VAN NIEL SYSTEMATICS OF THE BACTERIA AND BLUEGREEN ALGAE 109 



to recommend it, because it permits the ready assignment of a particular bac- 

 terium to a specific and small group as soon as its general morphological and 

 biochemical characters are known. Final identification then requires compari- 

 son with other members of only this assemblage. The advantage is, therefore, 

 of the same kind as that offered by Cohn's "form genera," and the categories 

 resulting from the combination of morphological and biochemical properties are, 

 in a sense, quite comparable though more numerous. In view of the great in- 

 crease in the number of different types of bacteria discovered in the course of 

 time this is a distinct benefit. Undoubtedly, such strictly utilitarian considera- 

 tions were responsible for the application of biochemical criteria in the manner 

 outlined above, as shown especially by the decision to use the "rarer" of two 

 otherwise equivalent characters. 



But if the homogeneous, morphologically and biochemically defined genera 

 cannot lay claim to phylogenetic significance, the superstructures of tribes, fami- 

 lies, and orders can do so even less. It follows that the existing systems of clas- 

 sification of the bacteria and bluegreen algae should not be considered "natural" 

 ones. If this be granted, the question whether retention of such systems is ad- 

 visable can be examined more critically. 



At first sight the now more or less generally accepted genera and families 

 of these organisms, even if devoid of phylogenetic meaning, might appear to 

 serve as a fully satisfactory framework for purely determinative purposes. This, 

 however, can be contested on the ground that they are too rigid, because the 

 families, tribes, and orders represent collections of genera grouped together on 

 the basis of only one set of arbitrarily chosen "primary" characters. While these 

 may be the most useful ones as determinative aids in some instances, in others a 

 different set of primary divisions would be preferable, thereby yielding a super- 

 structure of different composition. It is obviously inadmissible to include a par- 

 ticular "genus" in two or more different families, tribes, or orders. But if these 

 larger groups are considered as no more than convenient contrivances for rapid 

 identification, there is no need to insist on an "either-or" approach. By discon- 

 tinuing the use of families, tribes, and orders it becomes possible to construct a 

 diversity of groupings in which all the different opportunities for emphasizing 

 similarities in various respects can be expressed. It seems to me a dubious gain 

 to have all the photosynthetic bacteria assembled in a suborder, Khodobacteri- 

 ineae, if this practice eliminates the possibility of recognizing the existence of 

 the large group of "sulfur bacteria" comprising only some of the photosynthetic 

 bacteria in addition to organisms now incorporated in the orders Eubacteriales 

 (genus ThiohaciUus) and Chlamydobacteriales " (families Beggiatoaceae and 

 Achromatiaceae). Such an entity as the sulfur bacteria remains an extremely 

 useful assemblage, since it represents an ecological-physiological community of 

 all the conspicuous inhabitants of natural environments in which hydrogen 

 sulfide is present. 



It is not hereb}^ intended to dispute the probability that the photosynthetic 

 bacteria actually represent a phylogenetically related group, nor that the Beg- 

 giatoaceae might be similarly regarded. But the phylogenetic relationships of 

 the other "sulfur bacteria" are far less certain. Clearly, it is not imperative 

 that even the probable affinities of the first-mentioned organisms be given recog- 

 nition by uniting them into a family, tribe, suborder, or order; and if doing so 



