It would seem from the quantity of water that is sometimes ingulfed within these open breaks in the ice, that the glacier must occasionally be fissured to a very great depth.  I remember once, when boring a hole in the glacier in order to let down a self-regulating thermometer into its interior, seeing an immense fissure suddenly rent open, in consequence, no doubt, of the shocks given to the ice by the blows of the instruments.  The effect was like that of an earthquake; the mass seemed to rock beneath us, and it was difficult to keep our feet.  One of these glacial rivers was flowing past the spot at the time, and it was instantly lost in the newly formed chasm.  However deep and wide the fissure might be, such a stream of water, constantly poured into it, and daily renewed throughout the summer, must eventually fill it and overflow, unless it finds its way through the whole mass of the glacier to the bottom on which it rests; it must have an outlet above or below.  The fact that considerable rivulets (too broad to leap across, and too deep to wade through safely even with high boots) may entirely vanish in the glacier unquestionably shows one of two things,—­that the whole mass must be soaked with water like a wet sponge, or the cavities reach the bottom of the glacier.  Probably the two conditions are generally combined.

In direct connection with the narrower fissures are the so-called moulins,—­the circular wells on the glacier.  We will suppose that a transverse, narrow fissure has been formed across the glacier, and that one of the many rivulets flowing longitudinally along its surface empties into it.  As the surface-water of the glacier, producing these rivulets, arises not only from the melting of the ice, but also from the condensation of vapor, or even from rain-falls, and flows over the scattered dust-particles and fragments of rock, it has always a temperature slightly above 32 deg., so that such a rivulet is necessarily warmer than the icy edge of the fissure over which it precipitates itself.  In consequence of its higher temperature it melts the edge, gradually wearing it backward, till the straight margin of the fissure at the spot over which the water falls is changed to a semicircle; and as much of the water dashes in spray and foam against the other side, the same effect takes place there, by which a corresponding semicircle is formed exactly opposite the first.  This goes on not only at the upper margin, but through the whole depth of the opening as far down as the water carries its higher temperature.  In short, a semicircular groove is excavated on either side of the fissure for its whole depth along the line on which the rivulet holds its downward course.  After a time, in consequence of the motion of the glacier, such a fissure may close again, and then the two semicircles thus brought together form at once one continuous circle, and we have one of the round deep openings on the glacier known as moulins, or wells, which may of course become perfectly dry, if any accident turns the rivulet aside or dries up its source.  The most common cause of the intermittence of such a waterfall is the formation of a crevasse higher up, across the watercourse which supplied it, and which now begins another excavation.