Forty-four patients with senile dementia of the Alzheimer type were randomly allocated into double-blind treatment with either aniracetam (RO 13-5057) 1 g or placebo daily for 3 months. Neurological examinations were made before and after treatment and psychometric tests were performed before and after 1 month's and after 3 month's treatment. Treatment was interrupted due to occurrence of confusion in four cases in the aniracetam group and in one case in the placebo group. During treatment, an improvement was seen in several cognitive tests, especially those associated with memory, but this improvement occurred in the placebo as well as in the aniracetam-treated group. In clinical evaluation no difference was seen in efficacy between the two treatment groups.