Mycobacterium fortuitum and at least 1 unidentified species of soil mycobacteria were isolated from lymph nodes from 4 of 5 African buffalo (Syncerus caffer) that had been culled because of positive test results using the Bovigam assay. The buffalo were part of a group of 16 free-ranging buffalo captured in the far north of the Kruger National Park (South Africa) assumed to be free of bovine tuberculosis. No Mycobacterium bovis was isolated. To investigate the possible cause of the apparent false-positive diagnosis, the Mycobacterium isolates were inoculated into 4 experimental cattle and their immune responses monitored over a 13-week period, using the gamma interferon assay. The immune reactivity was predominantly directed toward avian tuberculin purified protein derivative (PPD) and lasted for approximately 8 weeks. During that period 3 of 4 cattle yielded positive test results on 1 or 2 occasions. The immune responsiveness was boosted when the inoculations were repeated after 15 weeks, which led to 2 subsequent positive reactions in the experimental animal that did not react previously. Including an additional stimulatory antigen, sensitin prepared from M. fortuitum in the gamma interferon assay, showed that it was able to elicit a detectable gamma interferon response in all 4 experimentally inoculated cattle when applied in parallel with bovine and avian tuberculin PPD for the stimulation of blood samples. The implications of occasional cross-reactive responses in natural cases of infection with environmental mycobacteria in the diagnosis of bovine tuberculosis in African buffalo and cattle in South Africa are discussed.