The natives who live near glaciers know that they advance and retreat, but the fact that the whole glacier is really a slowly flowing viscous mass--a sort of frozen but not immobile river--was only established by scientific observation in the last century. The frozen river is fed by the snow which falls on the higher mountain ridges, and is squeezed into the form of ice instead of snow powder by its own weight as it slips down the inclines, warmed by the unclouded sunshine. The big glaciers move much more rapidly (or perhaps one should say less slowly) in the middle than at the sides. The measurements which have been made differ in different glaciers and in different parts of the same glacier, and show smaller movement in winter than in summer. The advance of the sides is retarded, as in the case of an ordinary river of flowing water, by friction against the rocks, which enclose the glacier as its banks enclose a river. A good average case shows a flow downwards in summer of half a foot a day at the sides and a foot and a half in the middle. The distance below the snow-line to which the flowing glacier descends down a mountain gorge--before it melts away and becomes a river of liquid water--depends, as does the rate at which it moves, in the first place, on the temperature of the region and on the sharpness of the slope. A glacier will flow downwards (as will a lump of pitch) along a scarcely perceptible incline, but more slowly than down a steeper incline, and it will, consequently, get further down into the warm valley without altogether melting away when the slope is steep. But apart from these considerations, the bigger and thicker (or deeper) the glacier, that is to say, the more snow which each year falls at its starting-place and goes to making it, the further down will it flow before melting away; and it is the heavy snowfall of many years ago or of a series of years long past which has to-day reached in the form of ice the lower end of the glacier. So, though the lower end of the glacier may melt more quickly if the valley has become hotter, yet the heavy snowfalls of fifty years ago may only now have reached the valley, and may quite counterbalance the melting action of the warmer summers. Or reverse conditions, namely, less snow and lower or unchanged temperature in the valley, may prevail.