shore line advanced southwards, and over the deepwater deposits of the Bokkeveld were laid down once again shallow-water sands, the Witteberg rocks. The sands then, recently formed on the strip of coast land which had sunk and which now was laid bare again, became washed down into the water and formed the Witteberg sandstones. The grains, suffering thus a double erosion, became smaller, hence the grains of the Witteberg are smaller as a rule than those of the Table Mountain Sandstone. The iron content of the Witteberg one may assume to be due to the fact that on such a coastal ledge as that from which they were derived the drainage would be imperfect, ironstone gravel would accumulate under the soil as it does in the coastal ledges at the present day, and this ironstone gravel would be washed down and incorporated in the sediments forming offshore. It may be, also, that the deposit of iron has something to do with the water being fresh, as iron deposits are never formed in the sea. The Transvaal geologists have assumed that some of the Table Mountain Sandstone of this northern coast shelf was left and not washed down, and that it now forms the Waterberg Sandstone; but this is improbable if the order of events was anything like what has just been described. The Bokkeveld Series at the base is undoubtedly marine, but there is evidence that the open ocean was being enclosed, and in the Witteberg times the deposit was laid down in a freshwater lake, for the fossils of the Witteberg are those usually found in freshwater deposits. While the Bokkeveld was thus changing from a marine to a freshwater deposit the shore was continually oscillating, and when it advanced south shore-deposits were laid down, as evidenced now by the sandstone bands. When the
