The study assesses the reproducibility, typability and discriminatory power of several typing methods for Pseudomonas aeruginosa isolated from cystic fibrosis patients. 178 polyagglutinable Pseudomonas aeruginosa isolates from cystic fibrosis patients were serotyped using polyclonal sera and monoclonal antibodies, phage typed, pyocin typed and reverse phage typed. 31 of these polyagglutinable isolates, six monoagglutinable isolates and three nontypable isolates were also typed by means of hybridization using a DNA probe. In a comparison of the methods used, on polyagglutinable isolates only, typability was 0% with polyclonal sera, 90% with monoclonal sera, 94% with phage typing, 85% with pyocin typing, 36% with reverse phage typing and 100% with DNA-prope typing. Using monoclonal antibodies, the reproducibility was 75%, while that of phage typing was 88%, pyocin typing 53% and reverse phage typing 62%. Typing with the DNA probe was not repeated. using polyclonal sera, repeated typing showed that 94% of the isolates were polyagglutinable. Using phage typing, 40% of the isolates belonged to phage type 31, while 60% were distributed amongst 32 phage types. Using monoclonal antibodies, 71% of the isolates belonged to 0-group 3, and these isolates showed 16 different phage types. Subdivision of the phage types was further achieved by both pyocin typing and reverse phage typing. The DNA probe typing made it possible in some cases to discriminate between isolates which were otherwise found identical with the conventional typing methods, while in other cases typing with the DNA probe recorded as identical isolates which conventional methods had typed as being different. These differences may be due to a high mutation rate caused by the selection pressure of antibiotics, and by the host immune response. According to our results, investigations of reproducibility and typability of old and new typing methods are essential when they are used in clinical situations. The low reproducibility of some of the typing methods in the present study affects the reliability of epidemiological investigations in cystic fibrosis patients. Usage of only one method may not be sufficient in cases of polyagglutinable strains from cystic fibrosis patients.