In this vast field of snow the glacial ice slowly crept outward, possibly at an average speed of half a foot a day, but varying from almost nothing in winter at the north, to several feet a day in summer at the south.[44] The force which caused the movement was the presence of the ice piled up not far from the margins. Almost certainly, however, there was no great dome from the center in Canada outward, as some early writers assumed. Such a dome would require that the ice be many thousands of feet thick near its center. This is impossible because of the fact that ice is more voluminous than water (about 9 per cent near the freezing point). Hence when subjected to sufficient pressure it changes to the liquid form. As friction and internal heat tend to keep the bottom of a glacier warm, even in cold regions, the probabilities are that only under very special conditions was a continental ice sheet much thicker than about 2500 feet. In Antarctica, where the temperature is much lower than was probably attained in the United States, the ice sheet is nearly level, several expeditions having traveled hundreds of miles with practically no change in altitude. In Shackleton's trip almost to the South Pole, he encountered a general rise of 3000 feet in 1200 miles. Mountains, however, projected through the ice even near the pole and the geologists conclude that the ice is not very thick even at the world's coldest point, the South Pole. Along the margin of the ice there were two sorts of movement, much more rapid than the slow creep of the ice. One was produced by the outward drift of snow carried by the outblowing dry winds and the other and more important was due to the passage of cyclonic storms. Along the border of the ice sheet, except at the north, storm presumably closely followed storm. Their movement, we judge, was relatively slow until near the southern end of the Mississippi lobe, but when this point was passed they moved much more rapidly, for then they could go toward instead of away from the far northern path which the sun prescribes when solar activity is great. The storms brought much snow to the icefield, perhaps sometimes in favored places as much as the hundred feet a year which is recorded for some winters in the Sierras at present. Even the unglaciated intermontane Great Basin presumably received considerable precipitation, perhaps twice as much as its present scanty supply. The rainfall was enough to support many lakes, one of which was ten times as large as Great Salt Lake; and grass was doubtless abundant upon many slopes which are now dry and barren. The relatively heavy precipitation in the Great Basin was probably due primarily to the increased number of storms, but may also have been much influenced by their slow eastward movement. The lows presumably moved slowly in that general region not only because they were retarded and turned from their normal path by the