Z PROCEEDINGS OF THE NATIONAL MUSEUM vol 82 



which dam the median valleys of P^~*, no accessory folds occur on the 

 cheek teeth. Although the teeth are worn, there can be no doubt 

 of the presence of mures on P^"* as in D. armatum (Yale Peabody 

 Museum No. 10235) from the Upper Oligocene of the north fork of 

 White River (presumably in South Dakota rather than Nebraska). 

 The type (Y. P. M. No. 10003), from the John Day of Oregon, like- 

 wise has mures on P^~^ but differs in their absence on the fourth 

 premolars. The internal cingula are interrupted by both protocone 

 and hypocone in P* and M^ and are complete in P^~^ and M^. The 

 protocone of IVP is moderately demarked from the protoconule by 

 grooves, both anteriorly and posteriorly; that of M^ is barely demarked 

 from the protoconule anteriorly and is set ofl' by a moderate constric- 

 tion to the rear. This is the typical condition in members of the 

 Suhhyracodon-Diceratherium line. The nasal horn cores are somewhat 

 less elongated and rather more prominent and everted than in either 

 of the other two skulls of D. armatum. The infraorbital canal opens 

 above the anterior portion of P". 



The greatest apparent difference from the type specimen is that in 

 the Montana individual the nasals extend almost twice as far in front 

 of the horn cores. However, as indicated by Peterson (1920, fig. 10, 

 p. 415, and pi. 57), the anterior ends of the nasals are broken off in the 

 type so that this difference has no real significance. The Montana 

 specimen is almost exactly the same size as the type of D. armatum, as 

 shown by the measurements in Table 1 ; in fact, my measurements of 

 the Montana skull are as close to either Troxell's or Peterson's 

 measurements of the type skull as their measurements of that skull 

 are to each other. Inspection of these figures indicates that such dis- 

 crepancies as occur are due to choice of different loci for taking meas- 

 urements. The chief discrepancy is in the antero-posterior dimension 

 of the individual teeth. If the greatest measurement along the 

 ectoloph is taken, instead of the length along the midline of the tooth, 

 the agreement is considerably closer. There is, then, every reason for 

 referring the Montana skull to Diceratherium armatum Marsh, as even 

 the individual differences are of rather a minor character. 



From the viewpoint of paleogeographic distribution, this specimen 

 of D. armatum partially connects the two previously widely separated 

 occurrences of this species in Oregon and the Great Plains. 



This specimen's chief importance, however, is that it dates, in part 

 at least, the beds on the west side of the lower Madison River, which 

 are apparently unconformable below the late Miocene or early Pliocene 

 Madison Valley formation. Douglass, in 1903 (p. 149), perhaps re- 

 ferred to these beds, saying: "In the lower Madison Valley, where 

 the upper [White River] beds are so well exposed, the material is 

 mostly fine. The exact horizon of these upper beds is uncertain, as no 

 good mammals have been found, but they lie unconformably under the 



