134 PROCEEDINGS OF THE PALEONTOLOGICAL SOCIETY 



The chief advantage of the vertical phylogenetic classification is that it 

 brings animals together in similar or closely allied lines of evolutionary de- 

 scent ; it corresponds with the branches and subbranches of the family tree. 

 The chief difficulty with phylogenetic classification is a technical one, namely, 

 to harmonize it with the Liunjean and the prevailing zoological systems of 

 family, subfamily, and generic nomenclature, which are alike based on the 

 affinities displayed between the existing terminal twigs of the branches and 

 subbranches rather than on the phyletic ancestral lines which connect these 

 twigs with their several ancestral branches. Sometimes the subfamilies pro- 

 posed by zo()logists conform to the phyletic lines discovered by paleontologists ; 

 sometimes they do not. 



The present classification and nomenclature of the Proboscidea illustrate 

 afresh the confusion, at first glimpse apparently hopeless, resulting from the 

 morphological classification and nomenclature of Linnaeus and of various pale- 

 ontologists, following the zoological standards, such as were embraced by Cope. 

 Upward of forty generic names have been applied to the mastodons and ele- 

 phants, and, as pointed out by Matthew,® there is no uniformity in the usage 

 of these generic terms, nor has any principle of arrangement been worked out 

 by which we may at least hegin an advance toward a permanent system of 

 nomenclature of this highly important and interesting group. 



In the present paper, which is the result of studies begun in 1902 and of 

 observations carried on in American and European museums, with the valua- 

 ble aid of the recent rearrangement of the collections of Proboscidea in the 

 American Museum of Natural History by Dr. W. D. Matthew, I essay a phy- 

 logenetic classification. This attempt, aided by the recent observations of LuU,^ 

 Matthew,® and Barbour,^ is preliminary to a more thorough review which is in 

 preparation by the author.* 



It will probably subserve clearness to present at once the following key to 

 the proposed phy-logenetic classification, in which are shown at least eleven 

 distinct phyla of proboscidians, grouped into five subfamilies and three families. 



ORDER PROBOSCIDEA 



Families 

 DINOTHERIJS : 



I. Dinotheriidce, crested teeth, down-turned tusks. 

 II. Masfoclontidw, crested and cone teeth. 

 MASTODONTS. A. Buxolophodont, cone-and-crest-teeth mastodonts. 

 1. Bunomastodontince : 



5 R. S. Lull : The evolution of the elephant. Am. Jour. Sci., vol. xxv, Mar., 1908, pp. 

 169-212, figs. 1-27, 4 charts ; reprinted in Smiths. Report for 1908, No. 1909, pp. 641-674. 



^ W. D. Matthew : The generic nomenclature of the Proboscidea. Read before the 

 Paleontological Society, Pittsburgh. .Tan. 1, 1918. 



' E. H. Barbour : Mammalian fossils from Devils Gulch. Nebraska Geol. Survey, vol. 

 4, pt. i, Dec, 1913, pp. 177-190, pis. 1-13. A new lougirostral mastodon from Cherry 

 County, Nebraska. Nebraska Geol. Survey, vol. 4, pt. 14, Sept. 15, 1914, pp. 213-222, 

 pis. 1-6, figs. 1-6 (tailpiece). A new lougirostral mastodon from Nebraska. Tetrabelodon 

 osborni, sp. nov. Am. Jour. Sci., vol. xli, June, 1916, pp. 522-529, figs. 1-4. 



* A memoir on the phylogeny of the Proboscidea, with illustrations of the principal 

 American types of mastodon and elephants in the American Museum of Natural History. 



