The immune response of cattle to nonliving vaccines derived from Brucella abortus rough strain 45/20 was studied. Vaccines contained trehalose dimycolate and a derivative of muramyl dipeptide. N-acetylmuramyl-L-alpha-aminobutyryl-D-isoglutamine. A factorial experiment was designed to test the effects of type of antigen, quantity of antigen, and quantity of mineral oil on the immune response to porin. Muramyl dipeptide was kept constant at 5 mg per dose, and 1 part of trehalose dimycolate was incorporated for two parts of dry matter. Over a 10-week period, blastogenesis responses to porin were largest in cattle immunized with outer membranes; the highest antibody titers to the porin-lipopolysaccharide complex were achieved by immunization with detergent-extracted outer membrane proteins. There was no advantage in the use of 25, rather than 5, mg of any of the antigens, but antibody responses were improved by increasing the quantity of oil from 0.6 to 1.8 ml per dose. In other animals, blastogenesis and antibody responses were sustained at high levels longer than 3 months after two vaccinations with outer membrane proteins. Intradermal injection of porin evoked inflammatory reactions histologically consistent with delayed-type hypersensitivity. Cross-reactions in cases of delayed-type hypersensitivity occurred with porin derived from a smooth strain of B. abortus but were less extensive than in the blastogenesis test. The magnitude of the delayed-type hypersensitivity and blastogenesis responses induced by vaccination exceeded those observed after natural or experimental infections. No ill effects were observed after vaccination. These findings provide a basis for the use of trehalose dimycolate and muramyl dipeptide adjuvants in evaluating nonviable vaccines for bovine brucellosis.