On the 3rd of August I was again at the base of the Talefre cascade, and observed a fact the significance of which had previously escaped me. The rills which ran down the ice-slopes collected at the base of each protuberance into a stream, which, at the time of my visit, had hollowed out for itself a deep channel in the ice. At some places the stream widened, at others its banks of ice approached each other, and rapids were produced; in fact, _the channels of such streams appeared to be the exact moulds of the seams of white ice_. Instructed thus far, I ascended the Glacier du Geant on the 5th of August, and then observed on the wrinkles of this glacier the same leaning backwards and forwards of the blue veins as I had previously observed upon the Talefre. I also noticed on this day that a seam of white ice would sometimes open out into two branches, which, after remaining for some distance separate, would reunite and thus enclose a little glacier-island. At other places lateral branches were thrown off from the principal seam, thus suggesting the form of a glacier-rivulet which had been fed by tributary branches. On the 7th of August I hunted the seams still farther up the glacier; and found them at one place descending a steep ice-hill, being crossed by other similar bands, which however were far less white and compact. I followed these new bands to their origin, and found it to be a system of crevasses formed at the summit of the hill, some of which were filled with snow. Lower down the crevasses closed, and the snow thus jammed between their walls was converted into white ice. These seams, however, never attained the compactness and prominence of the larger ones which had their origin far higher up. I singled out one of the best of the latter, and traced it through all the dislocation and confusion of the ice, until I found it to terminate in a cavity filled with snow.