830 11. F. OSBOKN CLASSIFICATION OF THE MASTODONTOIDEA 



Superior canines, rounded upcurved Subfamilies 



tuslis, with enamel band on convex 

 outer surface, gradually disappear- 

 ing. 



(1) Longirostrimne, (2) Rhynchoros- 

 trina-, (3) Serridentinne, subfamily 

 nov., (4) Notorostrinai, subfamily 

 Subfamily, Mastodontince nov., (5) Brevirostrinie. 



(monophyletic) 



(Jenus, PalcBomastodon, Oligocene. 

 (ienus, Mastodon, Miocene to Pleisto- 

 cene. 



Family, Bunomastodontid^i: 



The Longirostriiiffi (1) and the Ehynchorostrinae (2), also the Brevi- 

 rostrinffi (5), have been clearly defined in previous papers. 



(3) The Serridentince apparently spring from M. turicense Schinz, 

 182-i, of the Middle Miocene of France and Switzerland — a rare animal, 

 probably because a forest dweller. The grinders are readily distinguished 

 by a prominent spur which projects from the inner apex of the crests in 

 the lower teeth and from the outer apex of the crests in the upper teeth. 

 It is recognized again in the Miocene Siwaliks of India. It appears again 

 in the Lower Pliocene {Trilopliodon serridens Cope) of the Clarendon 

 beds of Texas; also in the T. serridens cimarronis of Texas, represented 

 in the American IMuseum collection by a hitherto undescribed complete 

 skull, which proves that this phylum belongs near the Bunomastodontinse, 

 altliough the grinding teeth lack the trefoils characteristic of that phylum. 

 These animals have usually been placed with M. tapiro'ides of the lopho- 

 (jont or mastodontine phyla of France. They certainly possess cutting 

 teeth of bunomastodont type and are probably remote from the true 

 mastodontines. 



(4) The Notorostrince include all the Central and South American 

 brevirostrines, which are also abundant in the southwestern United 

 States — for example, Texas (Blanco formation), Mexico, the Andean, 

 and the Pampean regions. They were distinguished in my second paper 

 as "The brevirostrines of South America," embracing the classic species 

 M. andium Cuv., i¥. humholdtii Cuv. ; also M. tropicus Cope (Valley of 

 Mexico), M. successor Cope (Blanco of Texas). They possibly sprang 

 from or are related to Euhelodon morrilli Barbour, 1913, from the Lower 

 Pliocene, Nebraska. The fact that these animals are not only profoundly 

 separated from the otlier bunomastodonts by the progressive abbreviation 

 of the jaw, but that they are the only members of the great order Pro- 

 boscidea which, so far as known at tlie present time, entered the South 

 American continent, M-ill probably convince reluctant paleontologists that 



