looked upon as the ancestors from which the more specialized reptiles of later ages and the mammals are descended. The most generalized type is the great Pareiasaurus; the commonest form, however, is the Dicynodon, with two tusks, and the female form, Oudenodon, with no tusks. Towards the top there are the forms specialized on the lines of mammals in that the skulls have the teeth arranged as in a dog, with incisors, canines, and molars. Gomphognathus, Trirachodon, Cynognathus, and a host of similar forms are of this type. A little step farther and we have what is to all intents and purposes a mammal, Tritylodon. Although this fossil is usually said to come from the Karroo, the history of the specimen has been lost, and from an examination of the original I have no hesitation in saying that it comes from the Red Beds of the Stormberg Formation, that is, is lower Jurassic and not Triassic. The Karroo Beds extend to the sea at East London, and farther east they have been termed the Kentani Beds. Still farther east they are not very well represented, a fact which has led to confusion. The occurrence of reptilian remains in the shales along the Umkomazan River, in Western Natal, seems to be an isolated reappearance.
With the Beaufort Beds the Karroo Beds, as they occur normally in the Karroo, come to an end. A continuation of the sediments occurs in the Transvaal, and right up in the lake district the Drummond Beds are of the same age, though they were laid down in another lake basin. The whole southern hemisphere was united in one great continent, which has been called, by Suess, Gondwanaland, because the beds are all of the type that occur in the Gondwana System in India. In India the glacial conglomerate at the base is called the Talchir
