24 RELATIONSHIPS OF THE COCCACE^l 



to problems in theoretical biology and anthropology 

 which led us to attempt the solution of the vexed question 

 of bacterial relationships by these means. In the same 

 year in which our paper was published, Andre wes and 

 Horder (1906) in England presented a study of species 

 in the genus Streptococcus, based on identical principles 

 but apparently arrived at in independence of the work 

 of others in any field. The following statement of the 

 case by these authors shows an interesting parallelism 

 with the contemporaneous passage quoted from our 

 own paper on page 18; but Andrewes and Horder 

 have expressed the principle of biometrical classification 

 with peculiar felicity. After pointing out the diffi- 

 culties in the way of classifying the streptococci they 

 proceed as follows: 



" There was, however, one guide which, as in all such 

 taxonomic problems, proved of the greatest help — 

 namely, the numerical frequency of occurrence of any 

 given type. When any arbitrary set of characters is 

 taken as a basis for the classification of a group of 

 natural objects the same phenomena are usually seen — 

 large groups of like objects connected by small groups 

 which differ from them in only one or two particulars. 

 If the numerical frequency of each individual like the 

 group is represented by the proportional height of a 

 vertical line and the lines are arranged in series the 

 commoner types stand out boldly above the rarer ones. 

 Only in nature they are plotted out, not in linear series. 



