The structure of the ruins of the mushroom rock type is not radically different from that of the towers above described, they being exceptional only in their unusual sites. They occur on top of eroded pillars of rock, often enlarged on top, reminding one of mushrooms, like the so-called Snake rock at Walpi. They were once extensions or spurs of the mesa but are now rock pillars cut off by erosion so that they stand out isolated from the rim of the canyon. On account of the difficulty in reaching their tops, the ground plan of many could not be observed, but with a glass it was seen that as a rule they conform to the shape of the rim of the rock on which they stand. Considering the unusual sites of these inaccessible buildings, the question naturally arises, How could the ancient dwellers enter these rooms? Had they ladders or ropes, or were footholes cut in the side of the cliff to aid them? If the theory of footholes be correct we may suppose that these have been worn away, for no trace of them could be found. A geological question might likewise suggest itself to anyone seeing the evidences of erosion between the cliffs and pinnacles. Has the gap between the latter and the edge of the plateaux been ploughed out by the water since the building on the former were constructed? Although the cliffs show that the amount of the erosion has been enormous, it must be borne in mind that the prevailing rock is soft sandstone, the wearing away of which would not necessarily require a great period of time. It is not probable that these pinnacles have been separated by erosion from the cliff since man constructed the walls upon them, but this question involves the knowledge of a geological expert. To the same group of ruins as the mushroom type belongs one from a wholly different locality, shown in plate 12, _a_, a photograph of which was given the author by Mr. Chubbock. In this case the ruin is not built on top of a rock pinnacle, in the shape of an inverted cone, but in the horizontal fissure or constriction worn out under the harder stratum above it. The building in this cleft is in fact a kind of cliff house in which the front wall extends from top to bottom of the crevice, the rooms occupying a recess back of this wall. A somewhat similar form of habitation found in the side of a cliff has been described by the author.[19] It was discovered in the Verde Valley, Arizona, near Jordan’s ranch, about 6 miles from Jerome, Arizona. In his description it is classified as a “ledge house,” a type where the opening into the cave is completely walled up. Unlike a true cliff dwelling the rooms occupy the whole of a natural cave the top of which is its roof. It is not possible to determine from the illustration here shown whether or not the recess has been enlarged by artificial means, and as the author has not visited the ruin he has no idea of the arrangement of rooms. INVERTED CONE RUIN