Three perissodactyl phyla occur, namely the Hyracotheriinæ, Lophiodontinæ and Helaletinæ, whereas at the same period in America we find the Hyracotheriina, Tapiride, and Helaletine. Without exception in the Lutétien representatives of the perissodactyl families Lophiodontine and Hyracotherine the premolars are simpler than the molars and these animals are therefore in a stage of evolution corresponding with that which we find in the Wind River beds. The horses so far as I can judge from personal study, from FILHOL's descriptions and from figures, (GERVAIS, '59) P. Suillus, P. parvulus, P. duvalii, all belong to the primitive stage, namely, premolars simpler than molars, no mesostyle, and are therefore in a Wind River (Protorohippus) rather than Bridger (Orohippus) stage of development. FILHOL ('88, p. 182) lays great emphasis upon the fact that all the so-called Pachynolophus' of Issel, Pépieux and Lautrec have the premolars simpler than the molars. Furthermore in beds of undoubted Issel, Argenton or Buchsweiler age, no complete Anchilophus types of premolars (pm = m) occur. As for the oldest Artiodactyla in either country, COPE ('82, p. 71) has compared Lemoine's Lophiodocharus peroni of the Argiles-à-lignites with his Trigonolestes brachystomus, from the Wind River. Among the primates the little known Heterohyus armatus Gervais, distantly resembles Microsyops of the Bridger in its molar teeth only, the premolars being simpler than in the Bridger species.

These are significant facts. So far as they go they indicate that the known beds of Lutétien formation (having a thickness of 45-24 metres) are by no means equivalent to the Bridger Beds (having a thickness of 800 metres), as heretofore stated, but they merely correspond to a section of the Lower Bridger or more probably of the Upper Wind River formation.

It is true that in the Helaletine, or cursorial Lophiodonts, in the fauna of Egerkingen and Lissicu, namely H. cartieri, H. annectens (and perhaps Helaletes (Hyrachyus) intermedius of Selles-sur-Cher), the third and fourth premolars have double internal lobes like those of H. (Desmatotherium) guyotii of the Bridger. But it must be remembered as regards both Eger

kingen and Lissieu that they are composite faunæ, containing upper Eocene forms mingled with the middle Eocene forms, therefore, they cannot be cited at all as proofs of synchronism. Similar Helaletes-like teeth are described by FILHOL from Buchsweiler, Alsace, namely the type 3d and 4th premolars of his Palæotapirus buxovillanus ('88, p. 179, pl. XIX, fig. 4), which certainly belong not to the Tapiride but to the Helaletine, a sub-family of Lophiodontida. On the other hand, the upper molar and the lower jaw assigned to Hyrachyus intermedius (FILHOL '88, p. 114, pl. XXX, figs. 8 and 6) from Argenton resemble the Helaletine in a Wasatch or Heptodon stage of development because they are small and simple.

7. Middle Eocene, Bartonien, apparently Equivalent to the Lower Bridger

This substage receives its name from the Barton Clays of England (100 metres). The Sables de Beauchamp, marine (15 metres) is succeeded by the partly lacustrine Calcaire de Saint Ouen with which the fresh water Grès de Césseras (Hérault) are considered parallel.

From the Grès de Césseras a few mammals are recorded. The Cesserasictis antiquus (FILHOL '88, p. 182, pl. XIX, fig. 3) type is a small lophiodont jaw with molar teeth which resemble those of Helaletes of the Bridger except in the extreme simplicity of the supposed 4th premolar.

If FILHOL's identification and description is correct no comparison can be made with our Bridger Helaletes which has a partly compound fourth premolar. The Lophiodon cesserassicum FILHOL (L. occitanicum, GERVAIS) is judging by GERVAIS' figures (pl. 18, fig. 7), one of the Equidæ in a Bridger stage of development.

The American parallel of the Bartonien is probably Lower Bridger but it cannot be determined until we secure a more exact knowledge of the state of molar and premolar evolution of the few ungulate fossils which it contains.

The writer is chiefly indebted to Professor ALBERT GAUDRY for the arrangement of the lower Eocene in the accompanying 'Third Trial Sheet.'

8. Upper Eocene, Ligurien

The summit of the French Eocene is characterized geographically by the recession of the northern gulf on its western borders and by numerous small freshwater lake and river deposits in the south and southwest of France, in Switzerland, and on the German border (CANU, '95, Plate 44).

In the Paris Basin, made famous by the classic researches of Cuvier and Brogniart, is the Gypse de Montmartre (55 metres) partly marine, partly lacustrine; at its summit are 20 metres of gypsum which contain most of the mammals described by CUVIER. Above are the lacustrine Marnes de Pantin.

Parallel with the Gypse are the rich Lignites de la Débruge (Vaucluse, 2 metres).

Parallel with the Gypse in the South are the beds of St. Hippolyte de Caton (Gard) recently described by DEPÉRET; of Castlenaudry (Aude); of Lautrec (Tarn) described by NOULET ('63) also by GERVAIS ('69).

There are also the lacustrine limestones of Carcassonne, near the Pyrenees, and the localities Mas-Saintes-Puelles and Villeneuve-la-Comptal, Castres. To the west in Germany are the fissure deposits or Bohnerzen of Heidenheim (Mittelfranken) Ulm, Pappenheim, Fronstetten1 (Swabian Alps), Sigmaringen; to the south the older fissure deposits of the Phosphorites du Quercy, and the fissures of Egerkingen and Lissieu.

This period contrasts with all its predecessors by the superbly full fauna which it contains; we feel for the first time that the fossil record is approximately representative of the living fauna. It is greatly enriched by the composite parallel fauna of the Siderolithique de Mauremont and the newer portions of the composite fauna of Egerkingen and Lissieu.

Lautrec, undoubtedly Upper Eocene, contains a very large Lophiodon, L. lautricense of especial interest, because it is apparently the last of its race. It is probable that the large Lophiodon of Heidenheim, with complex premolars, is related to the Lautrec type. In the Heidenheim specimen the second and 1 Fronstetten fauna described by Jäger, Fraas, Quenstedt and v. Meyer. ANNALS N. Y. ACAD. SCI., XIII, July 19, 1900-2

third superior premolars have double internal cusps.

Lissieu as studied by Depéret is mainly middle Eocene but it contains some important Upper Eocene forms, while Egerkingen has a rich representation of Upper Eocene types.

The large L. rhinocerodes Rütimeyer, of Egerkingen is, however, not of the Heidenheim type because it has simple upper premolars associated with it; it is an older representative of the large race of Lophiodontida.

Mauremont is considered mainly, if not exclusively, of Upper Eocene age.

GENERAL CHARACTERS.

(1). This fauna is much more modern than that of the Grès de Césseras, or of the Calcaire Grossier and Issel; the great advance in the structure of the teeth especially seen in a comparison of Propaleotherium and Paleotherium is proof of modernization. Palæotherium is now the predominating type of Perissodactyl, although a large form of Lophiodon survives.

(2). Secondly, the composite beds of Egerkingen and Lissieu furnish the ancestry of certain types of Gypse Artiodactyla and in these beds we also find certain other forms transitional between the Issel stage and the Gypse stage.

(3). Thirdly, the Gypse, is a very highly specialized and differentiated fauna including many artiodactyls and other types the ancestry of which is known neither in Europe or America and has not thus far been found in Egerkingen or Lissieu.

(4). Fourth, the Ligurien is widely distinct faunally from the American Upper Eocene or Uinta with which it has been heretofore paralleled. At no period of the Tertiary were the Nearctic and Palearctic faunæ so widely separated. In fact a much wider gap exists between Western America and Europe in the Upper Eocene than in the preceding Lower and Middle Eocene or in the succeeding lower Oliogocene.

The resemblances or parallels with America are mostly limited to one genus of horses (Pachynolophus), which occur in both countries, to one Creodont Hyaenodon, and to the ancestors of the Canida and Viverride which occur in both countries.

(5). Contrasts. The Cheiroptera and Insectivora of these two

regions cannot be compared until the American forms named by Marsh are adequately studied. The Primates have no direct parallels. Among the Perissodactyla, Palæotherium, Palaplotherium, and Anchilophus have no parallels in America. The Selenodont Artiodactyla of the two continents are widely distinct; the Gypse selenodont Artiodactyla have no parallels in America. The bunodont Artiodactyla have not yet been carefully compared. (6). There are therefore comparatively few direct reasons for considering the Gypse and Uinta as nearly contemporaneous but there is a substantial indirect reason namely that they both closely underly Oligocene Beds in which there suddenly reappears a marked community of fauna in the Nearctic and Palæarctic regions. In other words the Gypse bears a relation to the Ronzon similar to that which the Upper Bridger bears to the Upper Uinta and White River.

The most significant fact is the apparent invasion of the Palæarctic region in the Upper Eocene by a great variety of Artiodactyla which mingled with the older phyla of France and Germany. Where did these animals come from? Not from Asia, certainly, because some of them would have found their way also into the Nearctic, probably therefore from Africa or the Ethiopian Region.

9. Composite, Imperfectly Stratified Fissure Deposits of Middle Eocene to Middle Oligocene Age

The most famous of these fissure deposits are those of Quercy, Egerkingen, Mauremont, Fronstetten.

In the Swiss Jura are the Bohnerzen, mainly non-calcareous reddish clay nodules with pisolithic iron grains. The siderolithic earths, Sidérolithiques, typically at Mauremont, found in Jurassic limestone fissures are so called because they contain grains of iron, imbedded in concretions probably of mineral spring origin, associated with travertines. A special type of fissure deposits, analogous to the above in certain respects are the Phosphorites, typically represented in Quercy but characteristic also of other periods.