For the new type I propose the name Noteosaurus africanus, g. et sp. nov.

The specimen was discovered in the district of Victoria West by Mr. Blake. There is no doubt it is from the same horizon as the South African species of Mesosaurus. Unfortunately only the pelvic region with the anterior part of the tail and the greater part of the two hind limbs are known; and further, though the feet are satisfactorily preserved the pelvis is almost hopelessly weathered away. A part of the ischium is preserved, a fragment of the pubis, and a fair portion of the ilium. There is no evidence as to whether the pubis was perforated or notched.

The femur measures 33 mm. in length. The ends resemble those of the South African Mesosaurs rather than the Brazilian in being less completely ossified. The tibia and fibula agree rather with those of M. brasiliensis. The fibula has, as in that species, the typical Cotylosaurian shape. It measures 21 mm. in length, and the head is 5·3 mm. wide. The tibia is rather slender. The tarsus is well preserved. The two large proximal elements are very similar to those in M. brasiliensis and M. capensis, the resemblance to those of the former being the more marked. Distally there appear to be only four tarsalia. The shape of the elements will be best understood from the figure.

The following are the lengths of the metatarsals and digits in this new species, and, for comparison, the lengths in the foot of the Nieuwoudtville Mesosaurus which is believed to be M. capensis.

In comparing the feet of the two forms the most striking differences are the relatively greater lengths of the 3rd, 4th, and 5th metatarsals in Mesosaurus, the greater length of the 4th toe in Mesosaurus, and the relatively greater length of the 5th toe in

Noteosaurus. The possession of 6 phalanges in the 5th toe of Noteosaurus is a striking difference from Mesosaurus, when certainly there are only 4 in the known specimens.

The phalanges of the 6th toe are slender, and the last one is pretty certainly not clawed, whereas those of the other toes are short and probably had some sort of nail or claw.

Noteosaurus africanus is a very near ally of Mesosaurus, and is an interesting modification. The increase in number of the phalanges of the 5th toe is exactly comparable to the increase in the number of phalanges in the manus of the Cetacea. It is remarkable to find this modification in one of the oldest known reptiles. The zone in which it occurs is certainly not later than Lower Permian. It may even possibly be Upper Carboniferous. It is further remarkable that only one toe should be thus specialised.

21.—A Revision of the Reptiles of the Karroo.-By R. BROOм, M.D., D.Sc., F.R.S.S.Af.

PROBABLY no fossiliferos deposit in the world will ever compare in importance with the extensive Permian, Triassic, and Jurassic beds forming the Karroo system of South Africa. We have here the continuous record of perhaps 3,000,000 years, and as they were the years which saw the birth of all the various reptilian orders, of the birds and of the mammals, the most important years in the world's history.

Though a large number of species are known, and though we have a very good idea of the general character of the fauna at different periods, the known species must be few in number in comparison to those which remain to be discovered.

In the last ten years the number of known types has been more than doubled, and one feels compelled from time to time to stop and take stock lest one gets overwhelmed with the multitude of forms.

In the following lists I have endeavoured to classify all the known reptiles of the Karroo. For many years it will be impossible to be certain of the validity of a few species founded on very imperfect remains. Fortunately most South African workers have been careful only to found species on skulls or good skeletal remains, and hence the confusion caused so frequently in America and Europe by having names given to vertebræ or limb bones scarcely exists with us, and the few such types as there are will be treated as they deserve to be.

As an Order can rarely, if ever, be as carefully defined when first proposed as a species or genus, it is impossible to apply the rules of nomenclature to the large groups. Doubtless it is right to commemorate the work of the pioneers by using the names proposed by them wherever possible, but it is quite impossible always to do so without causing hopeless confusion. For example, the name Cotylosauria was proposed by Cope in 1880 as a suborder of which Diadectes was the type. It was founded on an error and was given up. In 1889 it was re-established as an order to include the reptiles with roofed temporal regions. In 1888, however, Seeley had proposed the order Pareiasauria, and there has been much discussion and

difference of opinion as to which name has the priority and ought to be used. If the question is to be decided by strict priority then neither name must be used, but Owen's term Serratidentia, proposed in 1876 for the group to which Pareiasaurus belongs. As, however Owen believed Pareiasaurus to be a Dinosaur, and as no one has ever adopted Owen's term for the reptiles with a roofed temporal region it seems necessary to use one of the other two terms. As Seeley in 1888 manifestly did not mean his Pareiasauria to include any but Pareiasaurus-like animals and in his classification puts the Procolophonia and the Cotylosauria as other orders of equal value we can hardly use his term as a super-order. On the other hand, Cope in 1889 clearly made Cotylosauria an embracing order including all the primitive reptiles with roofed temporal regions. And as most later writers have continued to use Cope's term in this sense it seems best to use the term Cotylosauria as a super-order and to use the term Pareiasauria as a suborder of the Cotylosauria.

In the following list I have used the terms which seem to me those that have most claim. Very doubtful species I have omitted. Species that are a little doubtful but possibly good I have retained :—

P. indicates that the form is Permian; T. Triassic; and J. Jurassic.