THE FAMILIES AND GENERA OF THE BACTERIA 193 



recent report of the committee on generic types of the Botanical 

 Society of America, published in Science for April 4, 1919, has 

 urged that the application of generic names should be determined 

 by type species rather than by attempts at generic characteriza- 

 tion; and with this point of view the members of the committee 

 on characterization and classification of bacterial types are in ac- 

 cord. The committee was agreed that type species with proper 

 literature references should be included in all the genera listed, 

 but that it was also desirable to include brief characterizations of 

 the genera themselves. The situation with which we deal in 

 attempting to classify the bacteria is somewhat different from 

 that which exists among the higher plants. In the latter case 

 actual type species have been deposited in herbaria and are 

 available for reference ; while among the bacteria this is not the 

 case except for a few type species which have recently been de- 

 posited in the collection of the American Museum of Natural His- 

 tory in New York. In consideration of the uncertainty which 

 surrounds the description of many bacterial species it was felt 

 that it would be helpful to furnish at least tentative characteri- 

 zations of the genera presented, and this policy has therefore been 

 pursued in the preparation of section III of this report. 



It will perhaps be convenient to indicate briefly the general 

 changes in classification which distinguish the present report 

 from that of 1917. The main departures are as follows: 



The group recognized in 1917 as the family Mycobacteriaceae 

 has now been elevated to the rank of an order Actinomycetales, 

 and divided into two families, Actinomycetaceae and Myco- 

 bacteriaceae. To the former family we have added the genera 

 Actinobacillus and Erysipelothrix, and we have omitted Nocardia, 

 which Breed (1919) has recently shown should be combined for 

 the present with Actinomyces. To the second family we have 

 added the genus Pfeifferella. 



The Nitrobacteriaceae have been divided into two tribes, the 

 Nitrobactereae and Azotobactereae, and the definition of the 

 family has been modified to permit the inclusion of Rhizobium 

 which recent investigations have shown to possess peritrichous 

 flagella, but whose general characteristics ally it clearly with 



