it has been leached of its lime contents far in excess of the similar leaching of the upper moraine. These three lines of evidence from relative erosion, decay, and leaching all agree that the interglacial interval was much longer than postglacial time. Besides this we find evidence at numerous places of the existence of swamps, lakes, and forests in the interval between successive moraines. One of the most striking instances of this is at the city of Toronto. Here resting on an older moraine are lake beds containing numerous remains of animals and plants. The flora is that of the latitude of southern Ohio at the present time. Resting on these lake beds is another sheet of till or morainic material. It is clear that in the interval between the deposition of the two till sheets the latitude of Toronto had a climate as mild as that of southern Ohio and that this was subsequently displaced by a climate of sufficient severity to produce general glaciation. On the basis of these and similar observations it was established that the glaciation of the North American continent was not a simple event, but that the phenomena represent a sequence of glaciations with large interglacial intervals which, in some cases at least, enjoyed a climate even milder than the present.

While these ideas were being developed, explorations in northern Canada brought to light new observations of great interest. It was found that the western or Cordilleran glaciation was earlier than the glaciation of the central part of Canada, and that the movement of the ice had been outward in all directions from a center of dispersion in the northern Cordillera. It was discovered also that the movement of the ice in the region between the Cordillera and Hudson's Bay had been outward from a center of dispersion in the district of Keewatin; and that similarly the movement of the ice in Labrador had been outward in all directions from the center of that great peninsula. These observations

had the effect of breaking down the hypothesis of continental glaciation and substituting for it the idea of a series of at least three great ice fields each with its own independent center of dispersion. These have become known as the Cordilleran, the Keewatin, and the Labradorean glaciers, and some observations published within late years indicate that there may be a fourth, for which the name Patrician glacier has been suggested.

Now the interesting thing in this connection is that these great glaciers or centers of dispersion are not only discrete geographically, but that they are also separated in time. The successive moraines which have been discriminated on the southern margin of the glaciated area have been correlated with the ice from different centers of dispersion. The lower or older moraines have been shown to have been deposited by the Keewatin glacier, while the later overlying moraines are clearly referable to the Labradorean glacier, and there was probably a succession of invasions from each of these centers of ice dispersion. It has thus become clear that the glaciation of North America is referable neither to a polar ice cap nor a continental ice cap, but to a series of glaciations distributed both geographically and chronologically across the continent from west to east. And this fact must control any speculations we may indulge in as to the climate of the Glacial period. It is worth noting that the great Greenland glacier presents conditions which are analogous to those which must have prevailed in Labrador at the maximum of the Labradorean glacier, and in central Canada at the maximum of the Keewatin glacier; and that Greenland may represent the most easterly and latest of the series of centers of dispersion.

In other parts of the North American continent far beyond the limits of the continental glaciation, we have proofs of climatic fluctuations in Pleistocene time which are no less convincing than the fact of the great ice invasions. In the Great Basin, for example, to the west

of the Wasatch, we have in Great Salt Lake the remnant of what was once a vast inland sea, which in geological literature goes by the name of Lake Bonneville.

The surface of this lake was once nearly 1,000 feet higher than the present Great Salt Lake as is shown by the shore lines distinctly scored on the slopes of the surrounding mountains. At its maximum the area of Lake Bonneville was about 20,000 square miles and its level was determined by an outlet, north of Ogden, to the drainage of the Columbia. It had reached this outlet by the slow rise of its waters in a confined basin, due to the excess of inflow over evaporation. If for purposes of rough estimation we assume that the evaporation is six feet per year, then the annual inflow to Lake Bonneville at its maximum stage, just when it reached the outlet, must have been in excess of a prism of water six feet deep and 20,000 square miles in area.

The area of Great Salt Lake is about 2000 square miles; assuming the same rate of evaporation the inflow cannot exceed a prism of water six feet deep and 2000 square miles in area if the surface of the lake remain constant or if it be lowering in level. It thus appears that the inflow of water to Lake Bonneville at its maximum must have been at least ten times as much as that now flowing into Great Salt Lake. There are of course errors in the assumption which I have made, but these will not affect the general contrast between the climate of Bonneville time and of the present.

But the humid conditions thus indicated by the filling of Bonneville basin to its rim were preceded by a condition of aridity more intense than that of the present. The basin was completely dried out. This in turn was preceded by a still earlier period of humidity when the basin filled with water to within 90 feet of its outlet and then subsided. The evidence of this earlier lake is to be found in its sediments, which are quite distinct from those of the second lacustral epoch, and in the shore

lines which can be distinguished from those of the second lake. The evidence of the intermediate desiccation is found in the fact that the earlier lake deposits are buried by alluvial wash, and in the fact that the second lake was a fresh-water lake, as is shown by the fossils in its sediments; it could not have been fresh unless the earlier lake had been completely desiccated and its precipitated salts buried by alluvial wash beyond the possibility of re-solution.

We thus have in the geological history of the Bonneville basin the record of two humid periods alternating with two arid periods, one of the latter being the present. The period of time to which this record applies is the Pleistocene, and the climatic fluctuations to which it testifies are doubtless the same as those which caused the glaciation and deglaciation of the more northeastern part of the continent.

The succession of glacial and interglacial periods which marks the Pleistocene of North America has its counterpart in Europe, though there is much doubt as to the detailed correlation of these periods on the two sides of the Atlantic.

I may now proceed to point out that the peculiar climatic conditions of the Pleistocene are not unique in geological history, and that similar changes of climate have also occurred in the more remote past. Not many years ago the Pleistocene was the only glacial period known to geologists and it was, therefore, commonly referred to as the Glacial period. Today when we use this term we may be called upon to state which particular glacial period we are speaking of. For there are at least two others which are as well defined and as certain as the Pleistocene, and there are indications of others concerning which our knowledge is as yet very scant.

In the formations which were laid down in the Permian period, in the time transition from the Palaeozoic to the Mesozoic, we find abundant evidence of

glaciation in several parts of the world. This evidence consists not only of morainic accumulations with ice scratched boulders and all the other characteristics of Pleistocene moraines, but also of scored and polished hummocky rock surfaces upon which these moraines rest. In some places there is a series of such morainic formations alternating with ordinary marine sediments containing beds of coal, thus showing incontestably that the moraines were laid down near sea level, and that climates severe enough to produce glaciation alternated with climates which were mild enough to engender the abundant vegetation necessary for the formation of coal. That is to say, there were glacial and interglacial periods in the Permian period just as there were in the Pleisto

cene.

This Permian glaciation was far more extensive than the Pleistocene glaciation and had a most remarkable distribution. In Pleistocene time the glaciation was roughly circumpolar in both the northern and southern hemispheres, a fact which accords with our natural expectations; but in Permian time the glaciation was nearer the equator than the poles and was distributed on both sides of the equator. A brief reference to the principal regions where this Permian glaciation is displayed may be of interest.

In southeastern Australia the glacial deposits extend over a belt extending from Lat. 30° S to 40° S, and are intercalated with marine sediments and coal measures, the lowest moraine resting, however, on well glaciated surfaces. At the same geological horizon glacial deposits are again found in Tasmania so that the known range of glacial conditions is from Lat. 30° S to 42° S. In South Africa moraines of Permian age are spread out over the greater part of the width of the southern part of the continent and extend from Lat. 24° S to 34° S. Here we have the remarkable fact also revealed that the Permian ice invasion was from equatorial regions southerly.