He does not believe that the berg falls into the sea, broken by its weight from the parent glacier; he is of opinion that it _rises from_ the sea. The process is at once gradual and comparatively quiet. “The idea of icebergs being discharged, so universal among systematic writers, seems to me at variance with the regulated and progressive actions of Nature. Developed by such a process, the thousands of bergs which throng the Polar seas should keep the air and water in perpetual commotion, one fearful succession of explosive detonations and propagated waves. But it is only the lesser masses falling into deep waters which could justify the popular opinion. The enormous masses of the great glacier are propelled, step by step and year by year, until, reaching water capable of supporting them, they are floated off to be lost in the temperatures of other regions.” The Humboldt Glacier did not differ in structure from the Alpine and Norwegian ice-growths; and its face presented nearly all the characteristic features of the latter. The _overflow_, or viscous overlapping of the surface, was very strongly marked. “When close to the island rocks,” says Kane, “and looking out upon the upper table of the glacier, I was struck with the homely analogy of the batter-cake spreading itself out under the ladle of the housewife, the upper surface less affected by friction, and rolling forward in consequence.” * * * * * The crevasses bore the marks of direct fracture, as well as of the more gradual action of surface-drainage. The extensive watershed between their converging planes gave to the icy surface most of the hydrographic features of a river-system. The ice-born rivers which divided them were margined occasionally with spires of discoloured ice, and generally lost themselves in the central areas of the glacier before reaching its foreground. Occasionally, too, the face of the glacier was cut by vertical lines, which, as in the Alpine examples, were evidently outlets for the surface drainage. The height of this ice-wall at the nearest point was about three hundred feet, measured from the water’s edge; and the unbroken right line of its diminishing perspective showed that this might be regarded as its constant measurement. It seemed, in fact, a great icy table-land, abutting with a clean precipice against the sea. This, indeed, is the great characteristic of all those Arctic glaciers which issue from central reservoirs or _mers de glace_ upon the fiords or bays, and is strikingly in contrast with the dependent or hanging glacier of the ravines, where every line and furrow and chasm seem to indicate the movement of descent and the mechanical disturbances which have impeded and delayed it.