More accurate tests are required to test cattle which have reacted positively in the tuberculin skin test. For This purpose, a range of mycobacterial antigens, MPB59, MPB64, MPB70, MPB83, ESAT-6 and CFP10, were used either as recombinant proteins or as synthetic peptides in the whole blood interferon-γ (IFN-γ) test. Groups of uninfected cattle with typical 'non-specificity' problems were targeted, in particular animals with skin tuberculosis, animals vaccinated against Johne's disease and animals that were positive in the standard purified protein derivative (PPD)-based IFN-γ test. The two study groups consisted of 74 Mycobacterium bovis-culture positive animals and 72 uninfected animals, all of which tested positive in the caudal fold tuberculin skin test eight to 28 days before the blood test. The use of combinations of ESAT-6 and CFP10 antigens, either as recombinant proteins or peptides, detected similar percentages of M bovis-infected animals as the PPD-based IFN-γ test, but produced significantly fewer false positive reactions. The PPD-based IFN-γ test was very effective in differentiating animals vaccinated against Johne's disease that were skin-test positive from those with bovine tuberculosis, and the use of PPD or specific mycobacterial antigens minimised the number of false positive reactions in animals with skin tuberculosis.