ORIGIN, RELATIONSHIPS, AND PHYLOGENY 23 



indigenous fauna. However, it may be argued equally well that this family was evolving in the 

 same or neighboring areas before the Uinta and that a change in conditions of preservation brought 

 them thus suddenly into our range of observation. 



Throughout most of the Eocene there was a struggle between the indigenous and the immi- 

 grant mammals for supremacy which was finally won by the latter, in part because of their superior 

 mental equipment. Several genera of the merycoidodonts have preserved to a large degree many 

 primitive characters, common to most of the Eocene artiodactyls and presumably inherited from 

 creodont ancestry. A few years ago (1923) the writer pointed out some of these "carnivorous" or 

 primitive features in the skull and skeleton. 



Lydekker (1883) described a superior molar from the lower Manchhars of the Laid Hills of 

 Sind, India, which he was convinced belonged to A griochoerus . He definitely considered that it did 

 not belong to any of the then known Indian artiodactyls. I very much doubt this reference to 

 either the agriochcerids or the merycoidodonts, although I do believe that the ancestral stock of 

 both groups had an Eurasiatic origin, whence it spread into India and over the Eocene land bridge 

 into North America. Their distribution along the western half of this continent, including Canada, 

 may argue also in favor of this origin. North and South America likewise were united during much 

 of early Eocene time, but there are no indications of affinities between any of the South American 

 fauna? and the merycoidodonts. Ameghino (1889, pp. 576-578) described the palate of an artio- 

 dactyl (Diplotremus agrestis) from the lower Pliocene beds of Argentina and referred it to the 

 Merycoidodontidse. This reference is certainly erroneous and I believe that the form probably 

 belongs with the deer family, although of this I cannot be positive, as I have not seen the specimen. 

 Ameghino seemed to consider it the South American equivalent of Cyclopidius. 



Classification and Relationship: The family name, Merycoidodontidas, is used through- 

 out this monograph rather than Agriochceridas. Merycoidodon was described in 1848, while 

 A griochoerus did not appear until 1850. Again, the former family comprises more than eight times 

 as many genera and species as the latter. I believe that the family names of the merycoidodonts and 

 of the agriochcerids should have equal rank, and, if either be a subfamily, it should be the agrio- 

 chcerids on the basis of priority of description and of relative numbers of species included in the 

 group. Leidy ( 1 869) first used both the family names of Oreodontida; and Agriochcerida; and con- 

 sidered them of equal rank, the former on page 7 1 and the latter on page 131. The name Oreodon 

 was first used in 1851, three years later than Merycoidodon and one year after A griochoerus. 

 Therefore my conclusion is that the superfamily should be Merycoidodontoidea, with the two 

 families Agriochceridaj and Merycoidodontidas. 



This family has been classified with various diverse groups, such as the Suidse, Tylopoda, and 

 tragulines, with all of which it has certain characters in common, but when the characters are con- 

 sidered as a unit it is not close to any of them. Hay (1902, p. 662) placed it in the superfamily 

 Cameloidea, but in 1930 (p. 776) he erected a new suborder, Agriochceri formes, for the family. 

 Matthew (1929, p. 404) pointed out relationship with the Anthracotheres, thus following some 

 of the earlier writers. He proposed the group name of Ancodonta to include the Anthraco- 

 theriids, Anoplotheriidas, Camotheriidas, and Oreodontidas. I believe that this arrangement cor- 

 rectly expresses the relationships as far as the evidence warrants. Romer (1933) places these 

 groups in the Protoselenodontia, adding the Dichobunida; and the Homacodontidas but excluding the 

 Anthracotheriidas, which he believes lie in the stem stock of the Hippopotamida;. 



When we examine the European fauna of the Eocene, we find a certain amount of similarity in 

 the dentition of the more primitive types of anthracotheres and their relatives, the semiselenodont 

 artiodactyls, known as the ancodonts, or hyopotamids, the former group being known mainly from 

 fragmentary skulls. These skulls were long-muzzled, with very long, compressed molars. Seem- 

 ingly few descriptions of foot structure are published. The American Oligocene genera of these two 

 groups have advanced in the direction of pig-like specializations in the skull which remove them far 

 from the oreodonts, although in limbs and feet there is a marked resemblance. There are other 

 European Eocene genera showing suggestions of affinity with the oreodonts, such as Mixtotherium, 



