142 
Annals of the Transvaal Museum. 
cone. This may be the same process as the characteristic one of Eryops. 
The other is situated near the lower end of the anterior border and seems 
to be rather flat. 
The radius is a short bone with a square shaft of which the lateral 
sides are slightly concave, while the anterior side is flat ; there is a 
longitudinal groove on the distal half of the posterior side. The proximal 
end is thick and nearly square, the articulation surface being flat and 
standing at right angles to the axis of the bone. The distal end is not so 
thick as the proximal and is not square, the different angles being rounded 
off. The articulation surface, which is hollow, does not make right angles 
with the axis of the bone, the inner side of the shaft being longer than 
the outer. 
The ulna has the shaft bent with the concavity inwards. Its distal 
end is broad and flat and the articulation surface is strongly convex. 
The bone thickens towards the proximal end, which is broad and thick, 
the articulation surface being larger than that of the radius. This surface 
looks towards the radius, the concave inner side of the shaft being much 
shorter than the convex outer one. There is no olecranon process. 
The carpus is only represented by two small bones, which are mainly 
situated between the radius and the first or second metacarpal. As the 
fore-foot is present in three complete specimens and as all three have only 
two carpals, I can only, conclude that the others were not ossified. The 
proximal bone is most probably the radiale, while the other might be the 
first carpal. 
There are four metacarpals , of which the first is broadest and shortest, 
the others being longer and narrower. The four fingers w T ere short, and 
the hand gives a much weaker impression than the foot. The first and 
second fingers have two phalanges each, while the third and fourth have 
each three. In one specimen there is an irregularity with the third finger, 
which has only two phalanges. It seems that the end phalange is lost. 
The Pelvis. 
One pelvis is nearly completely preserved, while the two others are 
only represented by fragments and their outside cast in the matrix 
(PI. XVI and XXII). In all these cases the iliac portion of the right side 
has been pressed into the plane of the ischio-pubis. The left side in the 
two impressions with fragments is bent and generally distorted. Of the 
preserved specimen of the left side a large part of the ischio-pubis, 
bordering on the symphysis, is missing. The iliac portion stands at an 
angle of 45 degrees upwards and outwards. This seems to me to be about 
its normal position. 
The ischio-pubis portion is flat and joined to its fellow by a broad 
symphysis which is broader dorsally than ventrally. Its seems that the 
two innominates joined each other without forming an angle. The hinder 
border of the ischium is convex, so that the two united bones are separated 
distally for a short distance by a notch. The outer border is slightly 
concave until the ilium is reached, where a small convexity marks the 
beginning of this bone. The suture between the ilium and the pubis is 
further marked by a ridge on the dorsal surface, stai ting from this con- 
vexity and crossing the bone in a straight line inwards and forwards. 
The front border of the ischio-pubis is broader than the posterior 
border and slightly concave. There is an angular projection where the 
ilium meets the bone and the suture between the ilium and the innominate 
is marked by a slight ridge, which unites with the ridge starting on the 
