The effect of an autogenous vaccine against Staphylococcus aureus on S. aureus prevalence and mastitis, as well as on somatic cell count (SCC), was studied in a dairy herd with a high prevalence of S. aureus. The vaccination group (n = 35; 22 cows and 13 heifers) and the control group (n = 36; 23 cows and 13 heifers) received the vaccine or a placebo, respectively, according to the following protocol: all animals: basic immunization (twice, 3 weeks apart); cows: booster dose at the time of drying off, 5 and 2 weeks before calculated calving date; heifers: booster dose 2 and 5 weeks before calculated calving date. The vaccine or the placebo was administered subcutaneously in the area of the supramammary lymph nodes. Quarter milk samples were collected monthly and subjected to SCC and bacteriological evaluation. At this time, the animals were also checked for signs of clinical mastitis. Non-clinical S. aureus mastitis diagnoses were based on udder quarter SCC and a positive S. aureus culture. In order to compare the SCC in individual whole milk samples, records from the monthly milk quality testing were evaluated. Cow and udder quarter prevalence of S. aureus intramammary infections calculated for the experimental animals and quarters, respectively, did not differ between groups. However, during the lactation period following the boostcr dose, the prevalence of S. aureus increased in both groups (P < 0.05). The cumulative incidence of various mastitis diagnoses (clinical, subclinical, latent infection) due to S. aureus on an animal basis did not differ between groups. On an udder quarter basis, the cumulative incidence of subclinical mastitis was higher in vaccinated animals than in control animals (33.8 versus 26.0%; P < 0.05). This was mainly due to a higher cumulative incidence of subclinical mastitis in vaccinated than control heifers. The SCC in composite milk samples did not differ between groups, but increased as lactation progressed. The herd prevalence of S. aureus differed considerably throughout the study period, but declined consistently to below 10% at the end of the study period. Recent herd checks revealed a prevalence of S aureus infections of < 5%. It is concluded that the autogenous bacterin tested in this study did not have the desired effect on the prevalence of S. aureus infections and mastitis or SCC. The decline in S. aureus prevalence was very probably due to other factors than specific immunization against S. aureus.