The Hamilton Rating Scale for Depression (HAM-D) and DSM-III-R criteria were simultaneously employed to assess the prevalence of depression in 26 outpatients with dementia of the Alzheimer type and 26 age-matched normal control subjects. Both assessment methods evidenced a higher frequency of depression during the severe stages of Alzheimer's disease. Among the Alzheimer patients, the prevalence rate of depression produced by the HAM-D (38%) was higher than the rate produced by DSM-III-R criteria (23%). Such a difference was due to the weight given by the HAM-D to the vegetative symptoms reported by the Alzheimer patients with more severe dementia. In a subgroup of 14 Alzheimer patients who underwent computed tomography, the volumetric measurement of CSF spaces did not reveal any difference between the depressed and nondepressed patients. On the basis of these results, the clinical problems related to the assessment of depression in Alzheimer's disease are discussed.