WlNSLOW : A STATISTICAL CRITERION FOR BACTERIA 35 



recognize as species (Lancet 1906^: 708), Their work greatly 

 extended ours, as we had worked scarcely at all on the particular 

 genus with which they dealt. The combined effect of the two 

 investigations is to emphasize strongly the value of the statistical 

 method in bacterial classification ; and there seems no reason why 

 the same principles should not be of similar value in the system- 

 atic study of algae and fungi and Protozoa and other simple and 

 variable forms. 



Among other things that these studies emphasized is the 

 importance of physiological differences in bacterial classification. 

 Systematic bacteriology has been greatly retarded by the undue 

 emphasis laid on morphological characters. Reasoning from 

 analogy with higher plants, many bacteriologists have refused to 

 base generic distinctions upon anything but morphological differ- 

 ences. As there are very few morphological differences in the 

 groups there has been no rational generic classification. The dis- 

 tinction between physiological and morphological characters is 

 merely a superficial one. Both are presumably due to chemical 

 modifications of protoplasm ; and there is no reason to suppose 

 that a protoplasmic property which manifests itself in the size and 

 arrangement of parts is any more fundamental than one which 

 manifests itself in the ability to utilize a certain food stuff. It is 

 precisely along the lines of metabolism that the bacteria have 

 attained their extraordinary degree of differentiation. The higher 

 plants have developed complex structural modifications to enable 

 them to absorb food materials of certain limited kinds and to utilize 

 the sun*s energy in building them up into protoplasm. Meanwhile 

 the bacteria have maintained themselves by acquiring the power 

 of assimilating simple and abundant foods of the most varied sorts. 

 Evolution has developed gross structure in one case without alter- 

 ing metabolism ; it has produced a diverse metaboHsm in the other 

 case, without altering gross structure. There is as wide a differ- 

 ence in metabolism between the Pneumococci and the nitrifying 

 bacteria as there is in structure between a liverwort and an oak. 

 So-called physiological differences are quite as important in one 

 case as so-called morphological differences in the other. 



The characters which are of prime systematic importance will 

 naturally vary according to the particular group of bacteria which 



