BACTERIAL CLASSIFICATION 17 



red. Twort (1907) reports that he was able by con- 

 tinued cultivation in sugar media to develop fermenta- 

 tive power in bacteria of the B. enteritidis type, thus 

 breaking down the barrier between this group and 

 that of B. coll. Schierbeck (1900) records a perma- 

 nent loss of fermentative power in lactic acid bacteria 

 exposed to the action of phenol. 



In general, habitat species of definite characters are 

 so commonly . associated with particular environments 

 as to leave no doubt of their reality. If the same 

 type is found again and again in the same environ- 

 ment, it must practically receive systematic recognition, 

 whether its members are connected by direct phylog- 

 eny or not. 



Just what rank to give to various groups among the 

 bacteria is naturally not easy to determine. Species 

 may often be fixed among the higher organisms by the 

 test of mutual fertility; but here this criterion is lacking. 

 The application of a specific name to any particular center 

 of variation must therefore be a somewhat arbitrary act. 

 The Linnaean nomenclature has however great practical 

 advantages over any other system of notation. It 

 implies no artificial order but a varying degree of 

 family relationship; and thus corresponds most nearly 

 to the complex truth. The rank of larger systematic 

 units (families and genera) may well be reserved for those 

 profound modifications which alter the whole center of 

 gravity of the organism, while species may be defined by 



