PROSPECTS FOR A NATURAL SYSTEM OF CLASSIFICATION OF BACTERIA 



judged from its results. A system which unites in one genus Proteus 

 vulgaris and Lactobacillus delbruckii, and assumes close relationships of 

 these with Acetobacter, Rhizobium (all in one and the same subfamily 

 of the Eisenbergiinae) and with Sclerothrix [Mycobacterium tuberculosis) 

 and Corynebacterium by placing all these genera in one of the 15 fam- 

 ilies, condemns itself. 



Janke [1926], who in 1926 had already published a critical essay on 

 the subject of bacterial systematics, later developed a system which is 

 clearly influenced by Enderlein's views [Janke, 1929]. It seems super- 

 fluous to give here a complete survey of Janke's system. We will only 

 remark that we encounter in this system many well-known groups - 

 to which Janke assigns the rank of families - viz., Coccaceae, Bacillaceae, 

 Bacteriaceae, Corynebacteriaceae, Spirillaceae, Spirochaetaceae, Desmobacte- 

 riaceae and Myxobacteriaceae. A discussion of" the question whether the 

 establishment of these units as families is more or less justified may be 

 omitted here. Our remarks will be confined to the subdivision of the 

 families. Janke's critical attitude towards the use of physiological 

 characters for classificatory purposes leads him to restrict his genera 

 to those groups which can be identified with the aid of morphological 

 characters only. This means that e.g. in the family of the Bacillaceae 

 the common aerobic and anaerobic sporeforming bacteria are united 

 in the one genus Bacillus thus rejecting the genus Clostridium of va- 

 rious older systems. Moreover, many bacteriologists will be shocked 

 to meet with the genus Azotobacter in this family. The justifica- 

 tion of this procedure is found in the statement: 'Sporenbildende 

 Stabchen gehoren zum Entwicklungskreis (Lohnis)'. The rather 

 doubtful observations of Lohnis have, however, never been corrob- 

 orated. It is also worth mentioning that the external morphology 

 which is responsible for the maintenance of a separate genus Azoto- 

 bacter in Janke's system is frequently nearly duplicated in species 

 closely related to Bacillus megaterium [Dianowa and Woroschilowa, 

 1 931] which is accepted by this author as a type species for a sub- 

 group of the genus Bacillus. 



The trend of thought mentioned above also accounts for the occur- 

 rence of only two genera, Bacterium and Fusiformis, in the family Bacte- 

 riaceae. The weakness of Janke's principles of classification is clearly 

 evidenced by the heterogeneity of the genus Bacterium. The diversity of 

 organisms collected by Enderlein in the subfamily of the Eisenbergiinae 



299 



