116 THE CERATOPSIA. 



widely from those already described. The horn cores, instead of being short and transverse, as in the existing bisons, are 

 lon°- and elevated, with slender, pointed ends. They have large cavities in the base, but in the upper two-thirds are nearly 

 or quite solid. Their position is well shown in the cuts below [fig. 106]. The frontal region between the horn cores is broad, 

 somewhat convex, and very rugose. 



The remains of this species are found in the sandstones of the Denver group, at the eastern base of the Rocky Mountains, 

 where they indicate a well-marked horizon, which may be called the Bison beds. These deposits are more recent than the 

 Equus beds, and are probably late Pliocene. 



The locality of the type specimen is on the banks of Green Mountain Creek, near Denver, Colo., where it was found by 

 Georo-e L. Cannon, jr., of Denver. Portions of the same specimen were subsequently secured by Whitman Cross, of the U. S, 

 Geological Survey. Other remains were obtained by G. H. Eldridge, of the Survey, and all were sent to the writer for 

 examination. 



To the erroneous determination of the nature of these remains is due the error made by 

 Marsh in determining the age of the deposits constituting the Denver group, which are now 

 known to be late Cretaceous instead of late Pliocene. 



Nearly two years later" Professor Marsh recognized the real affinities of these remains 

 and removed them from the genus Bison to Ceratops. His remarks in this connection at that 

 time were as follows: 



The bison-like horn cores figured in this journal (Am. Jour. Sci., vol. 34, p. 324) probably belonged to a member of this 

 group (the Ceratopsidse) , as already suggested by the writer.6 They were sent to him from a locality in which he had himself 

 collected Mastodon remains and other Pliocene fossils. As they agreed in all anatomical characters with the remains of 

 cavicorn mammals from that formation, they were referred to the genus Bison, under the name B. alticornis. The writer 

 has since learned that they were found in the Denver beds, which, although regarded as Tertiary, are probably Cretaceous. 

 Under these circumstances this well-marked species may be known as Ceratops alticornis until additional remains make 

 certain its true nature. 



The previous suggestion referred to by Marsh in the quotation just given certainly does 

 not make it clear that he at that time considered B. alticornis as a member of the Ceratopsidse 

 or as a dinosaur; and since, in his original description of the species already quoted, he clearly 

 states that it was found "in the sandstones of the Denver group," it is clear that he was not 

 misled by the collectors as to its stratigraphic position. The error was clearly one of erroneous 

 determination of the nature of the animal to which the remains pertained, and was entirely 

 excusable, considering the little that was then known concerning this remarkable group of 

 dinosaurs. All that was at that time known concerning the comparative osteology of the 

 vertebrata suggested its relations with the bisons among the Mammalia and, without making 

 a microscopical examination he would have been a daring anatomist who would have ventured 

 to suggest from any external anatomical characters alone that these horn cores pertained to 

 a dinosaur or other member of the Reptilia. In the article last cited, in speaking of the resem- 

 blances between these horn cores and those of the Bovidse, Marsh aptly concludes with the 

 following remark: "This accurate repetition in later and still existing forms of the highly 

 specialized weapons of an extinct group of another class is a fact of much interest." He 

 might very properly have added, with no reflection upon himself, that in the present instance 

 it has led to an unfortunate error both as to the determination of the nature of the remains 

 themselves and of the age of the deposits in which they were found. 



Why Professor Marsh should have referred this species to Ceratops rather than Triceratops 

 is not at all clear. Its affinities are certainly with the latter genus, as will appear when we 

 come to discuss the synonymy of the various genera. 



TRICERATOPS Marsh. 1889. 



Type species, T. horridus. 



Original description in Am. Jour. Sci., vol. 38, August, 18S9, p. 173. 

 Baur, G., Science, vol. 17, 1891, pp. 21fP217; Am. Naturalist, vol. 25, 1891, p. 450. 

 Ed. Am. Geologist, Am. Geologist, vol. 8, 1891, p. 56. 

 Lambe, L. M., Sum. Kept. Geol. Surv. Canada, 1898, pp. 184, 187. 

 Lydekker, R., Nature, vol. 48, 1893, p. 304. 



a Am. Jour. Sci., vol. 3S, August, 1889, pp. 174-175. 'Am. Jour. Sci., vol. 37, April, 1889, p. 335. 



