CHAPTER XIII

THE CHANGING COMPOSITION OF OCEANS AND ATMOSPHERE

Having discussed the climatic effect of movements of the earth's crust during the course of geological time, we are now ready to consider the corresponding effects due to changes in the movable envelopes--the oceans and the atmosphere. Variations in the composition of sea water and of air and in the amount of air must almost certainly have occurred, and must have produced at least slight climatic consequences. It should be pointed out at once that such variations appear to be far less important climatically than do movements of the earth's crust and changes in the activity of the sun. Moreover, in most cases, they are not reversible as are the crustal and solar phenomena. Hence, while most of them appear to have been unimportant so far as climatic oscillations and fluctuations are concerned, they seemingly have aided in producing the slight secular progression to which we have so often referred.

There is general agreement among geologists that the ocean has become increasingly saline throughout the ages. Indeed, calculations of the rate of accumulation of salt have been a favorite method of arriving at estimates of the age of the ocean, and hence of the earliest marine sediments. So far as known, however, no geologist or climatologist has discussed the probable climatic effects of increased salinity. Yet it seems clear that an increase in salinity must have a slight effect upon climate.

Salinity affects climate in four ways: (1) It appreciably influences the rate of evaporation; (2) it alters the freezing point; (3) it produces certain indirect effects through changes in the absorption of carbon dioxide; and (4) it has an effect on oceanic circulation.

(1) According to the experiments of Mazelle and Okada, as reported by Krümmel,[97] evaporation from ordinary sea water is from 9 to 30 per cent less rapid than from fresh water under similar conditions. The variation from 9 to 30 per cent found in the experiments depends, perhaps, upon the wind velocity. When salt water is stagnant, rapid evaporation tends to result in the development of a film of salt on the top of the water, especially where it is sheltered from the wind. Such a film necessarily reduces evaporation. Hence the relatively low salinity of the oceans in the past probably had a tendency to increase the amount of water vapor in the air. Even a little water vapor augments slightly the blanketing effect of the air and to that extent diminishes the diurnal and seasonal range of temperature and the contrast from zone to zone.

(2) Increased salinity means a lower freezing temperature of the oceans and hence would have an effect during cold periods such as the present and the Pleistocene ice age. It would not, however, be of importance during the long warm periods which form most of geologic time. A salinity of about 3.5 per cent at present lowers the freezing point of the ocean roughly 2°C. below that of fresh water. If the ocean were fresh and our winters as cold as now, all the harbors of New England and the Middle Atlantic States would be icebound. The Baltic Sea would also be frozen each winter, and even the eastern harbors of the British Isles would be frequently locked in ice. At high latitudes the area of permanently frozen oceans would be much enlarged. The effect of such a condition upon marine life in high latitudes would be like that of a change to a warmer climate. It would protect the life on the continental shelf from the severe battering of winter storms. It would also lessen the severity of the winter temperature in the water for when water freezes it gives up much latent heat,--eighty calories per cubic centimeter. Part of this raises the temperature of the underlying water.

The expansion of the ice near northern shores would influence the life of the lands quite differently from that of the oceans. It would act like an addition of land to the continents and would, therefore, increase the atmospheric contrasts from zone to zone and from continental interior to ocean. In summer the ice upon the sea would tend to keep the coastal lands cool, very much as happens now near the Arctic Ocean, where the ice floes have a great effect through their reflection of light and their absorption of heat in melting. In winter the virtual enlargement of the continents by the addition of an ice fringe would decrease the snowfall upon the lands. Still more important would be the effect in intensifying the anti-cyclonic conditions which normally prevail in winter not only over continents but over ice-covered oceans. Hence the outblowing cold winds would he strengthened.[98] The net effect of all these conditions would apparently be a diminution of snowfall in high latitudes upon the lands even though the summer snowfall upon the ocean and the coasts may have increased. This condition may have been one reason why widespread glaciation does not appear to have prevailed in high latitudes during the Proterozoic and Permian glaciations, even though it occurred farther south. If the ocean during those early glacial epochs were ice-covered down to middle latitudes, a lack of extensive glaciation in high latitudes would be no more surprising than is the lack of Pleistocene glaciation in the northern parts of Alaska and Asia. Great ice sheets are impossible without a large supply of moisture.

(3) Among the indirect effects of salinity one of the chief appears to be that the low salinity of the water in the past and the greater ease with which it froze presumably allowed the temperature of the entire ocean to be slightly higher than now. This is because ice serves as a blanket and hinders the radiation of heat from the underlying water. The temperature of the ocean has a climatic significance not only directly, but indirectly through its influence on the amount of carbon dioxide held by the oceans. A change of even 1°C. from the present mean temperature of 2°C. would alter the ability of the entire ocean to absorb carbon dioxide by about 4 per cent. This, according to F. W. Clarke,[99] is because the oceans contain from eighteen to twenty-seven times as much carbon dioxide as the air when only the free carbon dioxide is considered, and about seventy times as much according to Johnson and Williamson[100] when the partially combined carbon dioxide is also considered. Moreover, the capacity of water for carbon dioxide varies sharply with the temperature.[101] Hence a rise in temperature of only 1°C. would theoretically cause the oceans to give up from 30 to 280 times as much carbon dioxide as the air now holds. This, however, is on the unfounded assumption that the oceans are completely saturated. The important point is merely that a slight change in ocean temperature would cause a disproportionately large change in the amount of carbon dioxide in the air with all that this implies in respect to blanketing the earth, and thus altering temperature.

(4) Another and perhaps the most important effect of salinity upon climate depends upon the rapidity of the deep-sea circulation. The circulation is induced by differences of temperature, but its speed is affected at least slightly by salinity. The vertical circulation is now dominated by cold water from subpolar latitudes. Except in closed seas like the Mediterranean the lower portions of the ocean are near the freezing point. This is because cold water sinks in high latitudes by reason of its superior density, and then "creeps" to low latitudes. There it finally rises and replaces either the water driven poleward by the winds, or that which has evaporated from the Surface.[102]

During past ages, when the sea water was less salty, the circulation was presumably more rapid than now. This was because, in tropical regions, the rise of cold water is hindered by the sinking of warm surface water which is relatively dense because evaporation has removed part of the water and caused an accumulation of salt. According to Krümmel and Mill,[103] the surface salinity of the subtropical belt of the North Atlantic commonly exceeds 3.7 per cent and sometimes reaches 3.77 per cent, whereas the underlying waters have a salinity of less than 3.5 per cent and locally as little as 3.44 per cent. The other oceans are slightly less saline than the North Atlantic at all depths, but the vertical salinity gradients along the tropics are similar. According to the Smithsonian Physical Tables, the difference in salinity between the surface water and that lying below is equivalent to a difference of .003 in density, where the density of fresh water is taken as 1.000. Since the decrease in density produced by warming water from the temperature of its greatest density (4°C.) to the highest temperatures which ever prevail in the ocean (30°C. or 86°F.) is only .004, the more saline surface waters of the dry tropics are at most times almost as dense as the less saline but colder waters beneath the surface, which have come from higher latitudes. During days of especially great evaporation, however, the most saline portions of the surface waters in the dry tropics are denser than the underlying waters and therefore sink, and produce a temporary local stagnation in the general circulation. Such a sinking of the warm surface waters is reported by Krümmel, who detected it by means of the rise in temperature which it produces at considerable depths. If such a hindrance to the circulation did not exist, the velocity of the deep-sea movements would be greater.

If in earlier times a more rapid circulation occurred, low latitudes must have been cooled more than now by the rise of cold waters. At the same time higher latitudes were presumably warmed by a greater flow of warm water from tropical regions because less of the surface heat sank in low latitudes. Such conditions would tend to lessen the climatic contrast between the different latitudes. Hence, in so far as the rate of deep-sea circulation depends upon salinity, the slowly increasing amount of salt in the oceans must have tended to increase the contrasts between low and high latitudes. Thus for several reasons, the increase of salinity during geologic history seems to deserve a place among the minor agencies which help to explain the apparent tendency toward a secular progression of climate in the direction of greater contrasts between tropical and subpolar latitudes.

Changes in the composition and amount of the atmosphere have presumably had a climatic importance greater than that of changes in the salinity of the oceans. The atmospheric changes may have been either progressive or cyclic, or both. In early times, according to the nebular hypothesis, the atmosphere was much more dense than now and contained a larger percentage of certain constituents, notably carbon dioxide and water. The planetesimal hypothesis, on the other hand, postulates an increase in the density of the atmosphere, for according to this hypothesis the density of the atmosphere depends upon the power of the earth to hold gases, and this power increases as the earth grows bigger with the infall of material from without.[104]

Whichever hypothesis may be correct, it seems probable that when life first appeared on the land the atmosphere resembled that of today in certain fundamental respects. It contained the elements essential to life, and its blanketing effect was such as to maintain temperatures not greatly different from those of the present. The evidence of this depends largely upon the narrow limits of temperature within which the activities of modern life are possible, and upon the cumulative evidence that ancient life was essentially similar to the types now living. The resemblance between some of the oldest forms and those of today is striking. For example, according to Professor Schuchert:[105] "Many of the living genera of forest trees had their origin in the Cretaceous, and the giant sequoias of California go back to the Triassic, while Ginkgo is known in the Permian. Some of the fresh-water molluscs certainly were living in the early periods of the Mesozoic, and the lung-fish of today (Ceratodus) is known as far back as the Triassic and is not very unlike other lung-fishes of the Devonian. The higher vertebrates and insects, on the other hand, are very sensitive to their environment, and therefore do not extend back generically beyond the Cenozoic, and only in a few instances even as far as the Oligocene. Of marine invertebrates the story is very different, for it is well known that the horseshoe crab (Limulus) lived in the Upper Jurassic, and Nautilus in the Triassic, with forms in the Devonian not far removed from this genus. Still longer-ranging genera occur among the brachiopods, for living Lingula and Crania have specific representatives as far back as the early Ordovician. Among living foraminifers, Lagena, Globigerina, and Nodosaria are known in the later Cambrian or early Ordovician. In the Middle Cambrian near Field, British Columbia, Walcott has found a most varied array of invertebrates among which are crustaceans not far removed from living forms. Zoölogists who see these wonderful fossils are at once struck with their modernity and the little change that has taken place in certain stocks since that far remote time. Back of the Paleozoic, little can be said of life from the generic standpoint, since so few fossils have been recovered, but what is at hand suggests that the marine environment was similar to that of today."

At present, as we have repeatedly seen, little growth takes place either among animals or plants at temperatures below 0°C. or above 40°C., and for most species the limiting temperatures are about 10° and 30°. The maintenance of so narrow a scale of temperature is a function of the atmosphere, as well as of the sun. Without an atmosphere, the temperature by day would mount fatally wherever the sun rides high in the sky. By night it would fall everywhere to a temperature approaching absolute zero, that is -273°C. Some such temperature prevails a few miles above the earth's surface, beyond the effective atmosphere. Indeed, even if the atmosphere were almost as it is now, but only lacked one of the minor constituents, a constituent which is often actually ignored in statements of the composition of the air, life would be impossible. Tyndall concludes that if water vapor were entirely removed from the atmosphere for a single day and night, all life--except that which is dormant in the form of seeds, eggs, or spores--would be exterminated. Part would be killed by the high temperature developed by day when the sun was high, and part, by the cold night.

The testimony of ancient glaciation as to the slight difference in the climate and therefore in the atmosphere of early and late geological times is almost as clear as that of life. Just as life proves that the earth can never have been extremely cold during hundreds of millions of years, so glaciation in moderately low latitudes near the dawn of earth history and at several later times, proves that the earth was not particularly hot even in those early days. The gentle progressive change of climate which is recorded in the rocks appears to have been only in slight measure a change in the mean temperature of the earth as a whole, and almost entirely a change in the distribution of temperature from place to place and season to season. Hence it seems probable that neither the earth's own emission of heat, nor the supply of solar heat, nor the power of the atmosphere to retain heat can have been much greater a few hundred million years ago than now. It is indeed possible that these three factors may have varied in such a way that any variation in one has been offset by variations of the others in the opposite direction. This, however, is so highly improbable that it seems advisable to assume that all three have remained relatively constant. This conclusion together with a realization of the climatic significance of carbon dioxide has forced most of the adherents of the nebular hypothesis to abandon their assumption that carbon dioxide, the heaviest gas in the air, was very abundant until taken out by coal-forming plants or combined with the calcium oxide of igneous rocks to form the limestone secreted by animals. In the same way the presence of sun cracks in sedimentary rocks of all ages suggests that the air cannot have contained vast quantities of water vapor such as have been assumed by Knowlton and others in order to account for the former lack of sharp climatic contrast between the zones. Such a large amount of water vapor would almost certainly be accompanied by well-nigh universal and continual cloudiness so that there would be little chance for the pools on the earth's water-soaked surface to dry up. Furthermore, there is only one way in which such cloudiness could be maintained and that is by keeping the air at an almost constant temperature night and day. This would require that the chief source of warmth be the interior of the earth, a condition which the Proterozoic, Permian, and other widespread glaciations seem to disprove.

Thus there appears to be strong evidence against the radical changes in the atmosphere which are sometimes postulated. Yet some changes must have taken place, and even minor changes would be accompanied by some sort of climatic effect. The changes would take the form of either an increase or a decrease in the atmosphere as a whole, or in its constituent elements. The chief means by which the atmosphere has increased appear to be as follows: (a) By contributions from the interior of the earth via volcanoes and springs and by the weathering of igneous rocks with the consequent release of their enclosed gases;[106] (b) by the escape of some of the abundant gases which the ocean holds in solution; (c) by the arrival on the earth of gases from space, either enclosed in meteors or as free-flying molecules; (d) by the release of gases from organic compounds by oxidation, or by exhalation from animals and plants. On the other hand, one or another of the constituents of the atmosphere has presumably decreased (a) by being locked up in newly formed rocks or organic compounds; (b) by being dissolved in the ocean; (c) by the escape of molecules into space; and (d) by the condensation of water vapor.

The combined effect of the various means of increase and decrease depends partly on the amount of each constituent received from the earth's interior or from space, and partly on the fact that the agencies which tend to deplete the atmosphere are highly selective in their action. Our knowledge of how large a quantity of new gases the air has received is very scanty, but judging by present conditions the general tendency is toward a slow increase chiefly because of meteorites, volcanic action, and the work of deep-seated springs. As to decrease, the case is clearer. This is because the chemically active gases, oxygen, CO_{2}, and water vapor, tend to be locked up in the rocks, while the chemically inert gases, nitrogen and argon, show almost no such tendency. Though oxygen is by far the most abundant element in the earth's crust, making up more than 50 per cent of the total, it forms only about one-fifth of the air. Nitrogen, on the other hand, is very rare in the rocks, but makes up nearly four-fifths of the air. It would, therefore, seem probable that throughout the earth's history, there has been a progressive increase in the amount of atmospheric nitrogen, and presumably a somewhat corresponding increase in the mass of the air. On the other hand, it is not clear what changes have occurred in the amount of atmospheric oxygen. It may have increased somewhat or perhaps even notably. Nevertheless, because of the greater increase in nitrogen, it may form no greater percentage of the air now than in the distant past.

As to the absolute amounts of oxygen, Barrell[107] thought that atmospheric oxygen began to be present only after plants had appeared. It will be recalled that plants absorb carbon dioxide and separate the carbon from the oxygen, using the carbon in their tissues and setting free the oxygen. As evidence of a paucity of oxygen in the air in early Proterozoic times, Barrell cites the fact that the sedimentary rocks of that remote time commonly are somewhat greyish or greenish-grey wackes, or other types, indicating incomplete oxidation. He admits, however, that the stupendous thicknesses of red sandstones, quartzite, and hematitic iron ores of the later Proterozoic prove that by that date there was an abundance of atmospheric oxygen. If so, the change from paucity to abundance must have occurred before fossils were numerous enough to give much clue to climate. However, Barrell's evidence as to a former paucity of atmospheric oxygen is not altogether convincing. In the first place, it does not seem justifiable to assume that there could be no oxygen until plants appeared to break down the carbon dioxide, for some oxygen is contributed by volcanoes,[108] and lightning decomposes water into its elements. Part of the hydrogen thus set free escapes into space, for the earth's gravitative force does not appear great enough to hold this lightest of gases, but the oxygen remains. Thus electrolysis of water results in the accumulation of oxygen. In the second place, there is no proof that the ancient greywackes are not deoxidized sediments. Light colored rock formations do not necessarily indicate a paucity of atmospheric oxygen, for such rocks are abundant even in recent times. For example, the Tertiary formations are characteristically light colored, a result, however, of deoxidation. Finally, the fact that sedimentary rocks, irrespective of their age, contain an average of about 1.5 per cent more oxygen than do igneous rocks,[109] suggests that oxygen was present in the air in quantity even when the earliest shales and sandstones were formed, for atmospheric oxygen seems to be the probable source of the extra oxygen they contain. The formation of these particular sedimentary rocks by weathering of igneous rocks involves only a little carbon dioxide and water. Although it seems probable that oxygen was present in the atmosphere even at the beginning of the geological record, it may have been far less abundant then than now. It may have been removed from the atmosphere by animals or by the oxidation of the rocks almost as rapidly as it was added by volcanoes, plants, and other agencies.

After this chapter was in type, St. John[C] announced his interesting discovery that oxygen is apparently lacking in the atmosphere of Venus. He considers that this proves that Venus has no life. Furthermore he concludes that so active an element as oxygen cannot be abundant in the atmosphere of a planet unless plants continually supply large quantities by breaking down carbon dioxide.

But even if the earth has experienced a notable increase in atmospheric oxygen since the appearance of life, this does not necessarily involve important climatic changes except those due to increased atmospheric density. This is because oxygen has very little effect upon the passage of light or heat, being transparent to all but a few wave lengths. Those absorbed are chiefly in the ultra violet.

The distinct possibility that oxygen has increased in amount, makes it the more likely that there has been an increase in the total atmosphere, for the oxygen would supplement the increase in the relatively inert nitrogen and argon, which has presumably taken place. The climatic effects of an increase in the atmosphere include, in the first place, an increased scattering of light as it approaches the earth. Nitrogen, argon, and oxygen all scatter the short waves of light and thus interfere with their reaching the earth. Abbot and Fowle,[110] who have carefully studied the matter, believe that at present the scattering is quantitatively important in lessening insolation. Hence our supposed general increase in the volume of the air during part of geological times would tend to reduce the amount of solar energy reaching the earth's surface. On the other hand, nitrogen and argon do not appear to absorb the long wave lengths known as heat, and oxygen absorbs so little as to be almost a non-absorber. Therefore the reduced penetration of the air by solar radiation due to the scattering of light would apparently not be neutralized by any direct increase in the blanketing effect of the atmosphere, and the temperature near the earth's surface would be slightly lowered by a thicker atmosphere. This would diminish the amount of water vapor which would be held in the air, and thereby lower the temperature a trifle more.

In the second place, the higher atmospheric pressure which would result from the addition of gases to the air would cause a lessening of the rate of evaporation, for that rate declines as pressure increases. Decreased evaporation would presumably still further diminish the vapor content of the atmosphere. This would mean a greater daily and seasonal range of temperature, as is very obvious when we compare clear weather with cloudy. Cloudy nights are relatively warm while clear nights are cool, because water vapor is an almost perfect absorber of radiant heat, and there is enough of it in the air on moist nights to interfere greatly with the escape of the heat accumulated during the day. Therefore, if atmospheric moisture were formerly much more abundant than now, the temperature must have been much more uniform. The tendency toward climatic severity as time went on would be still further increased by the cooling which would result from the increased wind velocity discussed below; for cooling by convection increases with the velocity of the wind, as does cooling by conduction.

Any persistent lowering of the general temperature of the air would affect not only its ability to hold water vapor, but would produce a lessening in the amount of atmospheric carbon dioxide, for the colder the ocean becomes the more carbon dioxide it can hold in solution. When the oceanic temperature falls, part of the atmospheric carbon dioxide is dissolved in the ocean. This minor constituent of the air is important because although it forms only 0.003 per cent of the earth's atmosphere, Abbot and Fowle's[111] calculations indicate that it absorbs over 10 per cent of the heat radiated outward from the earth. Hence variations in the amount of carbon dioxide may have caused an appreciable variation in temperature and thus in other climatic conditions. Humphreys, as we have seen, has calculated that a doubling of the carbon dioxide in the air would directly raise the earth's temperature to the extent of 1.3°C., and a halving would lower it a like amount. The indirect results of such an increase or decrease might be greater than the direct results, for the change in temperature due to variations in carbon dioxide would alter the capacity of the air to hold moisture.

Two conditions would especially help in this respect; first, changes in nocturnal cooling, and second, changes in local convection. The presence of carbon dioxide diminishes nocturnal cooling because it absorbs the heat radiated by the earth, and re-radiates part of it back again. Hence with increased carbon dioxide and with the consequent warmer nights there would be less nocturnal condensation of water vapor to form dew and frost. Local convection is influenced by carbon dioxide because this gas lessens the temperature gradient. In general, the less the gradient, that is, the less the contrast between the temperature at the surface and higher up, the less convection takes place. This is illustrated by the seasonal variation in convection. In summer, when the gradient is steepest, convection reaches its maximum. It will be recalled that when air rises it is cooled by expansion, and if it ascends far the moisture is soon condensed and precipitated. Indeed, local convection is considered by C. P. Day to be the chief agency which keeps the lower air from being continually saturated with moisture. The presence of carbon dioxide lessens convection because it increases the absorption of heat in the zone above the level in which water vapor is abundant, thus warming these higher layers. The lower air may not be warmed correspondingly by an increase in carbon dioxide if Abbot and Fowle are right in stating that near the earth's surface there is enough water vapor to absorb practically all the wave lengths which carbon dioxide is capable of absorbing. Hence carbon dioxide is chiefly effective at heights to which the low temperature prevents water vapor from ascending. Carbon dioxide is also effective in cold winters and in high latitudes when even the lower air is too cold to contain much water vapor. Moreover, carbon dioxide, by altering the amount of atmospheric water vapor, exerts an indirect as well as a direct effect upon temperature.

Other effects of the increase in air pressure which we are here assuming during at least the early part of geological times are corresponding changes in barometric contrasts, in the strength of winds, and in the mass of air carried by the winds along the earth's surface. The increase in the mass of the air would reënforce the greater velocity of the winds in their action as eroding and transporting agencies. Because of the greater weight of the air, the winds would be capable of picking up more dust and of carrying it farther and higher; while the increased atmospheric friction would keep it aloft a longer time. The significance of dust at high levels and its relation to solar radiation have already been discussed in connection with volcanoes. It will be recalled that on the average it lowers the surface temperature. At lower levels, since dust absorbs heat quickly and gives it out quickly, its presence raises the temperature of the air by day and lowers it by night. Hence an increase in dustiness tends toward greater extremes.

From all these considerations it appears that if the atmosphere has actually evolved according to the supposition which is here tentatively entertained, the general tendency of the resultant climatic changes must have been partly toward long geological oscillations and partly toward a general though very slight increase in climatic severity and in the contrasts between the zones. This seems to agree with the geological record, although the fact that we are living in an age of relative climatic severity may lead us astray.

The significant fact about the whole matter is that the three great types of terrestrial agencies, namely, those of the earth's interior, those of the oceans, and those of the air, all seem to have suffered changes which lead to slow variations of climate. Many reversals have doubtless taken place, and the geologic oscillations thus induced are presumably of much greater importance than the progressive change, yet so far as we can tell the purely terrestrial changes throughout the hundreds of millions of years of geological time have tended toward complexity and toward increased contrasts from continent to ocean, from latitude to latitude, from season to season, and from day to night.

Throughout geological history the slow and almost imperceptible differentiation of the earth's surface has been one of the most noteworthy of all changes. It has been opposed by the extraordinary conservatism of the universe which causes the average temperature today to be so like that of hundreds of millions of years ago that many types of life are almost identical. Nevertheless, the differentiation has gone on. Often, to be sure, it has presumably been completely masked by the disturbances of the solar atmosphere which appear to have been the cause of the sharper, shorter climatic pulsations. But regardless of cosmic conservatism and of solar impulses toward change, the slow differentiation of the earth's surface has apparently given to the world of today much of the geographical complexity which is so stimulating a factor in organic evolution. Such complexity--such diversity from place to place--appears to be largely accounted for by purely terrestrial causes. It may be regarded as the great terrestrial contribution to the climatic environment which guides the development of life.

FOOTNOTES:

[Footnote 97: Encyclopædia Britannica, 11th edition: article "Ocean."]

[Footnote 98: C. E. P. Brooks: The Meteorological Conditions of an Ice sheet and Their Bearing on the Desiccation of the Globe; Quart. Jour. Royal Meteorol. Soc., Vol. 40, 1914, pp. 53-70.]

[Footnote 99: Data of Geochemistry, Fourth Ed., 1920; Bull. No. 695, U. S. Geol. Survey.]

[Footnote 100: Quoted by Schuchert in The Evolution of the Earth.]

[Footnote 101: Smithsonian Physical Tables, Sixth Revision, 1914, p. 142.]

[Footnote 102: Chamberlin, in a very suggestive article "On a possible reversal of oceanic circulation" (Jour. of Geol., Vol. 14, pp. 363-373, 1906), discusses the probable climatic consequences of a reversal in the direction of deep-sea circulation. It is not wholly beyond the bounds of possibility that, in the course of ages the increasing drainage of salt from the lands not only by nature but by man's activities in agriculture and drainage, may ultimately cause such a reversal by increasing the ocean's salinity until the more saline tropical portion is heavier than the cooler but fresher subpolar waters. If that should happen, Greenland, Antarctica, and the northern shores of America and Asia would be warmed by the tropical heat which had been transferred poleward beneath the surface of the ocean, without loss _en route_. Subpolar regions, under such a condition of reversed deep-sea circulation, might have a mild climate. Indeed, they might be among the world's most favorable regions climatically.]

[Footnote 103: Encyclopædia Britannica: article "Ocean."]

[Footnote 104: Chamberlin and Salisbury: Geology, Vol. II, pp. 1-132, 1906; and T. C. Chamberlin: The Origin of the Earth, 1916.]

[Footnote 105: Personal communication.]

[Footnote 106: R. T. Chamberlin: Gases in Rocks, Carnegie Inst. of Wash., No. 106, 1908.]

[Footnote 107: J. Barrell: The Origin of the Earth, in Evolution of the Earth and Its Inhabitants, 1918, p. 44, and more fully in an unpublished manuscript.]

[Footnote 108: F. W. Clarke: Data of Geochemistry, Fourth Ed., 1920, Bull. No. 695, U. S. Geol. Survey, p. 256.]

[Footnote 109: F. W. Clarke: _loc. cit._, pp. 27-34 et al.]

[Footnote C: Chas. E. St. John: Science Service Press Reports from the Mt. Wilson Observatory, May, 1922.]

[Footnote 110: Abbot and Fowle: Annals Astrophysical Observatory; Smiths. Inst., Vol. II, 1908, p. 163.

F. E. Fowle: Atmospheric Scattering of Light; Misc. Coll. Smiths. Inst., Vol. 69, 1918.]

[Footnote 111: Abbot and Fowle: _loc. cit._, p. 172.]