The use of case history studies to differentiate potentially infected from potentially noninfected herds with reactors to Brucella abortus antigens.

Data on farm characteristics and the results of the first herd test for brucellosis were collected for 74 reactor and 74 negative herds in Wellington County, Ontario. Each reactor herd was classified as either probably infected or probably not infected using the occurrence of abortions prior to the first herd test or during the testing period, the total number of cattle removed and/or the spread of reactors within the herd as criteria of infection. Statistical techniques were used to select variables which were good "discriminators" between probably infected and noninfected herds. In general, reactor herds were primarily dairy herds and were somewhat larger than negative herds. The presence of only single suspicious reactors on the first test appeared to be a good predictor of lack of infection with Brucella abortus. Among the 37 farms in this category the single reactor was removed from only eight farms and no evidence o fthe spread of infection was observed. The presence of one or more positive reactors on the first herd test appeared to be a good predictor of the presence of infection. Of the 24 farms in this category, evidence of the spread of infection was present in ten farms and seven of these ten farms were eventually depopulated. The brucella milk ring test appeared to be the most effective means of identifying infected herds under the conditions present in Wellington County.