SIGNIFICANCE OF TAXONOMIC CRITERIA 311 



and Neumann, Bergey and his colleagues, and the earlier Committee of the 

 Society of American Bacteriologists which reported in 1917 and 1920. 



As a result of the activities of the American Society, the whole question of 

 bacteriological classification and nomenclature has been reopened during recent 

 years. It cannot be said that the system propounded by the American Com- 

 mittee (1920) has met with the entire approval of bacteriologists in general ; while 

 systematists in other biological sciences would probably question the validity of 

 the whole basis upon which the classification is founded. There does not, how- 

 ever, appear to be any compelling reason for the bacteriologist to abandon, for 

 purposes of classification, the criteria on which he has come to rely for purposes 

 of identification ; and few of us would be willing to admit that our systematic 

 grouping must have a purely morphological basis, simply because structural 

 differences have been found to afford adequate classificatory criteria in the case 

 of more highly differentiated plants and animals. While admitting that morpho- 

 logical differences must be given their full weight, and accepting them as the 

 natural basis for our primary subdivisions, we might argue that our differential 

 criteria, depending as they do on differences in chemical structure rather than on 

 the gross architecture of the cell, come nearer to the heart of the matter than do 

 those adopted by botanists or zoologists. 



We may note that our assessment of the significance of any particular 

 differential criterion rests largely on a statistical basis. Our first concern is to 

 determine the variability of a given character within a particular bacterial strain. 

 If it is constant, it may be of value for purposes of classification. If it varies, 

 but in such a way that the variation is itself characteristic, it may still have 

 classificatory value. If it varies in an entirely random and unpredictable fashion, 

 it cannot be used for purposes of identification or classification. Once it has been 

 shown that a given character is of service in identifying a particular strain, we 

 can examine the distribution of this character among a sample of strains which 

 possess other characters in common. In this way we gradually obtain a picture 

 of the frequency distribution of many different characters among large samples of 

 strains. The significance we attach to any particular character then depends in the 

 main on its association with other characters. If we find that a particular group 

 of strains resemble each other in several different characters, and differ in these 

 same characters from all other groups, we feel justified in regarding the group as 

 biologically valid and in attaching an added significance to each of the associated 

 characters, as a differential criterion within the larger group of which our homo- 

 geneous group forms a part. If, on the other hand, one particular charactef varies 

 independently of all other characters, within a group which has many other charac- 

 ters in common, we shall not in general attach the same significance to it, from 

 the point of view of classification. In assessing characters in this way, we shall 

 not of course accord all characters equal rank, a priori, and limit our considera- 

 tion entirely to their frequency distribution and degree of association. Some 

 characters will be given more weight than others, and our arrangement of char- 

 acters in descending order of importance will depend entirely on the function we 

 envisage for our classification of bacteria. The classification may be predominantly 

 utilitarian. For example, as medical bacteriologists, we might concentrate on 

 the outstanding features of the medically important bacteria, and ignore all other 

 bacterial species, excepting those which are sufficiently like the medically important 

 species to cause trouble in identification. This is, in fact, the working procedure 



