had been first formed among the highest mountains, and extended downwards into the plains, the largest and highest moraines ought to be the most distant, and to be formed of the most rounded masses; whereas the actual condition of the detrital accumulations is the reverse, the distant materials being widely spread, and true moraines being found only in valleys connected with great chains of lofty mountains. Again, all these moraines are within one another,--the most distant from the glacier to which they owe their origin encircling all those which are nearer and nearer to it within the same glacial basin. And as no glacier could reach to its farthest moraine without pushing forward all the intervening loose materials, it is self-evident that the outer moraines were first formed, and those nearer the glacier subsequently, in the order in which they follow one another from the lower valleys to the higher levels at which alone glaciers exist at present. Translating these facts into words, we see that the glaciers to which these ancient moraines owe their origin must have been retreating gradually while the moraines were accumulating. But a glacier while uniformly retreating forms no high walls of loose materials around its edges and at its lower extremity; as it melts away, it only drops the burden of angular rocky fragments which it carries upon its back over the loose fragments above which it moves, and which it grinds to powder, or to sand, or to rounded pebbles, in its progress. It is only where the glacier remains stationary for a longer or shorter period that large terminal moraines can accumulate; and they are generally found in such places in the valleys of the Alps as would naturally determine the lower limit of a glacier for the time being. There is no possibility of escaping the conclusion that the ancient glaciers must have begun that series of oscillations to which the accumulation of the moraines is to be ascribed, at a time when ice-fields already occupied the whole area which they have covered during their greatest extension. After we shall have seen how many centres of dispersion of erratic boulders existed in the northern hemisphere, similar to that of the Alps, we may perhaps be able to form some idea of the manner in which these ice-fields originated and gradually vanished.