On rocks mainly composed of silica, the lichen has a much harder substance to deal with, and one less easily affected by acids, though even silica may be dissolved in time. Uloth[853] concluded from his observations that the relation of plants to the substratum was chemical even more than physical, so far as crustaceous species were concerned. He found that the surface of the area of rock inhabited was distinctly marked: even such a hard substance as chalcedony was corroded by a very luxuriant lichen flora, the border of growth being quite clearly outlined. The corrosive action is due he considered to the carbon dioxide liberated by the plant, though oxalic acid, so frequent a constituent of lichens, may also share in the corrosion. Egeling[854] made similar observations in regard to the effect of lichen growth on granite rocks; and he further noticed that pieces of glass, over which lichens had spread, had become clouded, the dulness of the surface being due to a multitude of small cracks eaten out by the hyphae. Buchet[855] also gives an instance of glass which had been corroded by the action of lichen hyphae. It formed part of an old stained window in a chapel that was obscured by a lichen growth which adhered tenaciously. When the window was taken down and cleaned, it was found that the surface of the glass was covered with small, more or less hemispherical pits which were often confluent. The different colours in the picture were unequally attacked, some of the figures or draperies being covered with the minute excavations, while other parts were intact. It happened also, occasionally, that a colour while slightly corroded in one pane would be uninjured in another, but the suggestion is made that there might in that case have been a difference in the length of attack by the lichen. The selection of colours by the lichens might also be influenced by some chemical or physical characters. Bachmann[856] found that on granite there is equally a selection of material by the hyphae: as a rule they avoid the acid silica constituents; while they penetrate and traverse the grains of mica which are dissolved by them exactly as are lime granules.