If the glacier travelled on (and it certainly does travel on), and the water kept cutting back over the edge of the ice, there would be a great slit in front of the cascade; if the water did not cut back, the whole hollow and cascade, as you say, must travel on; and do you suppose the next season it falls down some crevice higher up? In any case, how in the name of Heaven can it make a hollow in solid rock, which surely must be a work of many years? I must point out another fact which Agassiz does not, as it appears to me, leave very clear. He says all the blocks on the surface of the glaciers are angular, and those in the moraines rounded, yet he says the medial moraines whence the surface rocks come and are a part [of], are only two lateral moraines united. Can he refer to terminal moraines alone when he says fragments in moraines are rounded? What a capital book Agassiz's is. In [reading] all the early part I gave up entirely the Jura blocks, and was heartily ashamed of my appendix (499/2. "M. Agassiz has lately written on the subject of the glaciers and boulders of the Alps. He clearly proves, as it appears to me, that the presence of the boulders on the Jura cannot be explained by any debacle, or by the power of ancient glaciers driving before them moraines...M. Agassiz also denies that they were transported by floating ice." ("Voyages of the 'Adventure' and 'Beagle,'" Volume III., 1839: "Journal and Remarks: Addenda," page 617.)) (and am so still of the manner in which I presumptuously speak of Agassiz), but it seems by his own confession that ordinary glaciers could not have transported the blocks there, and if an hypothesis is to be introduced the sea is much simpler; floating ice seems to me to account for everything as well as, and sometimes better than the solid glaciers. The hollows, however, formed by the ice-cascades appear to me the strongest hostile fact, though certainly, as you said, one sees hollow round cavities on present rock-beaches. I am glad to observe that Agassiz does not pretend that direction of scratches is hostile to floating ice. By the way, how do you and Buckland account for the "tails" of diluvium in Scotland? (499/3. Mr. Darwin speaks of the tails of diluvium in Scotland extending from the protected side of a hill, of which the opposite side, facing the direction from which the ice came, is marked by grooves and striae (loc. cit., pages 622, 623).) I thought in my appendix this made out the strongest argument for rocks having been scratched by floating ice.