JANUABT 16, 1914] 



SCIENCE 



87 



distinction between Micrococcus and Sar- 

 cina appeared to be of quite minor impor- 

 tance. Among the yellow ehromogens we 

 found a completely parallel series of packet- 

 formers and non-packet formers exactly 

 alike in all other respects. So with lique- 

 faction of gelatine. Among aurocoeei, 

 alboeocci, micrococci and sarcinffi were 

 strains having all other properties in com- 

 mon, but differing in this one respect. 

 KLigler (1913) has recently examined the 

 cocci in the American Museum collection 

 and has found that the fifty strains repre- 

 sented fall very clearly in the genera thus 

 outlined, although he concludes that the 

 species originally described should be modi- 

 fied in certain important respects. 



The general principles of statistical 

 classification were outlined as follows by 

 Miss Rogers and myself (1906) : 



We have first plotted the frequency curve for 

 each character in order to find whether the array 

 varies about one or several modes, and where these 

 modes are situated. ... In the second place, we 

 have calculated correlation factors for the most 

 significant pairs of characters. Each mode on the 

 curves of frequency may fa.irly be taken to mark a 

 natural species or variety, and the characters which 

 vary together must form the most important basis 

 for the establishment of the larger groups. By 

 such a method alone it is possible to locate those 

 mountain peaks in the chain of bacterial varia- 

 tions which rightly deserve generic and specific 

 names. 



In the same year in which this paper was 

 published, Andrewes and Horder in Eng- 

 land presented a revision of the species in 

 the genus Streptococcus (1906) based on 

 exactly identical principles at which they 

 had independently arrived. They say: 



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

 such taxonomic problems, proved of the greatest 

 help, namely, the numerical frequency of occur- 

 rence of any given type. . . . The common types 

 stand out as mountaiu tops above their fellows, 

 each mountain connected by valleys of intermedi- 

 ate types with many of its neighbors. 



Since these suggestions were first 

 made, the statistical method has been sys- 

 tematically applied by Howe (1912) to the 

 colon group, by Stowell, Hilliard and 

 Schlesinger (1913) to the streptococci, by 

 Dr. Morse (1912) to the diphtheria group 

 and by Rogers and Davis (1912) to the 

 lactic acid groups. 



There are many other serious investiga- 

 tions of bacterial relationships which might 

 be cited, many of them made before the 

 term ' ' statistical classification ' ' was 

 thought of in this connection, but charac- 

 terized by the fact that they include a care- 

 ful comparative study of many different 

 strains with due regard to the frequency 

 with which types occur and to the special 

 importance of correlated characters. 

 Among the earlier investigations, Bei- 

 jerinck's study of the acetic acid bacteria 

 (1898), Chester's on the aerobic spore 

 formers (1904) and Hefferan's on the red 

 pigment producers (1904) are worthy of 

 special mention. More recently Edson and 

 Carpenter have given us an excellent re- 

 vision of the group of fluorescent bacteria 

 (1912). Owen (1911) has added much to 

 our knowledge of the aerobic spore form- 

 ers. White (1909) has revised the B. hul- 

 garicus group and Dr. Claypole (1913) has 

 worked out some very striking correlations 

 between immunity relations and cultural 

 characters among the streptothrices. The 

 elaborate study of the dysentery group by 

 the late Dr. Hiss (1904) and Elser and 

 Huntoon's review of the Gram-negative 

 cocci (1909) should also be mentioned in 

 this connection. There has already been 

 accumulated a considerable mass of data 

 which when critically examined and codi- 

 fied should furnish a good basis for a sys- 

 tematic arrangement of many of the smaller 

 bacterial groups. 



So far as the general classification of the 

 bacteria into larger groups, families and 



