Surveillance for Johne's disease frequently relies upon culture of samples (tissue, fecal and environmental) for isolation of M. paratuberculosis (M. ptb.). This mycobacterial species is one of several genetically similar species belonging to the M. avium complex. PCR assays have been developed to discriminate among them. Accurate identification of members of this complex remains a difficult task however, especially if more than one mycobacterial species is isolated from a sample. Our aim was to evaluate the performance of three PCR assays that use different target sequences (2 university-developed: "A" and "B", and one commercially available: "C") to identify M. ptb. when an acid-fast organism is isolated from culture. The three PCR assays were evaluated with isolates obtained from radiometric culture of 261 fecal or tissue samples chronologically submitted for Johne's disease surveillance. Since both genetic target and growth pattern evidence was found of mycobacterial co-isolates from 74 cultures, these were analyzed separately. For the 187 cultures considered to contain only one mycobacterial isolate, PCRs A and B agreed 97% of the time with no statistically significant difference between their results. A comparison of PCR C with assays A and B yielded statistical significant differences with 67% and 68% agreement, respectively. When comparing each of the PCR assays (A, B, and C) for the 74 cultures believed to contain mycobacterial co-isolates, agreement was poor. (A:B - 31%, A:C - 78%, B:C - 53%.) Based on this study, isolation of more than one mycobacterial species is not rare, and their presence can interfere with the ability of the PCR assay to accurately identify M. ptb. The interpretation of results from molecular methods for diagnosis of M. ptb. from clinical samples should be done carefully. The use of multiple genetic targets is recommended as an effective practice to identify the presence of more than one mycobacterial species.