This study examined naming abilities in three groups of older adults with: I) major depression alone, II) major depression with reversible cognitive dysfunction, and III) dementia with depression. Groups I and II differed significantly from dementia patients in total correct responses to a visual-confrontation naming task (Boston Naming Test). Qualitative aspects of naming, specifically types of errors characterizing each patient group, were examined, but no statistically significant differences among groups were observed. The results support the contention that the presence of dysnomia may be useful in discriminating cognitive abnormalities secondary to dementia from cognitive dysfunction associated with depression.