Failing material for a more detailed comparison between the faunas of Tatrot and Pinjor, it may serve as an additional argument for keeping the two stages distinct if it can be shown that while the Tatrot and the Dhok Pathan faunas possess certain affinities to one another as listed above, which that of the Pinjor does not share, the latter on the contrary indicates very clearly that it belongs to the same epoch as the Villafranchian of Europe and of Nihowan in China and is obviously older than the Cromerian or Sicilian.

The abundance of Elephas (Archidiscodon) planifrons in the Pinjor, as in the Villafranchian of Europe, affords one of the strongest proofs of their identical age. This species coexists in the Villafranchian with Elephas (Archidiscodon) meridionalis, but the latter species alone persists into the Cromerian. Elephas (Loxodon) antiquus appears in Europe for the first time in the Cromerian. Except that E. meridionalis has not been certainly recognized in the Pinjor, although Depéret has suggested that E. hysudricus is the equivalent of a later mutation of that species which occurs at Saint Prest, and that E. cf. planifrons appears in the Tatrot, the geological distribution in India agrees perfectly with a correlation of the Pinjor with the Villafranchian. E. hysudricus probably occurs high up in the Pinjor and persists into the Boulder Conglomerate stage, while E. (Loxodon) namadicus is absent from the Pinjor but is plentiful in the Narbada beds and in the Lower Karewas of Kashmir. The occurrence of Mastodon arvernensis in the Villafranchian is paralleled by that of Mastodon sivalensis in the Pinjor. The presence of E. namadicus at Nihowan points to a somewhat younger age than the Villafranchian for those deposits.

The relationship of Rhinoceros platyrhinus to Diceros (or Dicerorhinus) etruscus is not proved. If the Indian species is a Coelodonta, as Colbert inclines to think, then it is certainly much more primitive than C. antiquitatis of the later Pleistocene. Matthew (1929, p. 535) observed that he would have expected that stage of evolution in the Miocene rather than in the Pliocene.

rine.

Following Hopwood (1937, p. 907) there are both Zebrine and Caballine types of Equus in the Pinjor, represented by E. sivalensis and E. namadicus which are Caballine and E. cautleyi which is ZebThe latter species is intermediate in size between E. robustus and E. greveyi. The small horse E. stenonis, which is much more common in Italy than the large E. robustus, has not been recorded from the Pinjor. In the Cromerian the Caballine type appears

to be much more plentiful than the Zebrine type, though a later mutation of E. robustus occurs. Except for the bare record by Lewis of the occurrence of Equus at Tatrot, I know nothing of its

nature.

Teilhard and Pivetean have provisionally referred a Chalicotheroid from Nihowan to the same genus, Nestoritherium, as the Pinjor N. sivalense, though the teeth are more hypsodont than those of the latter.

The Nihowan deposits have also yielded the large Paracamelus gigas, which according to Teilhard and Piveteau is a true Camelus, though in a slightly more primitive stage than the living forms and also than C. sivalensis and C. antiquus of the Pinjor.

They

The

The only Bovine known from the Villafranchian of Europe is Leptobos, represented by closely allied species to which the names L. elatus, L. etruscus and L. stenometopon have been given. are as plentiful as Leptobos falconeri is in the Pinjor of India. latter is for the most part more primitive than the European forms, except perhaps in the greater degree of deflection of the supraoccipital into the plane of the occipital. The genus is not known to have persisted in India beyond the Pinjor, but dentitions from the Cromerian of Süssenborn have been referred to it. In the Djetis stage of Java it certainly exists, but in a much more progressive stage.

From the Cromerian of Europe Bison priscus is recorded and possibly Bos primigenius, but these species are far more progressive than the Pinjor Bison sivalensis and Bos acutifrons. Bison palaeosinensis of Nihowan is perhaps more nearly the equivalent of the Indian species. On the other hand neither of these two extra-Indian areas contains the primitive Pinjor forms Hemibos, Bucapra and Platybos.

The Villafranchian antelopes of Europe are not as a rule comparable with those of the Pinjor. The Caprinae are represented in the former region by forms like Deperetia ardea, Procamptoceras brivatense, Nemorhaedus meneghinii, and Megalovis latifrons, and at Nihowan by a true sheep, Ovis shantungensis. The only Caprrinæ so far known from the Pinjor belong to the genus Sivacapra, which though it may be on the same lineage as Capra or Hemitragus is in a far more primitive stage, comparable only to that of Tossunoria of the Pontian of China.

The hyænas of the Pinjor are quite comparable with those of Europe. Crocuta sivalensis is closely allied to C. brevirostris of the Villafranchian. This is replaced in the Cromerian by C. spela.

Sivafelis brachygnathus is closely allied to both Felis arvernensis of Europe and to Cynaelurus pleistocaenicus of Nihowan. The type is absent from the Cromerian. Megantereon falconeri is a near relation of the wide spread M. megantercon of the Villafranchian of Europe, though slightly more progressive. The species does not persist into the Cromerian.

I have so far dealt with the views which are peculiar to De Terra, Teilhard and Lewis, which involve the Upper Siwaliks only, and which Colbert gives no indication of sharing. I now approach the main subject of this paper, the correlation of the Middle and Lower Siwalik stages, on which Colbert has laid the greatest stress. Lewis bases his correlation of this part of the series on the same premises as Matthew and Colbert and is more or less in agreement with those authors, though he considers each stage as somewhat more ancient than they do, thus approximating to my own view.

The correlation of the Dhok Pathan stage of India with the Pontian of Europe remains, as it has always been in my mind, the foundation of the whole correlation adopted. The other stages fall into place in the time scale, mainly because of the relation which their respective faunas bear to that of the Dhok Pathan, although strong confirmation is afforded both in the case of the Pinjor and the Chinji fauna by the indication which they afford of equivalent genera and species to those which occur at Perrier and La Grive Saint Alban respectively.

Matthew was less concerned to admit this close correspondence between the Dhok Pathan and the Pontian faunas than to demonstrate that the occurrence of Hipparion in India at an earlier date than the Dhok Pathan proved conclusively that the Dhok Pathan stage must be later than Pontian, from which he considered that its first appearance in both America and Europe dates: so much so that it seems as if he had often gone out of his way in an attempt to show that the Dhok Pathan species are more progressive than those of deposits like Pikermi, Samos, Maragha and the Red Hipparion beds of China.

He

Colbert (1935, pp. 21, 22) has taken a more liberal view. grants that the character of the respective stages of the Siwaliks when compared with those of European faunas seems greatly in favour of Pilgrim's views of correlation. In effect he admits that he would adopt the same correlation himself, were it not for the occurrence of Hipparion in the Lower Chinji.

Attempts have been made to reconcile the inconsistency of the faunal facies of the Siwalik stages with the assumption that Hipparion reached India from North America by: (1) on the one hand, advancing evidence that the earliest Hipparion of North America dates from the Sarmatian, and (2) on the other, stressing the fact that Hipparion appears in Europe in beds of Sarmatian age. The evidence in favour of the first of these has been succinctly summarized by Lewis (1937, p. 195). It should be remarked that McGrew and Meade (1938, p. 106) dispute Lewis' assumption that the Puente and Mint Canyon beds of California are Miocene. They, therefore, regard Hipparion as dating from the Lower Pliocene in America, from which they infer not only that the Chinji is but little older than the Pontain, but also that the Sarmatian of Europe should be placed into the Pliocene. Regarding the second the Upper Sarmatian age of the Sebastopol fauna has been commonly accepted and other occurrences of Hipparion in Eastern Europe go to confirm this. The most recent example of such Sarmatian occurrences of Hipparion in Europe has been admirably recorded by Tobien (1938, p. 177). He has also discussed other supposed Sarmatian finds of the same nature. No other identifiable mammalian remains were associated with Hipparion, except in the case of Sebastopol. There is, however, no reason to believe that any of these deposits are much older than the latter and in any case are sharply distinguished from the Tortonian of La Grive Saint-Alban. Nor is the contrary to be expected, since nothing seems to be so clearly established as the fact that this is a fauna which invaded Europe suddenly, and that these invaders, whether they arrived in the Sarmatian or the Pontian, more or less ousted and usurped the place of the previously existing fauna. In other words it is a fauna of Pontian not of Tortonian facies which entered Europe at the stage of Sebastopol. The question to be asked is: does the general composition of the Chinji fauna harmonize with such a facies? In Colbert's and my opinion it does not. Lewis does not enter into this side of the subject, but presumably he is disposed to accept Matthew's and Colbert's explanation of the facts. On the contrary I consider that the Chinji fauna might reasonably have occupied during the Tortonian, if not the same, at any rate an outlying part of the Central Asiatic region, which a stage later was to people Europe with Sarmato-Pontian migrants. I shall, therefore, endeavour to show that any other correlation which

differs materially from the one I have adopted is incredible, if not actually impossible.

It has been shown by Pilgrim, Bohlin and Colbert that the Chinji fauna contains many species which are very closely allied to Tortonian forms of Conohyus, Listriodon, Macrotherium, Dryopithecus, Amphicyon, or which are more primitive than allied species which make their first appearance in Europe in the Sarmatian or Pontian, such as primitive Tragelaphina (Sivoreas compared with Prostrepsiceros); primitive Boselaphinæ, (Strepsiportax gradiens and S. chinjiensis compared with Strepsiportax (?) latifrons); Giraffokeryx as compared with Palaeotragus quadricornis and the Helladotherines; Lycaeyna proava and L. chinjiensis as compared with L. chaeretis; Crocuta carnifer as compared with C. eximia.

According to Matthew's and Colbert's view, in spite of this ancient affinity and of its containing no species which is similar to or at the same stage as any Pontian species except Hipparion the Chinji stage is not Tortonian but Pontian. Similarly the Nagri fauna (see pages 47-50) which contains Strepsiportax, closely allied to S. (?) latifrons of the Sarmatian of Europe, and from which the typical Pontian Giraffoids, as well as Tragocerus and many other genera are absent, is not Sarmatian but Pontian or post-Pontian. The Dhok Pathan fauna which exhibits a very close parallel to that of the Pontian of Europe, with many almost identical species, is not Pontian but Middle Pliocene (Astian), the equivalent of Roussillon and Monpellier.

A detailed comparison of the Dhok Pathan fauna with that of the Pontian of Europe and China is unnecessary, since this was made by Pilgrim, first in 1913 (pp. 280-307) and later in 1931 (Pilgrim, 1931, p. 152), and their close correspondence may be seen from a study of the faunal lists of Colbert (1935, pp. 29-36) and Pilgrim (1938, pp. 5-7). For easy reference, however, the closest specific affinities found in the faunas under consideration may be tabulated here.

Mastodon (Choerolophodon) corrugatus to Mastodon (Choerolophodon) pentelici of Pikermi.

Mastodon (Tetralophodon) perimensis to Mastodon (Tetralophodon) longirostris of Eppelsheim, etc.

Listriodon pentapotamiae, a survival from the Chinji, to L. splendens, a survival from the Tortonian.

Tragocerus punjabicus to T. amalthea.