The annual incidence of newly detected P. aeruginosa from the respiratory tract of patients with cystic fibrosis is 15-17%, regardless of age. To understand risk factors for acquisition, a sensitive method which will detect variation among strains that have undergone clonal expansion and selection under different environmental circumstances for an extended time period is necessary. We compared two methods of strain identification: 1) polymerase chain reaction (PCR) based ribotyping and endonuclease subtyping (RTEST) with 2) arbitrary primer PCR based typing (APPCR) for their potential usefulness in such epidemiologic studies. We analyzed strains from widely separated locations in North America. We found that RTEST detected no length polymorphism differences among P. aeruginosa colonies from these distinct areas while APPCR did. By APPCR we detected no length polymorphism differences among DNA from 6 distinct subcolonies derived from single P. aeruginosa colonies. Sequential subculture showed that APPCR detected 0-1 length polymorphism differences among subcolonies derived from a single P. aeruginosa colony at 1, 10 and 40 generations. This provides evidence of the stability of these polymorphisms obtained by APPCR after clonal expansion of a given strain. P. aeruginosa strains recovered from the respiratory tracts of individual CF patients at intervals ranging from 6 mos to 4 years will be tested by APPCR. We conclude that APPCR may be useful in epidemiologic investigations of acquisitions of P. aeruginosa in patients with CF.