SUMMARYThe relatedness between several nomen species of the PseudomonasXanthomonas group and some other organisms was numerically fixed through D N A homology and D N A base composition. For Pseudomonas the numerically analysed strains proposed by Lysenko as neotypes were used. The mean Yo (G+C) was in the range 60-67.5. 14C-DNA from either Fseudomonasfluorescens or P. putida was hybridized with D N A from 17 different species centres and the D N A homology was in the range 50-100 %.Genetic species differentiation in the genus Pseudomonas seems justified. In three border cases (P. iodinum, P. diminuta and P. atlantica) D N A homology was only 28-50%, so the inclusion of these organisms in the genus Pseudomonas is uncertain. The species centres P. pavonacea and F. rubescens are omitted from the genus Pseudomonas because of their very low D N A homology and aberrant D N A base composition. Twentyeight nomen species of Xanthomonas all form a narrow group in the range 63.5-69y0 (G+C). With two exceptions D N A homology with a median strain Xanthomonas pelargonii was always over 75 yo and frequently nearly complete. 14C-DNA from P. fluorescens hybridized with Xanthomomas-DNA to the same extent as with the pseudomonads proper. The Xanthomonas cluster overlapped perfectly with part of the Pseudomonas group. It is proposed therefore to gather all xanthomonads in a single genetic species P. campestris. This is such a dense cluster that the preservation of separate species names for the border cases seems undesirable. About one half to two-thirds of Pseudomonas-and Xanthomolzas-DNA is identical. The genera Rhixobium, Axotobacter and Axomonas appeared to be rather closely related to Pseudomonas since they shared some 40-50y0 of their DNA. The genus Serratia appeared to be more closely related to Pseudomonas than to Escherichia. The genera Gluconobacter, Acetobacter, Serratia and Escherichia shared some 20-30y0 D N A with Pseudomonas, but Bacillus-DNA was almost entirely different. From a comparison between D N A homology and taximetric similarity, it appeared that most pseudomonads would not contain unused genes. The advantages of a classification based on % (G+C) and D N A homology are obvious.