And here again there is evidence that the deposits were laid down near sea level and that in part they were submarine. In peninsular India Permian glacial deposits are found in Lat. 18° to 19° N and again in Lat. 25° N; and in the Himalaya and Tibet as far as Lat. 36°. In South America, glacial deposits of Permian age are found in Brazil in Lat. 28° 30' S interstratified with formations containing coal seams and Permo-Carboniferous plants. We thus have abundant evidence of an extraordinarily widespread glaciation in Permian time that was not circumpolar, was common to the two hemispheres and was characterized by fluctuations similar to those of the Pleistocene. There can, therefore, be no doubt that the climatic conditions of our planet in the Permian period were different from those of the periods which immediately preceded and succeeded it; and, if we may judge from the extent and distribution of the glacial deposits, the changes involved both in the oncoming of glaciation and the recovery from it, were even more profound than those connected with the glaciation of Pleistocene time.

If now we carry our inquiry still further back into the remote past we find evidences of another well defined glacial period in Cambrian time. In the Lower Cambrian rocks of Norway in Lat. 70° 8′ N a boulder till occurs resting on a glaciated pavement. In China in the valley of the Yangtze about Lat. 31° N, there is a boulder till of Lower Cambrian age 170 feet thick, the boulders in which show glacial polish and striations. In South Australia a Lower Cambrian boulder till occupies an area extending from Lat. 30° to 35° S and from Long. 137° to 140° E. Here the glaciated boulders occur in three formations interstratified with marine limestones, so that it is clear that there were fluctuations in glacial conditions at this early period similar to the climatic fluctuations of the Pleistocene.

Still further back in the pre-Cambrian formations of south-central Canada there occur rocks which cannot

be distinguished from morainic deposits and in which there are glacially scratched boulders. The extent of this glaciation is as yet unknown, but the evidence, such as it is, is significant of a climatic change similar to that which brought on the Cambrian, Permian, and Pleistocene glaciations. Evidences of pre-Cambrian glaciation are also reported from Spitzbergen and South Africa.

Here it may be well to point out that we are not wholly dependent upon evidences of glaciation for our notions of climatic change in the past. There are other indications equally significant. To illustrate these briefly we may consider the Caspian Sea. This sea receives the waters of the largest river in Europe and the drainage of a large part of the lofty and snowy range of the Caucasus. Yet its surface is 82 feet below sea level and it is more salty than the ocean. The east side of the Caspian is more arid than the west and evaporation is more active. It results from this that a steady current from the open Caspian flows through a narrow strait into the Gulf of Kara Bugaz to make good the excessive evaporation in that part of the basin. This current brings into Kara Bugaz vast quantities of salts, which are accumulating as a chemical deposit on the bottom of the gulf.

These general conditions are determined by the extreme aridity of the climate of the region in which the Caspian basin is situated. If, therefore, we find among the formations of the earth's crust deposits of salts which can only be explained as due to conditions similar to those which now prevail in the Caspian basin, we may safely infer that the climate was arid in the region where they occur at the time of their deposition. For example, in the Silurian rocks of New York state and the adjoining region there are found at a certain horizon vast deposits of salt accompanied by beds of gypsum, which are taken to signify aridity of climate during that portion of Silurian time. Again in the formations that were laid down in the period of transition

from the Paleozoic to the Mesozoic, including parts of the Permian and the Triassic, there are extensive deposits of red sandstones and shales of continental origin. That is to say, they were not deposited on the floor of the ocean, but on the surface of low lying continental areas above sea level. These "red beds" are widely distributed over Europe and North America and comprise numerous extensive deposits of salt and gypsum. It is most probable that these red beds and their associated chemical precipitates signify a climate of great aridity which was common to Europe and North America during this period of time. It is interesting to observe that this period of aridity on both sides of the north Atlantic was, in its earlier stages at least, contemporaneous with the glaciation of the continents which enclose the Indian Ocean.

Still other means may be used to trace out the fluctuations of climate in time, such as the luxuriance or impoverishment of marine faunas as indicators of the temperature of the sea water; but the consideration of these would carry us far beyond the space available. If I have said sufficient to make clear the general fact that the climates of the past exhibit something of the fickleness of the weather of today, I will go on to say something of the hypotheses that have been advanced to explain the changes we are considering. But before entering upon this phase of the subject, it may be well to remark that the general lesson of the geological record in regard to climate is that there has been a normal climatic condition which appears to have been not very different from that of today. The glaciations in particular regions at particular periods have been departures from the normal; and there has in all cases been a recovery to the normal condition. The periods of abnormal climate have been widely spaced and the periods themselves have been relatively short. The great bulk of the sedimentary rocks which enter into the makeup of the earth's crust, representing the products of the

waste of the continental masses, have originated and have been transported and deposited under climatic conditions very similar to those of the present. The formations indicative of aridity or of glaciation are but a small percentage of the total volume of the stratified column. There is certainly no suggestion in the geological record of a general refrigeration of the planet, such as some have supposed to be our doom.

In considering the possible causes of climatic variation, and particularly in determining the most probable cause, it is well to bear in mind that there are three classes of hypotheses which have been formulated to explain the phenomena. In the first class the controlling idea is that the glacial periods are but extreme phases of a fluctuation which has been going on all through geological history as a wave-like departure on either side of a fairly constant normal mean, as a result either of astronomical influences or of rearrangements of land and sea affecting the circulation of the atmosphere. These hypotheses explain glaciation as the effect of a mere redistribution of climate and reject the notion that there has been any temporary increase or decrease of the total amount of energy received from the sun.

The hypotheses of the second class are based on the idea that, while the total amount of energy received from the sun may have been constant, the constitution of the atmosphere itself has been subject to change, and that this change has affected the climate of the globe as a whole.

The third class of hypotheses has to do with the assumption that the heat from the sun has varied and so caused radical but temporary changes in the earth's climate.

The most naïve hypothesis proposed to explain the glaciations of the past, and particularly the Permian glaciation, which seems to have had but little relation to polar regions as they now exist, is that the axis of rotation of the earth has not been constant in position.

It has been supposed that the south pole in Permian time emerged in the middle of the Indian Ocean, so that the glaciation of Australia, Tasmania, India, and South Africa was in reality roughly circumpolar, just as that of North America and Europe was in Pleistocene time. This would of course effect a redistribution of climate without altering the total receipt of heat from the sun, and would be a fairly satisfactory explanation of the conditions prevailing on the rim of the Indian Ocean in Permian time. But unfortunately the emergence of a pole in the middle of the Indian Ocean necessitates another in northern Mexico, and the Permian rocks of that region show no trace of glaciation. Those who advocate this hypothesis, make no attempt, moreover, to smooth out the difficulties in celestial mechanics which flow from it, or, if they do, they have convinced very few people of their success.

Another hypothesis is that glaciations have been induced by continental uplifts and the subsequent deglaciation by subsidence. This notion found considerable favor some years ago when the Pleistocene glaciation of North America and Europe was the only one known to geologists, and when this was supposed to have been caused by a single continental ice cap. It is based on the fact that glacial conditions are characteristic of high altitudes in any latitude, and on the fact that in pre-Pleistocene time portions of the North American continent were somewhat higher than at present, as shown by the submergence of the ancient valley of the Hudson River, and by similar physiographic conditions. It is assumed that a moderate elevation of Canada, for example, would induce a winter so severe that the ablation of summer would fail to remove the annual snow; and this, of course, is all that is necessary to bring on glaciation. But since the multiple character of the Pleistocene glaciation has been established, and particularly since it has become known that the different