A ligase chain reaction (LCR) DNA amplification method for the molecular diagnosis of Mycobacterium tuberculosis complex (Abbott LCx MTB) was evaluated in comparison with solid (Lowenstein Jensen), liquid (7H12, Bactec 460 system) phase culture and microscopic examination (ME) on 86 tissue samples collected from 86 intradermal tuberculin positive cattle and one pool from 4 guinea pigs experimentally infected with M. bovis. Overall, 48 samples (58.81%) were culturally positive for mycobacteria, and on the basis of biochemical characters, all the isolates were identified as M. bovis. Sensitivity was 83.92% for LCx, 53.57% for LJ, 85.71% for Bactec and 41.07% for ME. In 3 out of 25 "no visible lesion" tissue samples, M. bovis was detected only by LCx and Bactec but not by LJ and ME. The concordance in the determination of positives and negatives among the methods observed in pairs was calculated according to Cohen's K concordance coefficient and showed 81.1% of concordance of LCx vs Bactec, 68.8% LCx vs LJ, 72.2% LCx vs ME, 80.0% Bactec vs LJ, 66.7% Bactec vs ME, 85.5% LJ vs ME. Despite a certain variability in concordance rates, both Cohen's K concordance coefficients or standardized (Zk) values were statistically significant. Both LCx and Bactec appear not alternative but subsidiary to the other methods traditionally applied for direct diagnosis of bovine tuberculosis on tissue samples from cattle reacting to intradermal tuberculin test.