H. F. OSBORN — MAMMALIAN PALÆONTOLOGY 101 
lution of the four phyla of great Oligocene titanotheres. Among the 
Artiodactyla the rare Middle Eocene forms still require elucidation, but 
the Camelidæ have been traced definitely into the diminutive Upper 
Eocene (Uinta) Protylopus (Scorr', Worrmax?). Two distinct phyla of 
Oreodontidæ have also been traced back in the Upper Eocene into the 
genera Protagrichærus and Protoreodon (Scorr). Among the enemies of 
these animals, the Canidæ have been traced into the Upper Eocene genera 
Prodaphænus and Uintacyon, and Marsxs Middle Eocene Vulpavus has 
also proved to be a member of the true Canidæ, although, its relation- 
ships are not exactly determined (Worrman, Marruew). The supposed 
ancestry (Worrman) of the Felidæ in the Kocene in the problematical 
genus Ælurotherium has been disproved (MATTHEW). 
Still undiscovered or unrecognized in the Eocene both of America and 
Eurasia are the ancestors of the true Rhinocerotidæ which suddenly 
appear in the Oligocene. The Basal Kocene ancestry of the Rodentia is 
still satisfied only provisionally by the family Mixodectidæ, belonging to 
the somewhat hypothetical Proglires (OsBorx *); the teeth seem to be ap- 
proaching those of the Rodents but knowledge of the skeleton is neces- 
sary to determine whether they may not after all be remotely related to 
the Lemurs (order Cheiromyoidea) as Core and Worrmax have sug- 
gested. Worrmax is strongly of the opinion that the Eocene Primates 
(Notharctidæ, Anaptomorphidæ) are not Lemuroidea, and that the former 
family are distinctly South American; this also requires confirmation. 
Search for the exact relations and points of connection between the 
Carnivora and Creodonta, has thus far been entirely without definite 
success; in other words, the true Carnivora seem to be as separate from 
the Creodonta as the true Perissodactyla are from the Condylarthra. 
As regards the Artiodactyla, as yet very little is known of the Middle 
and Lower Eocene stages, among which it is especially important to test 
the truth of Scorr’s‘ broad generalization that the American Artiodactyla 
should all be regarded as affiliated to the Tylopoda as a stem group from 
which not only the Camelidæ evolved but also the other distinctively 
American Artiodactyls, such as the Oreodontidæ, and that even the tra- 
guloid forms are of tylopodous affinity and merely parallel or analogous 
to the true Tragulines of Eurasia. There is no doubt that such an adap- 
tive radiation from a Tylopod stem is possible and that there is conside- 
! The Selenodont Artiodactyls of the Uinta Eocene. Trans. Wagner Free Inst. 
Sci. Phil., vi, 1899, p. 100. 
? The Extinct Camelidæ. Bull. Amer. Mus. Nat. Hist., Vol. x, 1898,. pp. 93-142. 
3 Bull. Amer. Mus. Nat. Hist., XVI, 1902. 
# The Selenodont Artiodactyles of the Uinta Eocene. Trans. Wagner Free Inst. 
Sci. Phil., vi, 1899, p. 100. 
