Normalised resistance interpretation was applied to measure the effects of incubation temperature and times on the precision of the data generated by MIC and disc diffusion antimicrobial susceptibility tests performed at 35 degrees C for 16 h, 28 degrees C for 24 h, at 22 degrees C for 24 h, at 22 degrees C for 48 h and < 19 degrees C for >96 h. Analysis of 151 MIC data sets and 141 disc diffusion data sets demonstrated the precision of MIC data sets was not affected by the incubation conditions but that the precision of the disc data decreased significantly as the incubation temperature decreased and the time increased. For MIC data sets the standard deviation of the normalised distribution of the log(2) transformed wild-type observations, which were independent of the numerical values of those sets, were applied as measures of the relative data set precision. The mean values of these standard deviations for 59 data sets generated at 35 degrees C, 27 data sets generated at 28 degrees C, 28 data sets generated at 22 degrees C and 37 data sets generated at 18-19 degrees C were 0.76 +/- 0.22 log(2) mu gml(-1), 0.76 +/- 0.26 log(2) mu gml(-1), 0.77 +/- 0.20 log(2) mu gml(-1) and 0.76 +/- 0.19 log(2) mu gml(-1) respectively. For the disc diffusion data sets the relative precision was quantified by calculating the standard deviation of the normalised distribution of the wild-type observations. The mean values of the standard deviation for 40 data sets generated at 35 degrees C, 50 data sets generated at 28 degrees C, 25 data sets generated at 22 degrees C and 26 data sets generated at <= 18 degrees C were 2.15 +/- 0.61 mm, 2.65 +/- 0.74 mm, 4.56 +/- 1.09 mm and 6.70 +/- 1.70 mm respectively. Because the incubation times increased as the temperatures decreased it was not possible in this work to determine whether changes in either or both were responsible for the reduction in precision observed in the disc zone data.The low precision of disc diffusion data obtained after prolonged incubation at temperatures below 22 degrees C suggests, that for bacteria that require these conditions, preference should be given to MIC susceptibility testing methods.