160 RELATIONSHIPS OF THE COCCACE/E 



comparison and correlation, before their type centers 

 could be discovered; and fortunately this task did not 

 wait long. Andrewes and Horder (1906) have applied 

 the statistical method to the species of this group; and by 

 a vast amount of work, carried out from a sound biological 

 viewpoint, have at last brought order out of chaos. The 

 descriptions of twelve hundred strains of streptococci 

 were available to these authors — three hundred isolated by 

 Gordon from the mouth, three hundred of Houston's 

 from faeces, one hundred and seventy-two of Houston's 

 from milk, two hundred of Gordon's from air, and two 

 hundred of their own from diseased conditions (the last 

 including pneumococci as well). These records were 

 analyzed with a view to determining which of the reactions 

 were of fundamental, and which of subsidiary, importance. 

 Given strains were tested first as to the constancy of their 

 characters; and the investigators concluded that altho the 

 power to ferment salicin and mannite was sometimes lost 

 in cultivation, the constancy of a given strain was on the 

 whole surprisingly great. When they proceeded to 

 attempt the correlation of the various characters, how- 

 ever, their results at first appeared disheartening. The 

 number of types was bewildering, each variety being con- 

 nected with others by innumerable transitional forms. 



A study of the numerical frequency of various com- 

 binations showed, however, that certain types were of 

 much more common occurrence than others. Andrewes 

 and Horder's picturesque comparison of the distribution 



