For all these reasons, so long as storminess remained great, the Pleistocene snowfields, according to the solar hypothesis, must have deepened and expanded. In due time some of the snow was converted into glacial ice. When that occurred, the growth of the snowfield as well as of the ice cap must have been accelerated by glacial movement. Under such circumstances, as the ice crowded southward toward the source of the moisture by which it grew, the area of high pressure produced by its low temperature would expand. This would force the storm track southward in spite of the contrary tendency due to the sun. When the ice sheet had become very extensive, the track would be crowded relatively near to the northern margin of the trade-wind belt. Indeed, the Pleistocene ice sheets, at the time of their maximum extension, reached almost as far south as the latitude now marking the northern limit of the trade-wind belt in summer. As the storm track with its frequent low pressure and the subtropical belt with its high pressure were forced nearer and nearer together, the barometric gradient between the two presumably became greater, winds became stronger, and the storms more intense. This zonal crowding would be of special importance in summer, at which time it would also be most pronounced. In the first place, the storms would be crowded far upon the ice cap which would then be protected from the sun by a cover of fog and cloud more fully than at any other season. Furthermore, the close approach of the trade-wind belt to the storm belt would result in a great increase in the amount of moisture drawn from the belt of evaporation which the trade winds dominate. In the trade-wind belt, clear skies and high temperature make evaporation especially rapid. Indeed, in spite of the vast deserts it is probable that more than three-fourths of the total evaporation now taking place on the earth occurs in the belt of trades, an area which includes about one-half of the earth's surface.