A simple method for the bacteriocine typing of Serratia marcescens without the use of induction was sought. The results of a mutual inhibition experiment with 89 unrelated cultures indicated that a bacteriocine-susceptibility method would give more discrimination between strains than would a bacteriocine-production method. A cross-streaking technique for bacteriocine-susceptibility typing without previous induction was developed, and its performance was compared with that of another susceptibility-typing method in which cell-free lysates of the producer strains were obtained by induction with mitomycin C.
Replicate typing of the same collection of cultures by both methods indicated that small variations in pattern were common and that larger variations occurred occasionally. Differences in pattern of less than two strong reactions in the mito-mycin-C induction method, and of less than three strong reactions in the cross-streaking method, should therefore not be taken as evidence that strains can be distinguished.
Sets of cultures of Ser. marcescens, 178 in total, from a number of supposed incidents of infection in hospitals, were used to evaluate the two bacteriocine-typing methods; all of the cultures were also O serogrouped. Comparison of the typing patterns of members of the same O serogroup from clear-cut incidents of infection confirmed that results of acceptable reliability could be obtained by either bacteriocine-typing method by the application of the appropriate 'difference' rule. When so interpreted, the cross-streaking method appeared to be slightly the more discriminatory.
The greatest discrimination between strains was obtained by the use of a 'hierarchical' typing system in which the strains were first O serogrouped, and the cross-streaking method of bacteriocine typing was then used to make subdivisions within O serogroups.