THE MAIN OUTLINES OF BACTERIAL CLASSIFICATION

THE MAIN OUTLINES OF BACTERIAL CLASSIFICATION
R. Y. STANIER AND C. B. VAN NIEL
Hopkins Marine Station, Pacific Grove, California
Received for publication February 8, 1941

"Was diese Wissenschaft betrifft,
Es ist so schwer, den falschen Weg zu meiden,
Es liegt in ihr so viel verborgnes Gift,
Und von der Arzenei ist's kaum zu unterscheiden."

-GOETHE

1. INTRODUCTION
Although a great deal has been written on bacterial taxonomy during the past few decades, a perusal of the literature shows that for the most part this work has been restricted to the classification of the Eubacteriales alone. Since the early days of microbiology, comparatively little attention has been paid to the broader problem of delimiting and defining the Schizomycetes as a whole and the major groups contained therein. Nevertheless, it can hardly be contended that this is an unimportant aspect of bacterial taxonomy; on the contrary, a clear recognition of the larger natural groups of bacteria, their characteristics and relationships, would seem to be an indispensable basis for more detailed work. The increased use of Bergey's Manual of Determinative Bacteriology for purposes of identification, together with the obvious attempts made by the present Board of Editors to seek collaboration with specialists on various groups, make it likely that this Manual will ultimately become the internationally recognized and authoritative handbook on bacterial taxonomy. Nevertheless, in its main outlines the system used in Bergey's Manual is still far from satisfactory. There will in due course be a succeeding edition, and it is with the hope of contributing some constructive suggestions for its outline that the present essay is offered.

2. PHYLOGENY AND EMPIRICISM IN BACTERIAL SYSTEMATICS

In most biological fields it is considered a truism to state that the only satisfactory basis for the construction of a rational system of classification is the phylogenetic one. Nevertheless, "realistic" bacteriologists show a curious aversion to the attempted use of phylogeny in bacterial systematics. This is well illustrated, for example, by the statement of Breed (1939): Realistic workers have on their side been impatient with idealists who have introduced many ... unjustified speculations regarding relationships between the various groups of bacteria. To what may we ascribe this distrust of phylogeny? In part it is undoubtedly due to the unsatisfactory nature of certain systems, purportedly based on phylogeny, which have been proposed in the past. However, the mere fact that a particular phylogenetic scheme has been shown to be unsound by later work is not a valid reason for total rejection of the phylogenetic approach. Another important reason for the "realistic" attitude is the widespread belief that bacteria present too few characters on which schemes of relationships (and hence a natural system) can
be based. It is our belief that such pessimism is not entirely justified, and that at present some relationships can be recognized which can well be incorporated in a system of classification. Even granting that the true course of evolution can never be known and that any phylogenetic system has to be based to some extent on hypothesis, there is good reason to prefer an admittedly imperfect natural system to a purely empirical one. A phylogenetic system has at least a rational basis, and can be altered and improved as new facts come to light; its very weaknesses will suggest the type of experimental work necessary for improvement. On the other hand, an empirical system is largely unmodifiable because the differential characters employed are arbitrarily chosen and usually cannot be altered to any great extent without disrupting the whole system. Its sole ostensible advantage is its greater immediate practical utility; but if the differential characters used are not mutually exclusive (and suchmutual exclusiveness may be difficult to attain when the criteria employed are purely arbitrary) even this advantage disappears. The wide separation of closely related groups caused by the use of arbitrary differential characters naturally enough shocks "idealists," but when these characters make it impossible to tell with certainty in what order a given organism belongs, an empirical system loses its value even for "realists." It seems unnecessary to give here an exhaustive review of bacterial systematics. The reader is referred to Buchanan's scholarly treatise on general systematic bacteriology for an excellent survey of this field up to 1925. More recent literature has been briefly reviewed by Breed in the latest edition of Bergey's Manual. We shall, therefore, restrict ourselves to a critique of Bergey's system which illustrates well the weaknesses of the empirical approach.