But in the many cases observed in America it is not necessary to propose either the hypothesis involving such elaborate work as is suggested or one postulating enormous geological changes. The escarpment of cliffs is from time to time broken down by the action of the elements and the fragments fall to the base, frequently forming a talus of considerable height, on which it is easy to mount and incise or paint on the remaining perpendicular face of the cliff. When the latter adjoins a lake or large stream, the disintegrated débris is almost immediately carried off, leaving the drawings or paintings at an apparently inaccessible altitude. When the cliff is on dry land, the rain, which is driven against the face of the cliff and thereby increased in volume and force at the point in question, also sweeps away the talus, though more slowly. The talus is ephemeral in all cases, and the face of the cliff may change in a week or a century, as it may happen, so its aspect gives but a slight evidence of age. The presence, therefore, of the pictures on the heights described proves neither extraordinary skill in their maker nor the great antiquity which would be indicated by the emergence of the pictured rocks through volcanic or other dynamic agency. The age of the paintings and sculptures must be inferred from other considerations. Pictures are sometimes found on the parts of rocks which at present are always, or nearly always, covered with water. On the sea shore at Machias bay, Maine, the peckings have been continued below the line of the lowest tides as known during the present generation. In such cases subsidence of the rocky formation may be indicated. At Kejimkoojik lake, Nova Scotia, incisions of the same character as those on the bare surface of the slate rocks can now be seen only by the aid of a water glass, and then only when the lake is at its lowest. This may be caused by subsidence of the rocks or by rise of the water through the substantial damming of the outlet. Some rocks on the shores of rivers, e. g., those on the Kanahwa, in West Virginia, show the same general result of the covering and concealment of petroglyphs by water, except in an unusual drought, which may more reasonably be attributed to the gradual elevation of the river through the rise of the surface near its mouth than to the subsidence of the earth’s crust at the locality of the pictured rocks.