An increased incidence of bacteraemia due to Streptococcus faecalis in a Critical Care and Trauma Centre (CCTC) during November 1985 prompted investigation. During the epidemic 21 blood cultures from five CCTC patients were positive for S. faecalis. A point prevalence culture survey revealed two more strains from wounds from two of these five patients. Fifteen strains from blood cultures and the two strains from wound sites were further characterized by conventional biotyping methods and susceptibility patterns, but they could not be differentiated from 40 unrelated control strains by these methods. The API-20S and the API-ZYM systems, however, were able to distinguish the outbreak isolates from the control strains. This investigation supports the hypothesis that S. faecalis is capable of causing cross-infection, and it may be necessary to characterize enterococci beyond routine tests to identify an outbreak due to this organism.