PROSPECTS FOR A NATURAL SYSTEM OF CLASSIFICATION OF BACTERIA 



suchen, alle gegenseitigen Beziehungen der Organismen, die auf ir- 

 gendeine Verwandtschaft hindeuten, wiederzugeben'. 



The practical application of this line of thought obliges us to face 

 the question whether in the construction of the system morphological 

 or physiological characters should have priority in the expression of 

 relationships between those systematic groups which are no longer 

 morphologically and physiologically homogeneous. 



In our opinion the binomial nature of nomenclature accentuates the 

 demand that genera be systematic units which are characterized as 

 well by a more or less complete morphological homogeneity as by a 

 fundamental agreement in metabolic properties. This means, of course, 

 that for the distinction of species within a genus only characters of 

 secondary importance, such as biometric constants, hydrolytic abilities, 

 the occurrence of pigments not determining the metabolic type, etc. 

 can be applied. 



Consequently the question raised above must be answered in order 

 to make possible an arrangement of the genera in such a way as to 

 satisfy best the natural relationships existing between them. 



We have decided upon the use of morphological criteria as main 

 guiding principle in the creation of systematic units above the rank of 

 genera. In doing so we are fully aware that this choice may appear 

 arbitrary. The resulting system unintentionally suggests that a mor- 

 phological evolution has been primary and that in the various stages 

 of morphological development an independent, though sometimes 

 parallel, physiological differentiation has occurred afterwards. Yet it 

 does not seem excluded at all that in special cases the order of events 

 has been the reverse and that in reality parallel morphological evolu- 

 tions have taken place in two physiologically different groups. 



There is however, only a limited number of examples in which 

 morphological differentiation in a clearly defined physiological group 

 strongly suggests itself, whereas the instances are numerous that a 

 typical katabolic process is found in groups morphologically so un- 

 related that affinities on the basis of physiology seem fully incompatible 

 with the evolutionary idea. Moreover, in the former case the range of 

 morphological differentiation extends but over a small number of 

 closely related morphological units, as is e.g. clearly shown by the 

 groups of purple bacteria. Therefore in a mainly morphological system 

 such physiologically related groups will remain together, whilst, on 



305 



