The method of comparing premorbid versus current intellectual ability has become established clinical practice in the differential diagnosis of dementia versus depression. Recently, Schlosser & Ivison (1989) suggested that the comparison of premorbid ability versus current memory function may offer a more sensitive method of assessing early dementia. In the present study, a variety of within-subject discrepancy analyses comparing premorbid estimates with current measures of memory and intellectual functioning were compared across three groups: patients with dementia of the Alzheimer type, patients with major depression and healthy controls. The results revealed that, while mean group differences were easily demonstrated, the overlap between Alzheimer and depressed patients was large. It is concluded that none of the simple neuropsychological discrepancy analyses examined in the present study can be recommended for use in clinical practice for the differential diagnosis of dementia from major depression.