has been made for it. The sequence of tests employed has largely eliminated the necessity to treat an organism as positive or negative with respect to a character for which the information was not given. In the few" instances where this device has been employed, the species involved has been cited in the key, and the assumption made has been noted or the species has been traced through the key as far as the information permitted, a note having been made to this effect at the appropriate point.
Cowan's assertion (Canad. Jour. Microbiol., 2, 1956, 212-219) that the previous system broke down because of the use of the "positive or negative" approach, while apparently theoretically correct, was, in fact, not so. Such a device obviously could not have been used where genera with single species were involved. However, with the larger genera it proved to be valid where it was employed simply because the genera already had species, adequately described, which were either positive or negative with respect to the character, and the genera therefore appeared at two or more places in the key. The use of the "positive or negative" approach merely placed the poorly described species in one or the other of these categories.
In the present key, pathogenicity to animals has not been used as a sole differentiating character. It has been coupled with other tests in the separation of these genera. The plant pathogens and the Rhizobia have been separated on the basis of pathogenicity and nodule formation respectively. They have, in addition, been treated as organisms isolated from the soil, have been traced through the key as far as described characters would permit and have been cited at the appropriate points. Species of Pseudomonas producing water-soluble pigments have been separated on this characteristic. They were, however, also checked through the key as though not pigmented. With few exceptions those adequately described terminated at points where non-pigmented pseudomonads were located; the few which did not have been individually cited. Of fifteen species for which no sugar reactions were cited, thirteen terminated at Pseudomonas when the possible "positive or negative" combinations for glucose and lactose were applied. The others were non-motile and have been cited in the key.
The description of the genus Paracolohactrum as presented in the Manual contains too little information to be of use in the key. Reference was made to the original paper of Stuart, AVheeler, Rustigian and Zimmerman (Jour. Bact., 45, 1943, 101-119), and the strain descriptions used. While these strains do not appear as such in the Manual, the value of such strain specifications over generalizations for the species should be apparent.
The recommendations regarding designations of groups in the Enter ohacteriaceae published by the Enterobacteriaceae Subcommittee of the Nomenclature Committee of the International Association of Microbiologists (Internat. Bull. Bact. Nomen. and Taxon., 4, 1954, 1-94) are also indicated in the key by the insertion of the generic name followed by "Rome, 1953".
The primary division on the basis of cell width is purely arbitrary. A great deal of latitude has been allowed in respect to this character. Cells with widths
