During final phases of eradication programmes, strains of Mycobacterium sp. not belonging to the tuberculosis complex increase their relative frequency and are responsible for positive skin test reactions. Moreover, the specificity of any indirect diagnostic test, such as the skin test, is never completely accurate, therefore even when tuberculosis infection is completely eradicated, a number of false positive reactions are to be expected. The aim of this paper is to evaluate the performances of traditional isolation/typing techniques, automatic isolation/typing techniques based on fluorimetric detection of bacterial growth (Bactec), skin tests and the -interferon test. Samples examined for the evaluation of test sensitivities originated from 154 infected animals belonging to 32 infected herds. Samples used as negative controls in the evaluation of test specificities originated from 86 animals of nine officially infection-free herds. The automatic isolation/typing technique based on fluorimetric detection of bacterial growth showed higher sensitivity than the traditional isolation typing technique. Moreover, it allowed a safer processing of bacterial cultures, decreasing the risk for laboratory workers. The observed performance of the gamma-interferon test was considered beneficial in that it increased the sensitivity of individual diagnosis within an infected herd, especially in 'problem herds', but its poor specificity did not improve detection of infected herds compared to the skin test.