The objective was to compare the efficacy of two experimental Staphylococcus aureus mastitis bacterins and a currently marketed five-isolate-based Staph. aureus bacterin (Lysigin™, Boehringer Ingelheim Vetmedica, Inc.) with unvaccinated controls. Forty-seven Holstein-Friesian heifers were randomly assigned to one of four groups such that Group 1 (n=11) received a three-isolate experimental bacterin, Group 2 (n=11) received a five-isolate experimental bacterin, Group 3 (n=14) received Lysigin, and Group 4 (n=11) served as unvaccinated controls. Vaccinations were administered twice 28 d apart in late gestation. All groups were challenged with a heterologous strain of Staph. aureus (ATCC 29740) on days 6, 7, and 8 of lactation. Mastitis score, somatic cell count (SCC), milk culture yield, and total daily milk yield data were collected before and after challenge. All 47 cattle developed a Staph. aureus IMI post-challenge with three animals in Group 1 and one animal in Group 3 clearing their Staph. aureus IMI by the end of the study. However, there was no evidence of a difference between vaccinates and control with regard to Staph. aureus clearance rates post-challenge (P[ges ]0·214). Cattle vaccinated with Lysigin had a lower mean duration of clinical mastitis and lower total mastitis score post-challenge than controls (P=0·045 and P=0·046, respectively). Overall, there was no evidence that any of the vaccinated groups had a lower mean SCC than control (P[ges ]0·148) for the tested study days. Likewise there was no evidence that vaccinates had greater milk yield than controls post-challenge (P=0·617). Hence, there was no evidence that the vaccines reliably prevented Staph. aureus IMI, but Lysigin showed benefit in reducing the clinical severity and duration of clinical disease post-challenge. Neither of the experimental bacterins appeared to perform better than Lysigin.