It has been suggested that the 30-item version of the Boston Naming Test (BNT), in which either the odd or even items from the standard 60-item test are given, is the most psychometrically sound short form. However, no normative data are available for this version. We administered the Odd/Even BNT to 30 community-dwelling elderly individuals (age M = 72.93, range 61-84; education M = 13.73) in order to collect normative data. Odd and even forms were equivalent. The combined mean total correct score was 27.13 (SD = 2.06), a score consistent with that derived by retrospective extraction in the original odd/even test construction study. Each form discriminated normals from age- and education-matched patients with probable Alzheimer's disease, suggesting criterion-related validity.