H. F. OSBORN — MAMMALIAN PALÆONTOLOGY 109 
anomalous skull representing a new family (Desmostylidæ fam. nov.) ei- 
ther of hypsodont Sirenia or of Proboscidea, and Merrram' has recogni- 
zed as a similar form occurring on the coast of California the genus 
Desmostylus first noticed by Marsn. 
The phylogenetic series is all too limited, the horses being sparsely re- 
presented by species of Neohipparion (Ginrex?) and a doubtful Pliohip- 
pus, the Camelidæ by Pliauchenia, the Dicotylidæ by several species of 
Platygonus, the Carnivora by an Amphicyon and other doubtful species 
of Canidæ. The collateral lines of Camelidæ, so far as we know, died out, 
and the adaptive radiation of the true camels begins. 
However, no generalisations can as yet be made from this scanty fauna ; 
we are confronted with more gaps in our knowledge and more unsolved 
problems than in any other period. Among these, the direct ancestry of 
the South American cameloids (Auchenia) as well as of the true camels 
(Camelus) should be found. We also should find here the stages directly 
ancestral to the horse (Equus), because it now appears certain that Marsæ’s 
Phohippus was an Upper Miocene and not a Pliocene animal, and was, 
moreover, apparently on a side line not leading directly into Equus 
(Giozey, Fig. 4). Thus not only is the Pliocene plains fauna sparsely 
known but the Pliocene forest fauna is wholly unknown. 
Tue PrrISTOCENE FAUNA. 
Equivalent to (1) the Preglacial, Forest Beds of Norfolk (St. Prest, Dur- 
fort, Malbattu, Peyrolles), (2) Glacial, (Mid-Pleistocene, Lower Mid-Pleis- 
tocene), (3) Postglacial deposits of Northern Europe and Asia. 
Here again American palæontology is far behind that of Europe as to 
knowledge of the chronological succession of deposits, and a vast amount 
of work remains to be done in the discrimination of geological and fau- 
pal stages, in the comparison of Easternand Western cave — and sand — 
deposits, and in the coordination of the first appearance of man with that 
of the mammalian succession. 
The advent of the true Equus marks the base of our Pleistocene, as 
shown in the sand deposits of the Western plains in the so called ÆEquus 
beds. The geographical distribution and remarkable adaptive variation 
of the Pleistocene horses has now been fully worked out (GipLey *), pro- 
ving that there are ten species characteristic of different localities, and 
ranging in size from Æ. giganteus, larger than any modern horse, to the 
1 Science, n. s., Vol. xvi, Oct. 31, 1902, p. 714. 
? A New Three-toed Horse. Bull. Amer. Mus. Nat. Hist., Vol. xix, 1903, pp. 465-476. 
8 Tooth Characters and Revision of the Genus Equus. Bull. Amer. Mus. Nat. 
Hist., Vol. iv, 1901, pp. 91-142. 
