A quantitative carbon source growth assay, comprising ten carbon sources, was used to compare acinetobacter strains from three hospitals. The strains had been obtained during episodes of increased prevalence of isolations and were, for each hospital, assumed to be epidemiologically related. This assumption was supported by the electrophoretic protein profiles of the strains. Univariate analysis of growth data showed significant differences between strains from the three hospitals. Moreover, cluster analysis revealed that the major pattern in the data was related to the epidemiological origin of the strains. Exceptions to the epidemic-related pattern were observed. Thus, apart from epidemiological factors, other factors might contribute to carbon source growth profiles of the strains. It is concluded that the carbon growth assay may be useful to distinguish roughly between acinetobacter strains from different sites of origin. Further studies are required to analyse additional factors which influence carbon source growth of strains.