« PreviousContinue »
ARE THE EQUIDAE RELIABLE FOR THE CORRELATION OF THE SIWALIKS WITH THE COENOZOIC STAGES OF NORTH AMERICA? BY GUY E. PILGRIM, D.Sc., F.G.S., WITH AN APPENDIX BY A. T. HOPWOOD, D.Sc., F.L.S.
During the last decade views have become prevalent in America on the correlation of the various Siwalik stages with the mammal horizons of Europe and North America, which are wholly at variance with those which I have expressed since 1913 and which with a few modifications I still hold. These modifications may be seen if my paper of 1913 be compared with the correlation which I published in 1934 and are more succinctly expressed in a diagram published by Lewis in 1937 (p. 194). A further modification will be necessitated if De Terra's and Teilhard's work in India, involving the shifting of the Siwalik Boulder Conglomerate from the Lower into the Middle Pleistocene, can be substantiated.
These American views were first published in a paper by Matthew in 1929. They were reiterated and somewhat elaborated by Colbert in a short paper of 1935 and again in a larger work (1935, pp. 21-27). During 1936 De Terra published brief accounts of the work of an expedition sent by Yale University to Kashmir and north west India, and a fuller account of that portion of it which bears on the correlation of the Upper Siwalik and Later Pleistocene was published in 1936 under the associated names of De Terra and Teilhard de Chardin. Still later, Lewis (1937, pp. 191-204) published a correlation of the Siwaliks based partly on his own work and partly on that of
( 437 )
the De Terra expedition. In certain respects this differs less than that of Matthew and Colbert from my own correlation, but it is still based, as in their case, on the first appearance of the genera Hipparion and Equus in North America, which all these authorities regard as their earliest appearance anywhere in the world and as affording evidence of age which is irrefutible.
I have already briefly set forth some arguments in support of my own view and have suggested that a dual origin for Hipparion would explain the inconsistency which Colbert has so well indicated. Teilhard and Stirton (1934, p. 282) have also indicated their opinion that such a phenomenon might be possible, and their correlation of the Chinji and Dhok Pathan almost follows my own. The position to which they assign the Val D'Arno and Perrier faunas, in the Astian rather than the Villafranchian is, however, not easy to understand. As remarked above, the first author appears to have altered his opinion since his visit to India.
In view of the accumulation during the last few years of fresh facts which concern the Bovidae and which have only just been published, and on account of the scattered nature of certain faunal evidence which has caused it to escape the notice of at any rate some of my critics, I think it will be advantageous to review the whole question again in much greater detail.
A knowledge of the distribution of the fossil species throughout the various Siwalik stages is essential to the understanding of the evidence by which one must be guided in any attempt to correlate the faunas with those of other parts of the world. For this purpose Colbert's excellent faunal lists (1935, pp. 29-36) will serve for all the groups except the Bovidae, which was in the course of revision. when Colbert's paper was written. The table of the geological distribution of the fossil Bovidae in India, which occupies pages 5-7 of my 1939 memoir on the Fossil Bovidae of India will supply the omission.
I propose first to criticize De Terra's and Lewis' papers mentioned above and afterwards to consider how far Matthew's and Colbert's theory fits the facts. This theory, so far as I can gather, is accepted in a greater or less degree by the other three authors.
The most important results of De Terra's work, so far as it affects the correlation of the Siwaliks are:
1. The recognition of a complete glacial cycle in Kashmir, which corresponds closely to what is found in Europe.
2. The correlation of the Karewa beds, in which he has found fossils, with the events of this cycle.
3. The identification of the Siwalik Boulder Conglomerate with the second glacial period, and the existence of a great unconformity between the Siwalik Boulder Conglomerate and the Pinjor stage.
In connection with the third of these, De Terra's sections in figures 7 and 8, page 804, near Jammu and along the Tawi river are crucial to his case. One would feel more satisfied about the correctness of his identification of these and other horizontally bedded exposures of boulder conglomerate with the well-known Boulder Conglomerate of the Siwaliks if they contained determinable fossils. Unfortunately this appears not be the case. Fossils have, however, been found in the Siwalik Boulder Conglomerate in the Siwalik Hills by the Geological Survey of India as well as by Dr. Barnum Brown, so that we may hope that this lack may be supplied later. Conglomerates in India are, however, notoriously difficult to identify and one cannot but be struck by the fact that, as the Geological Survey of India have established, throughout the whole Sub-Himalayan region from Jammu to Nepal the Boulder Conglomerate shares completely in the folding which has affected the underlying Pinjor beds and no unconformity is observable. At the same time one cannot deny that there is the possibility of an overlap of the Boulder Conglomerate on to the lower beds of the Siwaliks, when we get away from the hills, and one is bound to accept provisionally the lithological resemblance on which De Terra relies. It seems, however, unwise to adopt the designation of Tawi stage for the Boulder Conglomerate, as Lewis suggests, so long as there remains any room for doubt that in the Tawi section we actually have to do with the true Siwalik Boulder Conglomerate.
De Terra's correlation of the Kashmir deposits appears to be based on the assumption that the glacial cycle began in the world generally with the advent of the Pleistocene. When, therefore, his observations proved to his satisfaction that the Boulder Conglomerate coincided with the second ice advance, it became evident that on the glacial chronology which he has accepted the Boulder Conglomerate must be middle Pleistocene. So far, if we are willing to admit that he was right in identifying the boulder conglomerate in contact with a ground moraine of the second stage with the