A prospective cohort study was undertaken in two commercial California dairies. The treatment group, 246 cows, received three doses of a whole cell bacterin of J5 Escherichia coli (mutant of E. coli O111:B4) plus Freund's incomplete adjuvant vaccine (two in the dry period and one after calving) while 240 unvaccinated cows served as controls. Thirty-five cases of clinical coliform mastitis were diagnosed, six in vaccinated cows and 29 in unvaccinated cows. Bacteria isolated from the clinical cases included 15 E. coli five Klebsiella pneumoniae, three K. oxytoca, three K. ozaenae, five Enterobacter aerogenes, three Serratia marcescens and one Serratia spp. Four control cows were culled, three of them because of chronic coliform mastitis and one because of postcoliform infection agalactia. Incidence rate of clinical gram-negative mastitis was 2.57% in vaccinated cows and 12.77% in unvaccinated cows. The estimated risk ratio, the measure of risk of having clinical gram-negative mastitis for vaccinated cows to unvaccinated cows, was 0.20 (p less than 0.005), indicating a strong relationship between vaccination and lack of clinical gram-negative mastitis. The results of this trial indicate that the administration of the E. coli J5 vaccine is protective against natural challenge to gram-negative bacteria, and reduces the incidence of clinical gram-negative mastitis in dairy cows during the first three months of lactation.