An increase in methicillin-resistant Staphylococcus aureus (MRSA) infections prompted a study of MRSA during a 21-month period in a 600-bed university hospital in southern Texas. MRSA cases were classified as community, nosocomial, or transfer cases. A case-control study of risk factors for community MRSA compared with community methicillin-susceptible S. aureus (MSSA) was performed. Pulsed field gel electrophoresis (PFGE) of whole cell DNA typing was used as a marker of strain identity for 31 consecutive isolates collected during the last 8 months of the study. During the 21 months there were 170 patients with MRSA infection or colonization, an incidence of 0.2 per 1,000 patient-days. Ninety-nine (58%) of 170 isolates were from community cases; the community to nosocomial case ratio was 2:1. No significant risk factors differentiated patients with community MRSA compared with community MSSA. Most community MRSA isolates studied (15 [68%] of 22) had distinct PFGE patterns, as did many nosocomial MRSA isolates (4 [44%] of 9). MRSA isolates were commonly present on admission to the hospital, and multiple MRSA strains were demonstrated among both community and hospital isolates.