scatter the short waves of light and thus interfere with their reaching the earth. Abbot and Fowle,[113] 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[114] 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