108 ^ CENTURY OF PROGRESS IN THE NATURAL SCIENCES 



But, while discrediting Prevot's contention, this argument does not mean that 

 a particular type of metabolism is a more reliable index of phylogeny than is the 

 gross morphology of the cells. The ability to carry out a lactic acid fermentation, 

 for example, is not the prerogative of the "lactic acid bacteria"; it has been 

 found also in some members of the facultatively anaerobic sporeformers. Simi- 

 larly, a typical alcoholic fermentation is produced by Sarcina ventriculi and by 

 Pseudomonas lind^ieri, and a propionic acid fermentation by Propionihacte- 

 rium species as well as by some anaerobic micrococci, anaerobic sporeformers, 

 and facultatively anaerobic myxobacteria of the Cytophaga type. In these 

 cases it is as difficult to find convincing grounds for the claim that the organisms 

 characterized by similarity in metabolism are phylogenetically closely related as 

 it is to assign natural relationships primarily on the basis of cell shapes. 



Awareness of this situation led Kluyver and van Niel (1936) to propose that 

 ■a bacterial genus be defined both morphologically and biochemically. In this 

 manner cross-relations in these two respects could find adequate expression, and 

 homogeneity in the composition of the individual genera was insured. However, 

 it did not solve the problem of a phylogenetic classification; once more it was 

 necessary to make a choice between morphological and physiological characters, 

 now for delineating families, and from the foregoing discussion it would appear 

 that a decision in this respect had to be an arbitrary one. 



Besides, another difficulty presents itself, even on the genus level, because not 

 all biochemical properties appeared equally suitable as generic characters. In 

 some cases a guiding principle can be found to aid in evaluating various fea- 

 tures. Thus, the lactic acid fermentation brought about by the lactic acid bac- 

 teria, the mixed acid fermentation of Escherichia coli and its relatives, the 

 ethanol-butanediol fermentation of Aerohacter and Aerohacillus, the propionic 

 acid fermentation, the butanol-acetone fermentation, the ethanol-acetone fermen- 

 tation of Bacillus macerans, the alcoholic fermentation of Sarcina ventriculi and 

 Pseudomonas lindneri, represent as many distinctive metabolic patterns. It was 

 therefore felt that they provide legitimate criteria for separate biochemical 

 genera, while the differential utilization of some particular members of the 

 class of carbohydrates, presumably depending merely on the presence or ab- 

 sence of specific carbohydrases, was deemed useful only for the demarcation of 

 species. There are, however, many instances in which the situation is more com- 

 plicated because one and the same bacterium may exhibit a number of different 

 metabolic patterns, each one of which would be suitable for the definition of a 

 "biochemical genus." This again implies the need for making a choice. As a 

 way out of the dilemma Kluyver and van Niel (1936, p. 389) suggested: 

 ... In those cases it is, of course, desirable to classify the organism in question 

 according to its most characteristic type of katabolism, that is, the type which permits 

 the distinction from otherwise related organisms. This implies that for organisms capable 

 of development under anaerobic conditions the katabolic process involved in this mode 

 of life has been determinative, regardless of the question whether or not the organism 

 also possesses a respiratory mechanism. If two different types of anaerobic katabolism, 

 e.g., saccharolytic and proteolytic, are represesnted, the latter, as being the rarer, has 

 been decisive. 



It will be superfluous to belabor the point that this passage contains nothing 

 to suggest a phylogenetic basis for the choice, nor does it seem likely that a 

 sound one can be discovered. Nevertheless, the classification proposed has much 



