that of some of the regions of Greenland which have only quite recently been vacated by the glaciers, while the composition of the shingle—the inclusion over so long a front of bowlders from beyond the first line of mountain heights, many of them most markedly grooved and polished—is also highly suggestive of glacial deposition. The gold sands, or sands that are worked for gold, are merely the ordinary materials of the beach, loose and incoherent like most seashore sands, and particularly defining horizons three to six feet below the surface. In regular stratified layers, with fine and moderately coarse gravel, they embrace four or five distinct layers of fragmented garnets (the components of the so-called "ruby sand"), and it is from these, and at this time almost exclusively from the
The Harbor of Snake River (Nome).
bottom layer of three to five inches thickness, which is popularly described as lying on "bed rock"—in most places merely a hard-pan of arenaceous clay or argillaceous sand, with no true rock to define it—that most of the gold is obtained. Each ruby band nearer to the top seems to contain less and less gold, and there is no question that the different layers are merely reformations by the sea from those of earlier deposition, just as surface shingle deposits generally are in part reconstructions of underlying beds. That the ocean is to-day depositing the ruby sand is unmistakably shown by the great patches of this sand lying on the surface and its incoming in the path of nearly every storm. Even these surface sands are mildly gold-bearing, showing that the gold, despite its high specific gravity, may be buoyed up and wafted in by such a
