Page images

Siwalik Boulder Conglomerate, there seems no escape from his conclusion. Where he seems to go beyond his facts is in assuming that the remainder of the Upper Siwalik, namely the Tatrot and Pinjor stages must correspond to the first ice advance and the first interglacial stage, and must, accordingly be classed as Lower Pleistocene. This does not follow. Since his observations show, again on the assumption that his interpretation of the Tawi section is correct, that there is an enormous erosion unconformity between the Siwalik Boulder Conglomerate and the underlying Siwalik stages, it is obviously not merely possible but likely that the Lower Pleistocene is represented by a break in deposition in India itself and that the Pinjor stage corresponds with the Upper Pliocene (Villafranchian) of Europe. How far this is consistent with the discovery by Dr. Barnum Brown of Bubalus platyceros from the top of the Siwalik Boulder Conglomerate one-half mile west of Siswan, Siwalik Hills, is another question. Other instances of typical Pinjor species occurring in the Boulder Conglomerate would certainly militate against the identification of the Tawi conglomerate which I have provisionally accepted.

On the Kashmir plateau itself De Terra appears to be convinced (1936, p. 686) that the Lower Karewas correspond to the first glacial period and are undoubtedly Lower Pleistocene. He infers that the fauna found in these beds justifies a correlation with the Pinjor

The only fossils, so far as I am aware, which he has recorded from the Lower Karewas are Equus, Elephas namadicus, Bos, Sus, Rhinoceros, Cervus, Felis, Sivatherium. Of these Elephas namalicus is undoubtedly post-Pinjor and is abundant in the older Narbada Alluvium, while it is not impossible that Sivatherium should have persisted beyond the Pinjor stage. Moreover, De Terra has found no evidences of glacial action in the Pinjor any more than there are in the Val d'Arno or Perrier, so that a correlation of the Pinjor stage with the Upper Pliocene (Villafranchian) of Europe, more in harmony with the character of its fauna and its stratigraphic and faunistic relations to the stages of Tatrot and Dhok Pathan, is more probable than De Terra's rather forced correlation with the Lower Karewas. A comparison of the Pinjor fauna with those of the Val d'Arno, Perrier and Senèze is discussed later in this paper.

To such authorities as Hopwood (1935, p. 46), with whom Lewis agrees, the advent of Equus, Elephas and Bos marks the beginning of the Pleistocene. Consequently both Pinjorand Val d'Arno


the case.

are Pleistocene, and each one of De Terra's glacial stages must be shifted to a later period. Colbert hardly goes as far as this, although the position in which he places the Val d'Arno on the boundary between Pliocene and Pleistocene would imply some disposition to adopt Hopwood's view. I doubt whether De Terra realizes the full implications of the correlation which he has adopted.

Hopwood's inclination to choose the advent of a few important invading types as marking the beginning of each main period is very attractive in the absence of geological evidence. In order that this criterion should be valid throughout the area of distribution of such types he assumes that the migration of any form is practically instantaneous. So far as actual travel is concerned, I quite agree with him that there is no reason to suppose that a species should take long to move from one area into another, but it seems obvious that physical or climatic conditions might alter

The Pontian fauna of Europe offers a classic example of a fauna which we have every reason to believe existed in Asia and America for at least a stage before it penetrated into Europe, its arrival into the latter continent being marvellously sudden. It follows that Hopwood's correlation test can only be rigidly applied in a single region, like Europe which has been thoroughly explored. It is liable to fail when an attempt is made to extend its application elsewhere, since we cannot be sure that the form in question invaded other parts of the world simultaneously with its invasion of Europe to say nothing of the area in which it originated, very probably an unknown one and to which for some reason or other it may have been confined for a considerable time. Further, more intensive collecting in any particular region may alter the opinion which we first formed as to the arrival or origin of a genus. A very apposite example of this is to hand in the case of the occurrence of the genera Equus and Elephus in India. Originally regarded as making their first appearance in the Pinjor stage, this seemed to afford an extremely sound reason for correlating the Pinjor with the Villafranchian of Europe in the Lower Pleistocene.

Now it appears that both genera also occur in the Tatrot stage, but it is quite unreasonable to consider the Tatrot as equivalent to the Villafranchian and so Lower Pleistocene, in defiance of the remainder of the fauna.

It is not of great concern to me here as to whether it is considered more appropriate to place the deposits of the Val d'Arno, Senèze and the Red Crag into the Pliocene or Pleistocene, or those of Pikermi and Samos into the Miocene or Pliocene. On the other hand, it is obviously of great importance from a zoological point of view that the Siwalik stages should be correlated with the mammaliferous stages of Europe with as great an approach to accuracy as possible. For this reason it is a pity that certain authors, in assigning an age to the Siwalik stages, should prejudge the question and use the terms Miocene, Pliocene and Pleistocene, thus leaving it doubtful with which stage they think that the Indian beds they may be discussing correspond. Lewis' correlation is in large part a

compromise between De Terra's, Colbert's and my own. He agrees with Hopwood that the appearance of Equus should mark the beginning of Pleistocene time. At the same time he considers that the onset of glacial conditions should equally mark it. These two standards, as far as we know at present, are incompatible. Equus appears in the Val d'Arno and Perrier, but there is, so far as I am aware, no indication of glaciation. Equus

Equus is also pre-glacial in certain Texas deposits. Like De Terra, Lewis places the Tatrot and Pinjor stages into the Lower Pleistocene. While De Terra and Teilhard have merely confessed to an inability to distinguish the Tatrot and Pinjor stages, though admitting the possibility of separating them if more was known about the Tatrot fauna, Lewis has definitely united them into a single stage with the name of Tatrot. Both Lewis and De Terra seem to have been led to consider the Tatrot and Pinjor faunas as one and the same, by their wish to minimize the fact that the Pinjor fauna corresponds so well with that of the Villafranchian of Europe that it is difficult to regard it as belonging to the later Cromerian or Sicilian stage. By assuming that the whole of the Pinjor fauna is spread through the Upper Siwaliks the situation is rendered somewhat easier. De Terra places the Dhok Pathan stage into the Upper Pliocene, whereas Lewis correlates the Dhok Pathan with the stage of Roussillon as Middle Pliocene, assuming that there was a long break between the Dhok Pathan and Tatrot corresponding to which there is no vertebrate fauna known in India. The Nagri stage becomes the equivalent of the Pontian, the Chinji of the deposits of Sebastopol, and the Kamlial of the combined La Grive Saint Alban and Sansan stages.

Lewis' statement as to the existence of a great unconformity between the Dhok Pathan and the Tatrot stages, makes no reference

to the fact, of which he was probably unaware, that I had previously (1913, pp. 274-277) stressed this very strongly. I was not, however, disposed to regard the break as universal even in the Salt Range area, and perhaps nowhere as of such long duration as Lewis thinks. Not only do the faunas of the two stages which are discussed below militate against the long time interval which Lewis predicates, but also the concordance of dip which, so far as I am aware, everywhere exists between them.

On the contrary Lewis has represented a very small break between the Pinjor stage and the Boulder Conglomerate, to which he has given the name Tawi stage. This is quite inconsistent if he accepts De Terra's interpretation of the Tawi section, where the Boulder Conglomerate is shown resting on the highly tilted denuded edges of the Pinjor strata. The remarkable change in the character of the rocks tells equally in the same direction, while the known fauna of the Boulder Conglomerate and the older Narbada Alluvium presents no features which can be held to be opposed to De Terra's correlation of the two.

A still more revolutionary idea is propounded by Lewis, when he places the deposits of Roussillon and Montpellier into the Plaisancian and leaves the entire Astian stage without any vertebrate fauna in Europe. The assumed absence of Astian land deposits in Europe is evidently intended by Lewis to correspond with the long interval which in India he assumed to have elapsed between the Dhok Pathan, ex hypothesi the equivalent of Roussillon, and the Tatrot, ex hypothesi the equivalent of the Villafranchian.

I think it may perhaps be advantageous briefly to summarize the arguments against removing Roussillon so far in time from Perrier and Senèze, not because Depéret's classic studies, which have been accepted in Europe for the last fifty years need any support from me, but because the arguments employed may be applicable in estimating the degree of affinity between the Tatrot and the Dhok Pathan on the one hand and the Pinjor on the other. The evidence of certain recent determinations, unknown to Depéret, may also not be without some value.

The character of the European fauna altered considerably between the Pontian and the stage of Roussillon. Numerous families and genera have disappeared. Dinotherium, Mastodon (Trilophodon), Mastodon (Tetralophodon) of the type of 1. longirostris, Chalicotheriidæ, Accratherium, Chilotherium, Tragulidæ, Giraflidæ, Listriodon, the large pigs of the Microstonyx type, Hippotraginae, Tragelaphinæ, Tragocerus, Criotherium, Amphicyon, Ursavus, Indarctos, Sivaonyr, Enhydriodon, Promeles, Promephitis, Ictitherium, Hyænidae of the type of Crocuta eximia, Machairodus, Paramachaerodus, Steneofiber, and Dryopithecus are among the chief of these.

There are very few surviving genera from the Pontian, and almost all of these differ specifically. Hipparion crassum is markedly different not only from H. gracile but also from H. mediterraneum and other Pontian species. Parabos cordieri and P. boodon have been referred to the same genus as Parabos (?) macedoniae from Salonika, but the material of the latter is hardly sufficient for comparison. Epimachairodus tarakliensis is probably ancestral to E. crenatidens, which abounds in the Villafranchian, but has also been identified by Schaub from Roussillon. Tapirus persists, but only as the different species, T. arvernensis. Hystrix primigenius is found both at Pikermi and Roussillon. Cricetus kormosi from the Pontian of Polgardi may be regarded as ancestral to C. angustidens of Perpignan. Mus gaudryi is not closely related to Mus donnezani.

Orycteropus depéreti from Roussillon is much larger than the Pontian 0. gaudryi and differs from it in other ways. Meles gennevauxi and Plesiogulo monspessulanus (see Viret 1939) are Asiatic migrants, the former related to Meles taxipater and the latter to Plesiogulo brachygnathus but of a more progressive type approaching Gulo. Meles taxipater is a Chinese Pontian form and Plesiogulo brachygnathus is found both in the Pontian of China and the Dhok Pathan of India.

On the other hand numerous species from Roussillon are identical or nearly so with Villafranchian forms. Such are : Hyuna arvernensis, Felis brevirostris, Megantereon megantereon, Epimachairodus crenatidens, Mastodon (Anancus) arvernensis, Tapirus arvernensis, Dicerorhinus leptorhinus of Roussillon occurs at Perrier. In the Val d'Arno it is replaced by D. etruscus. Cervus ramosus occurs at both hori

A species of Gazella from Perpignan has been referred to G. borbonica from Perrier (Roccaneyra) and the Red Crag.

Types which make their first appearance in Europe at Roussillon are : Alopecoid Canidæ, as Vulpes donnezani represented in the Villafranchian by the different species Canis megamastoides ; Ursus, as a variety ruscinensis of the wide spread Villafranchian species U. (Helarctos) arvernensis ; Cervus (Capreolus) is known by the species C. australis which is closely allied to C. cusanus from Perrier ; Potamochoerus, as P. provincialis, of which a variety, minor occurs both


« PreviousContinue »