Two groups of 4 cows were vaccinated subcutaneously with live Streptococcus uberis strain 0140J or a surface extract derived from the same strain, at 14 days prior to the cessation of lactation (drying off) and at calving. Both groups also received an intramammary administration of the surface extract 7 days after drying off. A third group of unvaccinated animals acted as controls. Following intramammary challenge of two quarters per cow with the vaccine strain, all quarters on control cows and those vaccinated only with surface extract developed clinical mastitis. However, only 12.5% of challenged quarters on cows which were vaccinated with live bacteria developed clinical mastitis. In addition, the numbers of bacteria in the milk following challenge were 10(5) times higher from the control and extract vaccinated cows than those which received live vaccine. Serum levels of S. uberis specific IgG2 were elevated in the animals vaccinated with the live organism when compared to that of either extract-vaccinates or controls, whilst S. uberis specific levels of IgG1 and IgM were similar in all groups throughout the experiment. Specific antibody levels in milk were unaffected by vaccination. Despite increased levels of IgG2, no increase in opsonic activity was detected in any serum or milk samples. Peripheral blood lymphocytes from animals vaccinated with live organisms showed a considerable increase in proliferative response to S. uberis antigen in vitro when compared with lymphocytes from control and extract-vaccinated animals. These results suggest that neutrophils and specific opsonising antibody may not form the major defence against infection with S. uberis.