BACTERIAL CLASSIFICATION 2/ 



tigations cover the family as a whole with reasonable 

 fullness. 



The success of Andrewes and Horder in classifying 

 the streptococci is strong evidence of the value of bio- 

 metrical methods in systematic bacteriology. Their 

 results and our own have led us to the conviction that 

 these methods offer. the key to the natural relationships 

 of the bacteria. A few general rules of procedure 

 must however be accepted before detailed application of 

 statistical methods can be fruitful. 



In the first place it must be admitted, before any 

 progress is possible, that physiological characters are as 

 important as morphological characters in this group ; and 

 that they alone, if of sufficient degree, may serve for 

 specific and generic definition. One of the greatest 

 drawbacks to systematic bacteriology ' has been the 

 attempt to find morphological differences where none exist, 

 and the corresponding failure to study physiological 

 characters with adequate exactness. Yet it is precisely 

 along the lines of metabolism that the bacteria have 

 attained their extraordinary degree of differentiation. 

 Other plants and animals, under the stress of natural 

 selection, have developed complex systems of organs 

 for procuring food of certain limited kinds and for 

 escaping harmful environmental conditions. The bac- 

 teria have maintained themselves by acquiring the 

 power of assimilating simple and abundant foods of 

 varied sorts, with a special resistance against cold and 



