o28 II. 1*\ OSBOKN CLASSIFICATION OF THE MASTODONTOIDEA 



As clearly pointed out in the second paper on this subject (Osborn, 

 1921. 514, pages 3-5), these four superfamilies are clearly distinguished 

 from each other by profound differences in the adaptations of the cutting 

 teeth, namely, the first and second pairs of superior and inferior incisors 

 characteristic of all Proboscidea. The author is not prepared at present 

 to add to what has been said in previous papers regarding the mceritheres 

 and diuotheres. 



Species 



Continued observation of the types on which the species in all parts of 

 the world have been founded and of the genotypes from which the genera 

 received their designations is gradually bringing order out of chaos. The 

 synonymy in the 47 generic names is most difficult of solution. There is 

 far less synonymy among the species ; in fact, the specific and subspecific 

 stages are far more numerous than has hitherto been supposed. It is not 

 proposed at present, however, to multiply the species; rather to attempt 

 to arrange the 170 or more species hitherto described in their natural 

 lines of phyletic descent. 



Phyla 



There are still wide differences of opinion about the reality of the poly- 

 phyletic division of the Proboscidea. Such division is still considered a 

 matter of theory, whereas it will be shown to rest on actual demonstration 

 of a very profound divergence of many lines, due to the principles of 

 adaptive radiation, local and continental. 



In 1918 I divided the Proboscidea into te7i phyla. In 1921 I divide 

 them into tivelve phyla. The application of generic names to these phyla 

 presents a very difficult problem in nomenclature, on Avhich I am securing 

 the cooperation and advice of all the experts on this subject, especially of 

 such leading authorities as Allen, Palmer, Matthew, Andrews, and Schle- 

 singer. The difficulty arises from the fact, pointed out in my first paper, 

 that the generic names were based on species which belonged within two, 

 three, and sometimes four distinct phyla. For example, the generic name 

 Tetralopliodon Warren, 1852, was based on the species M. latidens, M. 

 arvernensis, M. sivalensis, animals Avhich certainly belong to two distinct 

 phyla — stegodontines and brevirostrines. The same term, Tetralophodon, 

 was applied by Falconer in 1857 to animals belonging to four distinct 

 phyla, namely, M. lorifjirostris (a longirostrine), M. latidens (a stego- 

 dontine), M. andinm (an American brevirostrine), M. arvernensis and 

 M. sivalensis (two Eurasiatic brevirostrines). The name Tetralopliodon 

 Falconer would apply admirably to M. longirostris, but unfortunately it 



