance, had responded to either drugs or ECT. As the illness progresses, the illness episodes fail to respond to these treatments, or response is only partial with a continued low level of symptomatology. In essence, treatment resistance may develop with time. Their paper certainly gives a case for a more energetic approach to pharmacotherapy at the point in time of the patient's clinical history, where chronicity is established. Although, there are occasions when innovative pharmacotherapy produces dramatic, to the point of 'miraculous', cures, more frequently such treatments result in a diminution rather than extinction of depressive symptoms. It is perhaps appropriate to point out, in the setting of a belief that drugs are the complete answer to the problem, that these illnesses occur in human beings and consequently other factors which relate specifically to man have to be taken into account.A clue as to which areas one should focus on is given by studies which look at populations of patients who recover from depressive illness within 2 years compared with those who do not (Scott, Barker and Eccleston, 1988). The result of these studies shows quite clearly that chronicity has a multi-factorial origin and that certain predictions can be made about factors which predispose to chronicity. These can be divided, in very broad terms, into biological (physiological) and cognitive and psychodynamic (psychological). Even this division is somewhat arbitrary (Goodman, 1991). Akiskal et al. (1981) and Scott, Barker and Eccleston (1988) found that, on the biological side, chronic major depressives differed from depressives who recovered in that they were more likely to have unipolar disorder, had more previous episodes of illness, a higher family loading for affective disorder, and an increased incidence of thyroid dysfunction. Of the more specific factors that could be described as psychological were an increased incidence of independent undesirable life events, both before and after the onset of the index episode. Hirschfeld et al. (1986) confirmed the