The magnesium required for dolomitisation is derived from the magnesium sulphate and chloride of sea-water, calcium being removed during the change. C. Klement in particular urges that a concentrated solution of sodium chloride at 60° C. assists the process in the case of magnesium sulphate. Aragonite, the material of coral skeletons and of most molluscan shells, is more susceptible than calcite. The temperature of Klement's experiments may be realised in lagoons or between tide-marks, and Doelter suggests that the element of time in nature may allow the reaction to take place at lower temperatures. The intimate structure of modern dolomitic limestone, as exhibited in coral-reefs, satisfies us that many older or fossil dolomites were formed from marine calcareous deposits while these were still accumulating. In other cases we must admit that the dolomite has developed in the neighbourhood of joints after the consolidation of the rock. The view that dolomitisation results from the mere removal of calcium, the magnesium originally present in organic skeletons becoming thus more concentrated, is not borne out by recent observations. Skeats[15] has carefully compared the dolomite-rocks of Tyrol with the materials of recent coral-reefs. In both there is a striking absence of detritus of inorganic origin, and his work goes far to show that the much-discussed Alpine dolomites were formed under conditions which occur in the neighbourhood of existing reefs. This, however, does not solve the question as to whether we are dealing in Tyrol with fossil coral-reefs, or with the calcareous type of ordinary marine sediments, which might undergo the same kind of alteration. While Skeats finds in two dolomites from recent reefs 43 per cent. of magnesium carbonate, the substitution seems usually to terminate when 40 per cent. has been introduced. In Tyrol, however, the process has gone so far as to give rise to true dolomites, with 45·65 of magnesium carbonate. The dolomites of the Jurassic series in north Bavaria are massive rocks almost devoid of fossils, traversed by shrinkage cracks, and associated with richly fossiliferous stratified limestones. The relations of these two types of rock are those of coral-reefs to the bedded deposits on their flanks, and the dolomite seems to merge horizontally into the stratified series. As in Tyrol, fossils and corals are rare in the bosses of dolomite, but the structural evidence is strongly in favour of their having originated as steeply sided reefs.