Biochemical confirmation of the identity of Burkholderia pseudomallei in Singapore previously relied on the API 20NE panel of tests. After introducing an alternative proprietary biochemical panel, the Microbact 24E (MedVet, Adelaide, Australia), we noted that the API panel identified some presumptive B. pseudomallei isolates as other species. We therefore compared the performance of the API 20NE against the Microbact 24E with 50 distinct clinical isolates of B. pseudomallei, after 24 hours and after five days incubation of primary cultures. The API panel correctly identified 40 isolates. Four results were unacceptable or uninterpretable. Six isolates were misidentified as other species; the commonest being Chromobacterium violaceum. One of these was again identified as C. violaceum by the repeat API panel. Fourteen isolates, including the six misidentified isolates and four isolate pairs from separate sources in four separate patients, were typed using PCR amplification of repetitive extragenic palindromic sequences (REPS). The isolates identified as C. violaceum appeared to have identical REPS patterns, suggesting that some of the errant API results may be due to a single locally prevalent strain of B. pseudomallei. A previous suggestion that C. violaceum may produce a melioidosis-like illness may therefore be due to laboratory misidentification of B. pseudomallei with the API 20NE biochemical test panel.