JANUABT 16, 1914] 



SCIENCE 



83 



tively far more important than others. Of 

 540 dextrose fermenting bacilli freshly iso- 

 lated from the intestine Howe found that 

 practically all attacked lactose. Saccharose 

 divided the group into two approximately 

 equal subgroups, 58 per cent, attacking this 

 sugar and 41 per cent, failing to do so. Of 

 the 58 per cent, all but 5 per cent, also at- 

 tacked raffinose so that the dextrose-laetose- 

 saccharose forms may be considered inter- 

 mediate variants between the two main di- 



COMBINATION OF FERMENTATIVE 



REACTIONS mM, 



COLON BACILLI 



(Howe) 



visions which ferment either dextrose and 

 lactose alone or all four sugars (Fig. 3). 

 It seems clear that such quantitative stud- 

 ies furnish the surest basis for deciding 

 which sugars are primarily important in 

 classification and we may safely conclude 

 that so far as these four sugars are con- 

 cerned there are only two important divi- 

 sions of the colon group. The use of 

 dulcite which occupies an equaUy promi- 

 nent place in the classification of MacCon- 

 key (1909) and Jackson (1911), rests on 

 no such biometrieal basis and we have no 



good reason to think it deserves special im- 

 portance. In the same way the admirable 

 study by Stowell, Hilliard and Schlesinger 

 (1915) of the streptococci shows that the 

 five groups fermenting respectively dex- 

 trose only, dextrose and lactose, dextrose, 

 lactose and saccharose, dextrose, lactose, 

 saccharose, and rafEinose and dextrose, lac- 

 tose, saccharose, rafiinose and salicin are 

 quantitatively of special importance and 

 include between them 68 per cent, of 240 

 strains studied. 



A second conception, of much assistance 

 in the classification of bacteria, is the prin- 

 ciple that special weight should be given to 

 characters quantitative or qualitative, 

 which are found to be correlated with each 

 other in a number of different types. The 

 principle of numerical frequency offers a 

 basis for characterizing the individual 

 types and the principle of correlation a 

 basis for classifying these types in accord 

 with their biological relationships. 



The early bacteriologists established a 

 dozen genera, such as Streptococcus, Sar- 

 cina, Bacillus, Bacterium, and the like, 

 based entirely on a few obvious morpho- 

 logical characters. Some of these genera 

 are undoubtedly valid. Others like those 

 which are based only on the presence or 

 absence of flagella are quite as certainly 

 invalid. No one familiar with the colon 

 group can hold that it is reasonable to 

 place the common type of motile colon 

 bacillus in the genus Bacillus along with 

 B. mycoides, B. aerogenes, B. anthracis, 

 B. prodigiosus, B. radicicola and B. tetani 

 and to place an organism having all its 

 other properties identical but lacking fla- 

 gella in the genus Bacterium. The same 

 arguments hold true against the genera 

 Planococcus and Planosarcina among the 

 cocci. "We find in several of the major 

 groups motile and non-motile forms which 

 are precisely alike in half a dozen respects 

 and are clearly minor varieties of the same 



