58 OEOLOGICAL SURVEY OP TEXAS. 



that is, it has four well separated cross-crests, and a very rudimental heel. 

 The external half of each cross-crest wears into a trefoil, while the inter- 

 nal half is simple and undivided, and is a little anterior in position to 

 the external half. The tooth continues its width posteriorly, so that the 

 transverse diameters at the first and fourth cross-crests are equal. A 

 marked character of the species is the elevation of the anterior part of 

 the ramus and the decurvature of the symphysis, from which it results 

 that the superior face of the symphysis, or the spout, descends very 

 steeply to its extremitj^ from the second true molar quite as in the proxi- 

 mal part of the spout of Dinotherium. It has a very short horizontal 

 portion anterior to the second true molar to represent the long horizontal 

 production in Dibelodon tropicus, Mastodon americanus, etc. The sym- 

 physis is also much compressed above, so that the spout is narrow. The 

 extremity of the spout is produced and conti-acted, and slightly recurved 

 at the extremity, and tliere issues from the right side a well developed, 

 vertically compressed mandibular tusk. On the left side is the empty 

 alveolus of its counterpart, which was of much larger diameters than 

 that of the right side. The portions of superior tusks found in Cali- 

 fornia are stated by Leidy to possess an enamel band, and as there are 

 various reasons for supposing that they belong to the T. shepardii, I refer 

 provisionally the species to Tetrabelodon rather than to Dibelodon, as 

 heretofore. 



The contracted symphysis steeply descending forwards, and expanding 

 downwards at the base, from a posterior elevation, distinguishes this 

 species from any of those of the genera Tetrabelodon, Dibelodon, or 

 Mastodon, known to me. Hence there is no question of its difference 

 from the T. productus, Cope, with which it was identified by Leidy,* 

 where the symphysis is flat and much longer. In the Dibelodon andium, 

 according to D' Orbigny and Burmeister, the symphysis is not elevated 

 behind, is produced and decurved at the extremity, and has a wide spout; 

 all characters quite different from what is seen in T. shepardii. The 

 symphysis is totally distinct in the D. tropicus, as can be seen in the ac- 

 companying plates, and the last true molar is much more complexed in 

 the latter. I formerly identified the species so abundant in the Equus 

 beds of the valley of Mexico with this species, t This determination must 

 now be reconsidered, since the form of the mandibular symphysis is en- 

 tirely different; there is no mandibular tusk, and the last lower molar is 

 not identical in form,t though I formerly thought it not so different as 



*By reference to it of specimens of T. productus from Santa Fe. New Mexico, 

 in Eeport U. S. Geological Survey Terrs., 1, 1873, p. 235. 



tProceeds. Am. Philosoph. Soc, 1884, p. 5. 



JA figure of this too,th is given by Felix in the Paleontographica, Vol. 

 XXXVII, PI. 30, 1891. 



