acetabulum to the most anterior part of the iliac crest is 190 mm. The narrowest part of the ilium is 65 mm.

The acetabulum is large, and but for a marked development of the upper iliac border and of the ischial border would be rather shallow. Its greatest width is 115 mm., and its vertical measurement 95 mm. The three bones forming it are so completely anchylosed that their limits cannot be with certainty determined.

The pubis and ischium are so completely anchylosed that no trace of the dividing suture can be detected, and the bones of the two sides are likewise completely anchylosed along the symphysis. The bones are massive, and represent a somewhat specialised modification of the plate-like type.

The pubis is considerably longer vertically than antero-posteriorly. From the anterior part of the symphysis to the point where it probably unites with the ilium is about 170 mm. The greatest antero-posterior measurement is probably not more than 100 mm. There is a large pubic foramen situated as shown in the figure. As the bone lies in the body the pubic canal passes almost directly upwards. The outer half of the front edge of the pubis is much. thickened and bent downwards and forwards, and I think there is little doubt that it supported a prepubic cartilage. This surface for the cartilage measures about 90 mm. in length, and varies from 20 to 30 mm. in width. The lower end of the surface rises abruptly from the inner part of the anterior border of the pubis.

The ischium measures about 180 mm. by about 135 mm., and but for the great vertical depth would be practically of the plate-like type. The symphysis is very thick. In the pubic region it reaches a thickness of 68 mm., but becoming thinner in the ischial region it averages about 55 mm., and at the posterior part of the ischium it is only about 35 mm.

Femur.

The right is preserved in perfect condition. It resembles pretty closely the femur of Propappus omocratus, but differs from that of Pareiasaurus serridens in a number of important features. It agrees with the femur of Propappus omocratus in having the axis of the upper third of the bone making an angle with the lower part of the shaft of about 45°. It also agrees with it, and differs from that of Pareiasaurus in having the middle of the shaft constricted and nearly round. And lastly, it agrees with the femur of Propappus omocratus in that the trochanter minor is directed backwards and outwards, so that a deep concavity is formed between it and the

trochanter major. The femur of Propappus rogersi differs from that of P. omocratus in being shorter and stouter; in having the head more largely cartilaginous; in having the great trochanter more directly continuous with the articular surface; and in having the outer condyle relatively larger.

The following is a comparison of the measurements of the femur in the two species:

The right tibia and fibula are both in beautiful preservation. The tibia resembles considerably that of Propappus omocratus, but differs in a few points. The proximal articular end had probably a considerable pad of cartilage. As preserved it is a large flattened. surface rising up to a prominent boss in the middle. It measures 112 mm. by 77 mm. as against 110 mm. by 85 mm. in Propappus omocratus. The distal articular surface is kidney-shaped and measures 94 mm. by 55 mm., as against 90 mm. by 60 mm. in Propappus omocratus. The total length of the bone is 177 mm.: in P. omocratus it measures 200 mm.

The fibula is the most perfect Pareiasaurian fibula known. It is a slightly curved bone, with a slender shaft and dilated at each end. The proximal end is expanded and flattened antero-posteriorly and articulates on a distinct facet on the outer condyle of the femur. The distal end is flattened in the opposite direction to the upper and has a large convex articulation for the outer side of the proximal tarsal bone. The total length of the bone is 206 mm. The proximal end measures 63 mm. by 36 mm., and the distal end 68 mm. by 42 mm. The shaft is 27 mm. in diameter.

Proximal Tarsal Bone.

The proximal tarsal bone, which certainly represents the conjoined tibiale and fibulare with possibly the intermedium, is of the wellknown Pareiasaurian type.

The Pose of the Limbs.

Though most of the bones of Pareiasaurus have been known for twenty years, and though four skeletons are now known from South Africa, and at least as many from Russia, there are still a good many points in dispute in regard to the pose of the animal. Seeley had the British Museum specimen mounted with the limbs so bent that the abdomen must have been almost touching the ground, with the interclavicle in advance of the plane of the occiput, and with the scapula pointing almost directly backwards. Amalitzky has set up his skeletons with the limbs straighter, with the scapula directed as much upwards as backwards and with the shoulder girdle situated some little distance behind the head. In giving a restoration of Pareiasaurus in 1903 from the South African Museum specimen, I also placed the shoulder girdle a little distance behind. the head, and made the animal stand well off the ground. Boulenger, in discussing Telerpeton in 1905, criticised my restoration, pointing out that in his opinion Pareiasaurus had no more neck than a Salamander.

The very perfect bones of Propappus, though they may leave a little doubt on a few points, enable us to have a much clearer idea of the pose of the animal and of its movements than was previously possible. As all the limb bones have their articular surfaces perfectly preserved, we can fit the limbs together with certainty.

The peculiarly shaped glenoid cavity, with its large posterior concavity and the small anterior flattened surface, so exactly corresponds to the head of the humerus if we allow for a cartilaginous pad that we can articulate the humerus on the shoulder girdle with confidence, and the position of the bones when the animal is standing must be as shown in the figures given. The movement of the humerus must have been practically limited to an up and down direction, and when the animal lay down on its front the anterior limbs would lie outwards and forwards by its sides. In the standing position the humerus, radius, and ulna would be in the positions shown in the drawings. One interesting thing about the anterior limb is that it becomes quite manifest that the scapula must have been directed much more upwards than backwards, and as the shoulder girdle is about twice as long as the head is deep it cannot have been situated close up to the head, and therefore Parciasaurus and Propappus must be regarded as having had distinct necks.

The hind limb was probably more freely movable than the

anterior, but the movement was mainly between the femur and the innominate. The head of the femur must have had a large rounded pad of cartilage which fitted into the acetabulum, and allowed of considerable freedom of movement. In the knee-joint the amount of movement must have been very limited. Both femur and tibia doubtless had articular cartilages, but the evidence is that these were comparatively thin, and the articular surfaces of the femur are not larger than those of the tibia. It seems probable that the amount of movement of the tibia on the femur was limited to about 45°. The relative positions of the femur, tibia, fibula, and the proximal tarsal bone when the animal is standing are shown in the figures given. The degree of movement between the tarsal bone and the tibia and fibula was probably very limited, and the ankle movement must have been mainly between the distal tarsals and the proximal tarsal bone.

Pareiasaurus and Propappus were heavily built animals which probably walked with slow, deliberate movements such as we see in the large tortoises. They were land animals, and it seems more likely that they lived even on the dry land than that they frequented the marshes. The structure of the claws and the humerus would seem to indicate that they were digging animals, and probably, like Echidna, they defended themselves from their carnivorous enemies by digging into the ground. The occurrence of bony plates along the dorsal region of Pareiasaurus would serve to protect the most vulnerable part of the body when the animal had dug its way into the ground for safety. In Propappus, a smaller and feebler animal, bony plates were situated all over the back and sides and apparently even on the outer sides of the limbs, but the plates on the back are very much larger and strengthened by a powerful median boss in each.