26 RELATIONSHIPS OF THE COCCACE^E 



statistical method was essential for the solution of the 

 problem. 



Andrewes and Horder analyzed the records of 1200 

 different strains of streptococci which were available 

 through the painstaking investigations of Gordon and 

 Houston. The data at their disposal included tests of 

 fermentative power in a considerable number of carbo- 

 hydrate media. The observations were unfortunately 

 not quantitative; but a large number of substances were 

 tested. In spite of the bewildering. variations exhibited 

 by individual cultures the numerical frequency of 

 certain types was at once apparent. Seven such 

 types, marked by definite combinations of charac- 

 ters, formed peaks in the mountain chain of varieties, 

 the rarer variations grouping themselves about the main 

 centers and differing from them by comparatively slight 

 degrees. Furthermore, each type center was associated 

 with a definite habitat, one with the intestine of the 

 Herbivora, one with the human throat, one with patho- 

 logical conditions, etc. The results of this study will 

 be reviewed more fully in Chapter VII. It is pioneer 

 work and will doubtless require revision in many respects; 

 but its authors have succeeded in mapping out for the 

 first time a classification of the streptococci which 

 bears evidence of correspondence with the natural phy- 

 logenetic relations of the group. It complements our 

 own work which dealt mainly with the staphylococci 

 and the saprophytic forms. Together, the two inves- 



