Lymphocytes from cattle were tested in a blastogenesis test with outer membrane proteins isolated from smooth strain 2308 and rough strain 45/20 of Brucella abortus. The titration assay developed for measuring blastogenesis to microbial antigens (Baldwin, Antczak and Winter, this issue, pp. 319-333) was used to assess the response to both group 2 (porins) (Douglas et al., 1984) and group 3 proteins (Verstreate et al., 1982). Blastogenesis was evaluated for distinguishing cattle infected with virulent B. abortus strain 2308 from unimmunized cattle, cattle vaccinated with attenuated strain 19, or inoculated with Escherichia coli 0116:H31, known to cause serological cross-reactions with B. abortus (Nielsen et al., 1980). Strain 45/20 porin was the most effective for this purpose and data analyses utilizing the titration assay were better than those relying on a single point assay. When compared with BASA, an antigen preparation used in other studies (Kaneene et al., 1978a), responses to porin provided a more specific index of infection with B. abortus. Reactions to 45/20 porin occurred, however, in some heifers vaccinated as adults with strain 19 or inoculated with E. coli 0116:H31. Furthermore, nonpregnant heifers had negligible or only transient blastogenesis responses to the porin during the first 14 weeks after infection even though they developed strong 0 antibody responses. We do not recommend the blastogenesis test in its present form as a useful adjunct to serological tests, and could allow measurement of cell mediated immune responses relevant to protective immunity.