A survey was made of 274 non-institutionalized persons aged 70 and over living in Hobart. The prevalence of dementia and of depression was measured by interviewing subjects using a modified version of the Geriatric Mental State Schedule (GMS) (Copeland et al. 1976) and the Mini Mental State Examination (MMSE) (Folstein et al. 1975). Rates of morbidity were derived from different diagnostic procedures. These were: diagnoses made by a psychiatrist (A.S.H.) directly from the interview schedules and audiotapes, and rated as mild, moderate or severe; the criteria laid down in DSM-III, converted into algorithms describing 3 degrees of severity; and the algorithms for pervasive dementia and depression proposed by Gurland et al. (1983), and from these authors' rational scales. In addition, the relation between scales for dementia and for depression and the diagnosed categories was examined. Some problems in applying these methods to aged persons in the community are discussed. It is concluded that more detailed specification of criteria is desirable if the comparative epidemiology of dementia and depression in old age is to advance.