that while the reception and dissipation of heat by any portion of the earth's surface are equal, the physical properties of ice are such that ice-covered regions lose their heat immediately without being sensibly increased in temperature. The effect on temperature of accumulations of ice and snow must, therefore, be enormous. "Were it not for the ice," strikingly remarks Dr. Croll, "the summers of Greenland. . . would be as warm as those of England; but, instead of this, the Greenland summers are colder than our winters. Cover India with an ice-sheet, and its summers would be colder than those of England." Even on our own coasts the grounding of a single berg appreciably lowers the temperature and greatly increases the danger of frosts.
On the hemisphere whose summers occurred in aphelion during the period of high eccentricity, an exactly opposite tendency would be manifested: the snow and ice would gradually melt and perhaps entirely disappear, and vegetation might flourish even under the pole. That hemisphere would then enjoy an interglacial period. These periods occupy an important place in the theory under examination.
Dr. Croll then proceeds to show how the accumulation of ice in polar regions would affect the general oceanic circulation: "Owing to the difference between the temperature of the equator and the poles there is a constant flow of air from the poles to the equator. It is to this that the trade-winds owe their existence. Now, as the strength of these winds, as a general rule, will depend on the difference of temperature that may exist between the equator and higher latitudes, it follows that the trades on the cold hemisphere will be stronger than those on the warm. . . . Suppose, now, the northern hemisphere to be the cold one. The northeast trade-winds of this hemisphere will far exceed in strength the southeast trade-winds of the southern hemisphere. The median-line between the trades will consequently lie to a very considerable distance to the south of the equator. . . . Let us now consider how this would affect the Gulf Stream. The South American Continent is shaped somewhat in the form of a triangle, with one of its angular comers, called Cape St. Roque, pointing eastward. The equatorial current of the Atlantic impinges against this corner, but, as the greater portion of the current lies a little to the north of the corner, it flows westward into the Gulf of Mexico and forms the Gulf Stream. . . . Now, it is perfectly obvious that the shifting of the equatorial current of the Atlantic only a few degrees to the south of its present position—a thing which would certainly take place under the conditions we have been detailing—would turn the entire current" to the south of Cape St. Roque, and thence along the Brazilian shores and into the Southern Ocean, and "the Gulf Stream would consequently be stopped."
Now, it is quite manifest that if the wind theory of oceanic circulation is incorrect—and, as already shown, its correctness has not been
