It is difficult to determine the stage in the history of a rock at which silicification has set in. As A. Jukes-Browne[24] remarks, solution of the silica skeletons may be accelerated by pressure, _i.e._ by the depth of water in which the bed accumulated. Yet, in comparison with the calcareous shells of foraminifera, radiolarian and diatomaceous remains are only slowly soluble, and are found in the deepest spots reached by soundings. H. B. Guppy[25], on the other hand, has observed silicification of modern corals in reefs in the Fijis, and believes that the process went on during the elevation of the area, when waters containing silica became concentrated, and parts of the mass were exposed to evaporation. The instability of the non-crystalline siliceous skeletons in geological time makes it probable that a rock cannot long retain them when buried among other strata in the earth. It is clear that there is no support for the view, current from the time of James Hutton onwards, that nodular flints are formed by matter in hot solutions entering pre-existing cavities in limestone rocks. But there must be cases where the silicification of limestone has arisen through its penetration by hot springs. The presence of tabular flint in joints of the Chalk shows that water has imported silica along easy lines of passage from some other portion of the rock. Just as stems of trees become replaced by chalcedonic silica, so may beds of limestone be converted into flint, especially in volcanic areas. A. W. Rogers[26] records that recent limestones formed in the Cape province by the evaporation of ascending waters have already become silicified. These flinty rocks have been found in the Kalahari Desert and elsewhere, though not south of the Orange River; the chemical change is probably due to the character of local water rather than to temperature. Yet it is remarkable how, in the vast majority of instances, the partial or complete silicification of a limestone may be traced to an intermediate resting stage of the silica in the form of skeletons of the vegetable diatoms or the animal sponges or radiolarians. The decay of flint itself, by the removal of part of its substance in solution, is the cause of the white surface on specimens from the Chalk, and of the crumbling white residues found in certain gravels. This process has been fully discussed by J. W. Judd, who believes that the material removed is silica in the opaline condition[27]. LIMESTONE AND SCENERY