The exact prevalence of depression in old age remains uncertain, in large part because of the variations in interview instruments and diagnostic criteria selected as suitable to detect the myriad presentations of 'geriatric depression'. The Epidemiologic Catchment Area (ECA) Study, for example, found one-year prevalence rates for both DSM-I11 major depressive disorders (MDD) and dysthymia in 5000 persons aged 2 65 yrs to be only %-% of rates for middle-aged adults (Weissman and Myers, 1978). The more general finding, however, is a slightly increased rate of MDD in the elderly with a substantial increase in reports of minor depressive symptoms. Blazer and Williams (1980) found 3.7% of 1000 persons 2 6 5 yrs to have MDD with a further 11% reporting minor symptoms of 'senile dysphoria'. Physical illness is highly associated with depressive disorder (Gurland et al., 1983) and increased rates of MDD in medically ill patients are recognized. The mechanisms underlying this relationship are uncertain-one possibility is that the continuing stress of disability may produce hypothalamo-pituitary-adrenal axis dysregulation in physically ill individuals. Murphy's studies (eg Murphy, 1986) have demonstrated the importance of pre-existing vulnerability factors including physical health problems, social deprivation and lifelong personality traits, all of which contribute to an increased likelihood of a depressive reaction to the negative life events common to old age. The possibility arises that an additional vulnerability may be produced by the ageing of brain mechanisms or structures, reducing the capacity of the individual to accommodate or adapt to life stresses. Although such a mechanism might operate to some degree in most if not all elderly people developing depression, it would be expected to be