The Boston Naming Test (BNT) is frequently used to assess for word-retrieval difficulties, but its use outside North America is restricted by items such as beaver and pretzel. A total of 136 older Australians were tested on the BNT. Those items which frequently produced naming errors were identified and alternative items trialled. Age, years of schooling, visual acuity and forward and backward digit span assessments were correlated with BNT scores, and errors were analysed. The mean BNT score for Australian subjects was two to five items below the reported mean scores of North American subjects. Percentages correct for beaver and pretzel were greater than two standard deviations below the mean BNT score, and alternative items are proposed. Age, years of schooling, visual acuity and backward digit span were all found to be significantly correlated with BNT scores, and semantically related errors were the most frequent type of error followed by don't-know responses. This study suggests that examiners need to be cautious when interpreting BNT scores of Australian and possibly other non-American subjects. Cut-off scores need to take into account word-frequency differences between cultures, and examiners need to be aware of the effects of age, educational level, visual acuity and memory.