ice invasions came from centers of dispersion that migrated across the continent in chronological progression, the elevation hypothesis has not been so fashionable. The recognition of the fact that the Permian and Cambrian glaciations were in part, at least, at sea level has destroyed the elevation hypothesis as a general explanation.

A third hypothesis ascribes variations in climate to astronomical causes. It is based on the fact that the axis of the earth continuously changes its angular relation to the plane of the earth's orbit, just as the axis of a top sways as it spins. This swaying of the axis of the spinning earth, as it revolves about the sun, is periodic, and causes the northern hemisphere to receive less heat and the southern more heat, and vice versa, at regularly recurring intervals of 21,000 years. The total receipt of heat, however, is constant. The objection to this hypothesis is that, if in itself the variation in the obliquity of the ecliptic be adequate to induce glaciation, we should have had several thousand glaciations in the northern hemisphere alternating with as many in the southern hemisphere in the course of geological time, which is of course not so.

But it is urged this periodically most favorable relation of the earth's axis to the plane of the earth's orbit may be combined with maximum eccentricity of the latter, which occurs at larger intervals, and the combination would intensify the conditions conducive to glaciation. It is pointed out, further, that these astronomical conditions would not only affect the obliquity of the ecliptic so as to distribute unequally the sun's heat between the northern and southern hemispheres, but that this migration of the thermal equator would also greatly modify the trend of oceanic currents, particularly in the Atlantic, so that, at the very time that the northern hemisphere receives least heat from the sun, the warm equatorial currents would be deflected toward the Antarctic, thus

intensifying the unequal distribution of heat. But even under this form of the hypothesis there should have been many more glaciations than we have any trace of in geological history. We may, therefore, admit the tendency to alternations of refrigeration and warmth in the two hemispheres, but doubt seriously whether the tendency has been effective for the production of glacial climates.

Two hypotheses are based on the possible variation in the composition of the atmosphere, particularly in respect to its content of carbon dioxide and the vapor of water. It is a well-known fact that both of these gases play the rôle of a thermal screen in the lower strata of the atmosphere. The heat reaching the earth from the sun tends to escape by radiation but is absorbed by the carbon dioxide and vapor of water. The proportion of carbon dioxide in the atmosphere is very smallonly about .03 per cent, but its importance in determining the habitability of our planet is large. The extent to which it traps the sun's heat and controls the temperature of the earth's lower atmosphere may be gathered from the fact that, according to Arrhenius, if the content of carbon dioxide were increased to .09 per cent, the amelioration of Arctic climate would be equivalent to a rise of the mean annual temperature of from 12° to 16° F; while, if the amount now in the atmosphere were reduced by half, the climate of the temperate zones would become glacial. It becomes a matter of interest, therefore, to inquire for a moment into the possible causes of the variation of the amount of this constituent in the atmosphere. First, as to depletion. It is well known. that the vast deposits of coal which exist at various geological horizons in the strata of the earth's crust are due to accumulations of vegetable matter and that this vegetation has derived all its carbon from the carbon dioxide of the atmosphere. The formation of coal is therefore a process which tends to deplete the air of

carbon dioxide. Again, limestone contains 44 per cent by weight of carbon dioxide. This comes directly from the sea water, very commonly through the agency of organisms, such as corals, crioids, foraminifera, etc., but is derived ultimately, so far as we know, from the atmosphere, since all carbonates in the sea are the product of the decomposition of rocks exposed to the atmosphere, and the carbonatization of the lime, magnesia, etc., which they contain. Now, the limestones known to exist on the continental areas are many times more extensive than the coal seams and many times thicker, and besides these there are very extensive limestone deposits on the floors of the oceans of unknown but doubtless great thickness. It is thus certain that the formation of limestone tends to deplete the air of its carbon dioxide far more effectively than the formation of coal. It is therefore not unreasonable to suppose that after periods characterized by the deposition of coal, such as the Carboniferous period, and a fortiori after the great limestone forming periods, the atmosphere would contain less carbon dioxide than at the beginning of those periods; and that consequently it would be less competent to retain the sun's heat in its lower layers. The result of this would be, of course, a reduction of the mean annual temperature over the entire globe. There can be no doubt that this tendency is operative, but its quantitative effect is uncertain. The effect is certainly not nearly so large as might be inferred from any estimates of the amount of carbon dioxide that has been fixed in limestone and coal; for there are other important factors in the problem. For example, there is about fourteen times as much carbon dioxide in the sea as there is in the air, and the amounts in the air and in the sea are always in balance. It follows from this that whatever quantity of carbon dioxide is removed from the atmosphere and fixed in coal or limestone fourteen-fifteenths of this is made good from the ocean. Or to state the case another way, if one

part of the carbonic dioxide in the sea water be fixed as limestone only one-fifteenth of this amount will be withdrawn from the air to make good the balance, or maintain the ratio of fourteen to one.

Then we must consider the replenishment of the carbon dioxide which may be thus removed from the air. Volcanoes belch forth vast quantities of gases, and among these carbon dioxide is perhaps the most abundant. In certain epochs of the earth's history volcanoes have been far more numerous than they are today, and far more active. We might reasonably suppose that, after such an epoch, the content of carbon dioxide in the air would be augmented, and that the mean annual temperature would consequently be higher than before. But most of this added carbon dioxide must enter into solution in the sea in order to maintain the balance. Here again the tendency to climatic amelioration can scarcely be doubted, but the quantitative effect is entirely unknown.

The hypothesis that the fluctuation of the content of carbon dioxide in the air is an adequate cause of glaciation and deglaciation is a most attractive one. It involves the notion of climatic change affecting the whole globe. This quality of universality, however, affords a means of testing its validity. We can readily understand that, after the deposition of the great limestone and coal formations in the Carboniferous, the air might have been sufficiently depleted of its carbon dioxide to produce the refrigeration necessary for glaciation on the rim of the Indian Ocean in Permian time; but such depletion does not explain the absence of glaciation and the prevalence of aridity during the same period in North America and Europe. Moreover, the Tertiary, which immediately preceded the Pleistocene, was a period of widespread volcanic activity, so that the air of Pleistocene time should have been enriched in carbon dioxide. In spite of this enrichment

we had extensive glaciation. Again it is difficult by this hypothesis to explain the relatively rapid oscillation of conditions within the glacial periods, as indicated by alternations of glaciation and deglaciation. In the face of these and similar criticisms the hypothesis fails of winning general acceptance.

Another hypothesis explains glaciation and deglaciation by a variation in the amount of heat received by the earth from the sun. It is based on the fact established by the observations of Langley and Abbott that the heat of the sun actually varies in short periods by as much as ten per cent. It is argued that, if variation in the sun's heat is a fact, the limits of that variation being unknown, there is nothing to preclude the possibility that there may have been notable variation in geological time. The universality of the temperature changes involved in this suggestion again affords the means of checking its value. The distribution of glacial conditions in Permian time does not accord with the notion of a general lowering of the temperature of the earth's surface, since under this condition the high latitudes would be more likely to be glaciated than the lower, and this expectation is not realized in the geological record.

These brief references to supposed causes of climatic variation show that none of them is entirely satisfactory. Those who argue in favor of any one of these hypotheses can of course make out a very much better case for them than I have been able to do in the space at my disposal. But those who criticize them adversely are also more trenchant and convincing than I have been in my attempt at a very condensed summary. In general I can only relate briefly the fact that, while the several hypotheses put forward have all had a sympathetic and interested reception, and while some of them have been distinctly popular for a time, none of them appears to be able to withstand the skeptical criticism that new