A new method has been devised to trace cross-infection by Pseudomonas aeruginosa. Unknown strains growing logarithmically in liquid media were treated with mitomycin C to induce the liberation of pyocin and phage. The lysates were then tested against 27 selected indicator strains, and the zones of clearing were differentiated as to killing by pyocin or lysis by phage. Twenty-four standard pyocin-phage lysates were then applied to each of the unknowns, and the sensitivity pattern was recorded. Thus, an "epidemiological fingerprint" consisting of 51 operational characteristics was established for each isolate. Organisms from the same source had identical or similar fingerprints, but organisms from different origins could easily be distinguished. Pyocin production, pyocin sensitivity, and phage production were found to be stable genetic characters; however, spontaneous mutations in phage sensitivity were frequently encountered. The epidemiological fingerprint has proven to be a sensitive tool in establishing the identity or dissimilarity of unknown strains. This method has been of great value in tracing the epidemiology of P. aeruginosa in the hospital environment. Each of the 157 P. aeruginosa strains tested has been typable by this method. In this decade, Pseudomonas aeruginosa is MATERIALS AND METHODS Pseudomonas strains. The Pseudomonas strains used in this study were from the following sources. P. aeruginosa 45, P. fluorescens 12 (biotype A), P. putida 90 (Biotype A), P. acidovorans 14, P. testosteroni 78, P. alcaligenes 142, P. pseudoalcaligenes 417, P. multivorans 382, P. stutzeri 221, and P. maltophilia 67, all described by Stanier, Palleroni, and Doudoroff (12), were obtained from R. Rippka, University of California, Berkeley, Calif. Thirteen pyocin-sensitive strains of Pseudomonas aeruginosa, Ti to T13, and five pyocin-producing strains, P1, P5, P9, PlO, and P16, were obtained from R.