I believe that circumstances similar to those determining the more rapid advance of the glaciers from higher to lower levels at that point where the alternate thawing and freezing, the infiltration of water and consequent expansion of the ice under frost, are greatest, would also determine the motion of a large body of ice from north to south, since it would be along its southern limits that these conditions would prevail; while the great reservoir of snow at the north would correspond to the upper troughs of the present glaciers, from which their lower ranges are constantly fed. The change of snow into ice is owing to alterations of temperature, to partial melting and subsequent freezing, constantly renewed,--and also to the sinking of the mass upon itself in consequence of its own weight, the lower portions being thus forced out by the pressure of the superincumbent ice. Upon an inclined plane the movement consequent upon these changes will of course be downward; but what would be the result, if a field of snow many thousand feet thick, corresponding, except in its greater bulk, to the accumulations by which the present glaciers are caused, were stretched over an extensive level surface? The moisture from the upper superficial layers would permeate the larger mass as it now does the smaller one, trickling down into its lower portions, while the pressure from above would render the bottom hard and compact, changing it gradually into ice. If this should take place under climatic conditions which would keep the whole as a mass in a frozen state, the pressure from above would force out the lower ice in every direction beyond its original circumscription, thus enlarging the area covered by it, while the whole would subside in its bulk. Let us for a moment assume that such an accumulation of snow takes place around the northern and southern poles, stretching thence over the northern and southern hemispheres to latitude forty, and that this field of snow acquires a thickness of from twelve to fifteen thousand feet. Such a mass would subside upon itself in consequence of its own weight; it would be transformed into ice with greater or less rapidity and completeness, according to the latitude determining the surrounding climatic influences and the amount of moisture falling upon it as rain or dew, the alternations of temperature being of course more frequent and greater along its outer limit. In proportion as, with the rising of the temperature, these alternations became more general, a packing of the mass would begin, corresponding to that observed in the glacial valleys of Switzerland, though here the action would not be intensified by lateral pressure; an internal movement of the whole mass would be initiated, and the result could be no other than a uniform advance in a southerly direction from the Arctic toward the more temperate latitudes in Europe, Asia, and North America, and from the Antarctic toward South