194 



University of California Publications. [Geology 



Miocene and the Columbia Lava in the Middle Miocene, since 

 the former rests on the Upper Eocene and the latter is overlain 

 by beds determined from their flora as Upper Miocene. Follow- 

 ing the four-fold division which is becoming more and more 

 generally accepted by American geologists, the John Day is 

 referable mainly if not entirely to the Oligocene, and a study 

 of its fauna has shown the propriety of such reference. "Taken 

 together, the Canidae and Felidae of the John Day represent 

 a stage of evolution somewhat more advanced than that reached 

 in the White River and less advanced than that of the Loup 

 Fork. Compared with the known faunas of Europe they appear 

 to be not older than the Middle Oligocene of Fontainbleau and 

 not as young as the Middle Miocene of Sansan'V"' It is not 

 yet possible to speak so definitely regarding the degree of ad- 

 vancement of other phyla in the John Day fauna, but the per- 

 sistence of primitive types {ElotJierium, Agriochoerus, Mesohip- 

 pus, Protapirus) and the entire absence of the hypsodont camels 

 and horses characteristic of the Middle and Upper Miocene of 

 North America is certainly striking, and, when combined with 

 the stratigraphic evidence just presented, favors the reference 

 of the John Day formation mainly to the Oligocene, although 

 it does not exclude the possibility that the Upper John Day may 

 overlap in part on the Lower Miocene, of which the fauna is at 

 present incompletely known 51 . 



50 Merriam, J. C. Univ. Gal. Publ. Geol., Vol. 5, p. 64. 



=1 Peterson, O. A. The Agate Spring Fossil Quarry. Annals Carnegie 

 Museum, Vol. 3, pp. 487-494, 1906. The Miocene Beds of Western Ne- 

 braska and Eastern Wyoming and their Vertebrate Fauna. Annals Carnegie 

 Mus., Vol. 4, No. 1, March 21, 19(17. 



Matthew, W. D. A Lower Miocene Fauna from South Dakota. Bull. 

 Am. Mus. Nat. Hist., Vol. 23, p. 169, March 14, 1907. 



The recent papers of Peterson and Matthew, appearing since this article 

 was prepared, have shown the existence on the eastern side of the Cordil- 

 leran range of early Miocene faunas closely related to that of the John Day. 

 In both western Nebraska and South Dakota faunas are found 

 containing generic types previously known only from the John Day; but 

 in both of these regions the stage of evolution of the forms most closely 

 related to John Day types, as also of the total assemblage of forms, seems 

 a little more advanced than that of the John Day. As the eastern and 

 western faunas were quite distant from each other geographically, and 

 might therefore be expected to differ somewhat, it is not impossible that 

 the earlier portion of the Miocene beds east of the Cordilleran range cor- 

 responds in time to the upper portion of the John Day. A more complete 

 knowledge of all of these faunas will be necessary before their exact rela- 

 tionships can be determined, and most important in this connection is the 

 study of scattered patches of middle Tertiary beds which are geographically 

 intermediate. 



