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,[99] 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.[100] 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,[101] 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[102] when the partially combined carbon dioxide is also considered. Moreover, the capacity of water for carbon dioxide varies sharply with the temperature.[103] 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.[104]

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