Fifty-four cattle were sensitised to Brucella antigens either by vaccination with Brucella abortus strain 19 (S19) or B. abortus 45/20 (S45/20) and 24 of these were challenged 12 weeks after mating with virulent B. abortus strain 544 (S544). A further 12 cattle which were not vaccinated were exposed to S544. After 40 weeks, all these cattle (66), together with 5 cattle which were not sensitised by vaccination or challenge were subsequently inoculated with one dose of S45/20 and the anamnestic response was measured by the complement fixation test. Ten to 15 weeks later the cattle were slaughtered and tissues cultured. Of the 52 (2 died) vaccinated cattle, 35 gave a positive anamnestic response and 20 of these were not challenged. Of the 17 unvaccinated cattle, one gave a positive response and this animal had been exposed to S544 prior to the inoculation with S45/20. The results indicated that the method had a level of sensitivity of 75% and specificity of 100% in serologically negative cattle that had been exposed previously to Brucella antigens. An evaluation of the method for detecting serologically negative, but infected cattle was not possible as the number of cattle suitable for examination in this study was too low.