This study evaluates the performance in isolating methicillin-resistant Staphylococcus aureus (MRSA) of three media: the reduced salt formulation of mannitol salt agar plus oxacillin (MMSAO); CHROMagar Staph aureus plus ciprofloxacin (CHRAC); and Halifax MRSA medium (HMO), against the previously recommended mannitol salt agar (7% salt) plus oxacillin (OMSAO) and Baird-Parker medium plus ciprofloxacin (BPC). MRSA screening swabs were plated out onto the five selective media and the plates examined at 24 and 48 h. Suspected colonies were confirmed as MRSA by detection of heat-labile DNase, coagulase and/or protein A, and by confirming resistance to methicillin. Of 719 specimens examined, 191 grew MRSA on at least one medium. The relative sensitivities of the five media at 48 h were as follows: BPC, 94%; CHRAC, 70%; OMSAO, 61%; HMO, 56%; and MMSAO, 46%. In addition, BPC gave the least number of unnecessary investigations for non-MRSA isolates. The current advantage of BPC when performing direct culture for MRSA was confirmed. The other ciprofloxacin-containing medium also produced reasonable results. Of the two mannitol salt agar media, the formulation with 7% salt gave better results. HMO proved unreliable at isolating MRSA.