Abstract
A total of 190 strains of Y. enterocolitica from environmental sources, humans and other animals were scored on biochemical, nutritional and resistance tests. These were run in duplicate at 30 and 35.degree. C, and the results were analyzed by numerical taxonomy. The computer-generated phenograms derived from simple matching coefficients and unweighted average linkage analysis delineated 7 phenons from the 30.degree. C data and 5 phenons from the 35.degree. C data at a similarity value of 85%. A centrostrain was derived from each cluster in both matrixes and these centrostrains were used as reference strains for subsequent DNA hybridizations. The relative binding ratios of the centrostrains to 1 another and to other organisms of their own phenons, were determined. The genetic relationships as denoted by relative binding ratios were compared with the phenetic relationships developed by numerical taxonomy. Taxonomic relationships generated by numerical taxonomy do not necessarily correlate highly with those derived by DNA hybridizations in either a linear or a numerical sense (correlation coefficients, 0.48 and 0.40; mean correlation ratio, 0.60). Numerical taxonomies may show considerable fluctuations in phenomic composition, which are dependent on the temperature of incubation. DNA hybridizations among the various strains of this study divided the species Y. enterocolitica into at least 4 DNA relatedness groups. The level of intergroup homology was high enough (relative binding ratio, > 20%) to justify the inclusion of all groups as species in the genus Yersinia.