Two interpretations of the Boston Naming Test (BNT) discontinuation rule of "six consecutive failures" were uncovered in an informal survey. The rigorous interpretation includes correct responses to phonemic cues in the count of failures, whereas the lenient interpretation does not. Using both methods to score the same protocols, BNT final scores were changed in 3% of 655 normal elderly persons and in 31% of 140 patients with Alzheimer s Disease (AD). Among normals, discrepant scores were most often found in persons age 80 years and older, and scores differed by up to 16 points. Among AD patients, discrepant scores did not preferentially occur in any age group, but scores differed by up to 10 points. The clinical impact of BNT scoring differences occasionally can be substantial and warrants attention.