623One hundred and fifty pink-pigmented facultatively methylotrophic bacteria (PPFMs) of the 'Pseudomonas extorquens' type, 28 other facultative methylotrophs and 16 non-methylotrophic marker strains of the genera Pseudomonas, A lcaligenes, Mycoplana and Microcyclus were compared in a numerical phenetic study using 140 unit characters. Data were analysed using the simple matching (SsM) and Jaccard (S,) coefficients and single, complete and average linkage algorithms. Cluster composition was largely the same with each similarity coefficient and linkage method. Four major and seven minor clusters ,containing 187 of the 194 strains were defined above the 80% similarity level. The non-pink methylotrophs were recovered in two major, four minor and two single-membered clusters. One of the major clusters could be equated with Microcyclus aquaticus but strains in the other were not identified. Strains in two of the minor clusters were identified as Pseudomonas species but members of the other two minor and one of the single-strain clusters were not positively identified. They had pseudomonad properties but were not closely associated with any marker strains. The other single-membered methylotroph cluster was thought to be a cytophaga or flexibacter. The remaining minor and single-membered clusters contained only marker strains. All the PPFMs were recovered in the two remaining major clusters which were closely related to each other but not to the rest of the organisms studied. The generic assignment of the PPFMs is discussed and the suggestion is made that the genus Methylobacterium may be the most appropriate place for them, despite their apparentinability to utilize methane.

SUMMARYNumerical analysis has been carried out on I I O features of 158 named and unnamed coryneform bacteria. At the 30 yo S-level, four phena of unequal size were formed, the largest of which (phenon II) was divided into two subphena at the 45% S-level representing the genera Arthrobacter and Nocardiu. Phenon III was divided into two subphena at the 45 % S-level (IIIA and IIIB). Subphenon IIIA was made up largely of Gram-positive strains received as Flavobacterium. Subphenon IIIB contained a variety of strains including a group of cellulomonas-like organisms. Phenon IV was divided into two subphena at the 35% S-level representing the animal-pathogenic corynebacteria and Microbacterium j?avum respectively. Phenon V contained six strains of which five were plant pathogenic corynebacteria. DNA base composition determinations were carried out on representative strains and the values obtained generally correlated well with the numerical groupings. Considerable reorganization of most coryneform genera was considered necessary and suggestions for the reclassification of species and of particular strains have been made.

The present definition of the genus Methylobac.terizrm excludes organisms which do not utilize methane facultatively. However, reports have indicated that cultures of the type strain of the type species, Methylohacteririnz orgunophilirm, can lose this feature easily. Furthermore, studies have shown that M . organophilum is phenotypically highly similar to the methane-nonutilizing, pink-pigmented, facultatively methylotrophic bacteria and that the latter should be excluded from the various genera to which they have been assigned previously. Therefore, in accordance with an earlier suggestion, we propose that the description of the genus Methylobacterium be emended to permit the inclusion of methane-nonutilizing organisms which are otherwise highly similar to the type species. We further propose that all of the pink-pigmented, facultatively methylotrophic bacteria, which in c I u de the specie s Ps e ir d o m on us rho do s , Ps er i dom o r 1 as ra dio rn , an dPseudomonas inesophilica, be transferred to the genus Mt.th~lobac~teririrn.

No abstract

Fatty acid profiles of 202 coryneform and nocardioform bacteria were recorded by gas chromatography. Strains were grouped according to their profiles using mean linkage cluster analysis and similarity measures based on the correlation coefficient, the angular separation between vectors in a multidimensional space and the degree of overlap between superimposed traces. Comparisons using both real and hypothetical data showed the last of these measures to be the most effective. Strains were divided into two major groups, depending on whether they contained predominantly straight chain or iso-and anteiso-branched acids. The first group was divided into two subgroups according to the relative proportions of the characteristic acids present ; one subgroup had six clusters containing the rhodococci, nocardiae, mycobacteria and caseobacters, and the other had two containing the xanthobacters and true corynebacteria. The second group was divided into one subgroup containing strains of Arthrobacter simplex, Arthrobacter tumescens and Arthrobacter duodecadis, and one having three clusters. One cluster from this latter subgroup contained cellulomonads, one contained brevibacteria and curtobacteria and one contained arthrobacters, oerskoviae and kurthiae. Identification to generic level by fatty acid composition alone may not be feasible, but fatty acid analysis coupled with morphological examination may be sufficient to identify Corynebacterium, Arthrobacter, Cellulomonas, Oerskouia, Brevibacterium, Caseobacter, Kurthia and the A. simplexltumescens taxon. Distinction is not easy between Curtobacterium, Microbacterium and the diaminobutyric acid-containing coryneforms and between Rhodococcus, Mycobacterium and Nocardia.