On the north side of this peninsula, i. e., on the south shore of the northern half of the lake, about midway between the east and west shores, occurs one of the two sets of hieroglyphic markings. Lying off shore at a distance of a quarter to a half a mile, and making with it a long sheltered channel, is a chain of islands, trending east and west. On the south side of one of these islands, less than a mile to the west of the first locality, is to be seen the other set of inscriptions. The first set occurs on the top of a low, glaciated, projecting point of rock, which presents the characters of an ordinary roche moutonnée. The rock is a very soft, foliated, green, chloritic schist, into which the characters are more or less deeply carved. The top of the rounded point is only a few feet above the high-water mark of the lake, whose waters rise and fall in different seasons through a range of ten feet. The antiquity of the inscriptions is at once forced upon the observer upon a careful comparison of their weathering with that of the glacial grooves and striæ, which are very distinctly seen upon the same rock surface. Both the ice grooves and carved inscriptions are, so far as the eye can judge, identical in extent of weathering, though there was doubtless a considerable lapse of time between the disappearance of the glaciers and the date of the carving. The island on which were found the other inscriptions is one of the many steep rocky islands known among the Indians as Ka-ka-ki-wa-bic min-nis, or Crow-rock island. The rock is a hard greenstone, not easily cut, and the inscriptions are not cut into the rock, but are painted with ochre, which is much faded in places. The surface upon which the characters are inscribed forms an overhanging wall protected from the rain, part of which has fallen down. The Indians of the present day have no traditions about these inscriptions beyond the supposition that they must have been made by the “old people” long ago. The sketches published as copies of these glyphs show spirals, concentric circles, crosses, horseshoe forms, arrow shapes, and other characters similar to those found on rocks in the southwestern part of the United States, and also to petroglyphs in Brazil, examples from both of which regions are presented in this work, under their appropriate headings. BRITISH COLUMBIA. Dr. Franz Boas (_a_) published an account of a petroglyph on Vancouver island (now presented as Fig. 3) which, slightly condensed, is translated as follows: [Illustration: FIG. 3.--Petroglyph on Vancouver island.]