the materials which were blown about in one of the earliest deserts recorded in the history of the rocks. This desert differed markedly from all the large ones which are known to us at present, in having a predominance of carbonates instead of silicates in the "sand" grains. We must not, however, push the doctrine of uniformitarianism too far and insist that all the deserts in the past must have been composed of siliceous grains, because that is the rule in modern large deserts. On a small scale limestone deserts are forming now, and if large areas of limestones could be exposed in the arid regions of Africa or Arabia these limestone deserts would form on a vast scale. But there is now too much diversity in the rocks of the earth's crust, because throughout most of the world the continents have in large part been above sea level during the Tertiary and Quaternary, and erosion has been going on so that many types of rocks are exposed and particularly large areas of crystallines, and when any or all of these are brought under arid climatic conditions, grains of a great range in composition are exposed to the sorting action of wind. In the Middle Siluric of North America, on the other hand, a land area which had been covered by limestone was subjected to arid conditions, and there is no escape from the fact that dominantly lime grains were formed by the prolonged exposure during which mechanical processes alone were active, and decomposition played no part.
Succeeding the arid or semi-arid climatic conditions of the Salina was a period of greater rainfall and of expansion of the epi-continental seas. The rivers became permanent in response to the rains of a pluvial climate, and there followed upon the period of rock destruction in situ a period of transportation of material from the land into the sea. The prolonged disintegration of the limestones and dolomites with local shales had provided a vast soil covering which must have extended to a considerable depth, and which, because of fineness and friability could easily be removed by streams. Even the weakest little rivulet would be able to carry a small load of this material, which was so conveniently prepared. With the increased moisture in the air decay became active in further breaking down the mechanically disintegrated rocks, and in this way the igneous rocks that were exposed through erosion would yield a certain amount of silica and alumina as would also the shale bands in the limestones. Thus, while the rivers carried material which was dominantly calcareous or magnesian, certain impurities were also included. Some difficulty has been offered by the high amount of alumina, to account for
