22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 128 



Ml and M2 with pronounced hypoconulid developed on crest or saddle between 

 hypoconid and entoconid: Diacodexinae (generally weak in Bunophorus) . 



Ml and M2 with weak hypoconulid often on both cingulum posteriorly and in 

 saddle between hypoconid and entoconid, very close to hypoconid : Helohyus. 



Ml and M2 of forms having crescentic lower molars with hypoconulid often 

 though not invariably developed as a posteriorly directed spur from poste- 

 rior crest of hypoconid: AGRIOCHOERIDAE and OROMERYCIDAE. 



Ml and Ma of forms having crescentic lower molars with hypoconulid often 

 though not invariably developed as an enlarged or emphasized postero- 

 internal extremity of the crest of the hypoconid : Leptotragulinae. 



Ml and M.i essentially without hypoconulid: Parahyus, Achaenodon, and Poii- 

 brotheriinae; sometimes AGRIOCHOERIDAE and Leptotragulinae. 



SYSTEMATIC REVISION 

 Family DICHOBUNIDAE Gill, 1872 



As arranged by Simpson (1945) this family includes the two sub- 

 families Dichobuninae and Homacodontinae. Within the Dicho- 

 buninae he has included, along with the various European genera, 

 the three North American lower Eocene forms, Diacodexis, Wa- 

 satchia, and Bunophorus. Structurally these three are similar to one 

 another and are properly grouped together, but basically they are 

 less like the dichobunes, as represented by such genera as Dichobune, 

 Meniscodon and M ouillacitherium, than are the homacodonts. I am 

 convinced that their degree of relationship is best indicated by sepa- 

 rating them into a separate subfamily within the Dichobunidae and 

 designated, as shown in the accompanying chart, by the name of 

 Diacodexinae. 



Among the middle Eocene forms, Simpson, as tentatively sug- 

 gested by Matthew and Granger (1925), has included Helohyus in 

 the Choeropotamidae. I am rather inclined to believe, as Stehlin 

 (1906) has indicated and as followed by Sinclair (1914), that 

 Helohyus has affinities with the Dichobunidae. The degree of rela- 

 tionship, however, I again regard as best represented by a subfamily 

 separation, including the end product Achaenodon, an arrangement 

 anticipated by Matthew and Granger (1925) as an alternate possi- 

 bility. Nevertheless, their indication that under such an arrangement 

 the forms included would comprise the family Helohyidae is followed 

 only to a degree, because, referring to Matthew's (1910, p. 41) earlier 

 thinking in connection with Eotylopus: "If we adopt the 'linear' 

 system and ignore the more important and obvious structural differ- 

 ences between animals, on the plea that they are merely stages in 

 specialization, if we scatter apart a closely related group of ancestral 

 forms among widely divergent types to which they have given rise. 



