The drainage of a country recently affected by glaciers is always imperfect. Owing to the irregular erosion of the bed rocks, and to the yet more irregular deposition of the detritus, there are very numerous lakes which are only slowly filled up or by erosion provided with drainage channels. Though several thousand years have passed by since the ice disappeared from North America, the greater part of the area of these fresh-water basins remains, the greater number of them, mostly those of small size, have become closed. Where an ice stream descends into the sea or into a large lake, the depth of which is about as great as the ice is thick, the relative lightness of the ice tends to make it float, and it shortly breaks off from the parent mass, forming an iceberg. Where, as is generally the case in those glaciers which enter the ocean, a current sweeps by the place where the berg is formed, it may enter upon a journey which may carry the mass thousands of miles from its origin. The bergs separated from the Greenland glaciers, and from those about the south pole, are often of very great size; sometimes, indeed, they are some thousand feet in thickness, and have a length of several miles. It often happens that these bergs are formed of ice, which contains in its lower part a large amount of rock _debris_. As the submerged portion of the glacier melts in the sea water, these stones are gradually dropped to the bottom, so that the cargo of one berg may be strewed along a line many hundred miles in length. It occasionally happens that the ice mass melts more slowly in those parts which are in the air than in its under-water portions. It thus becomes top-heavy and overturns, in which case such stony matter as remains attains a position where it may be conveyed for a greater distance than if the glacier were not capsized. It is likely, indeed, that now and then fragments of rock from Greenland are dropped on the ocean floor in the part of the Atlantic which is traversed by steamers between our Atlantic ports and Great Britain. Except for the risks which they bring to navigators, icebergs have no considerable importance. It is true they somewhat affect the temperature of sea and air, and they also serve to convey fragments of stone far out to sea in a way that no other agent can effect; but, on the whole, their influence on the conditions of the earth is inconsiderable.