Annals of the Transvaal Museum, 
137 
the three ridges on the hypocentrum of I rimer orhachis occupy practically 
the whole of its lower surface, excepting the sides, they only occupy its 
central half in Myriodon. Both ends of the hypocentrum project slightly 
below its outer surface. The sides of the hypocentrum rise upwards, 
terminating in a point. The outer surface of these parts is concave and 
their hinder border shows a concavity, which has most probably served as 
articulation surface for the rib. 
For so far present the presacral vertebrae have all got pleurocentra, 
one on each side. These elements are bent, so as to partly envelope the 
chorda. Their outer surface is hollow and shining and covered with a 
coarse net-like structure. The other surfaces are rough and were probably 
covered by cartilage. Most of the pleurocentra are out of their original 
position and have suffered through fossilization. Only a few are still near 
their natural position and have retained a normal shape. From these it is 
seen that their hinder border is slightly convex, whilst the front border is 
formed by two sides meeting at a very obtuse angle at the middle. The 
upper side stands convexly outwards, while the ilower is concave outwards 
at its upper end. It could not be made out whether the rib articulated 
with the upper front part of this bone. The transverse dimension of the 
pleurocentra is greater than the longitudinal. 
The neural arch consists of a coalesced bone with a neural spine. 
The coalescing has taken place above the neural canal , but not below. 
Between the neural canal and the chorda the sides of the neural arch just 
meet, but they do not coalesce. In the centre of this “ suture ” there is a 
round opening, uniting the neural canal with the chorda. The sides of 
the neural arch, enclosing the neural canal, first form two parallel, 
perpendicular walls and then suddenly expand into a broad flat basis. 
The lower surface of this basis is rough, indicating that it was covered 
with cartilage. The neural canal is very narrow, having a diameter of 
about 1 mm. The neural spine hangs slightly backwards out of the 
perpendicular. It is a broad, flattish bone, broader at the top than at its 
base and with a longitudinal ridge along its middle on either side. The 
condition at the lower end of this ridge is not quite clear. In some 
instances it seems to flatten down to the common surface of the spine, in 
others it seems to split into two ridges directed towards the zygapophyses. 
The free end of the spine is thicker than the rest and ends with a flat 
and smooth surface. The lower anterior zygapophyses project outwards 
without seeming to deviate from the horizontal. They meet the lower 
and outer side of the higher posterior zygapophyses of the foregoing 
vertebra. The anterior zygapophyses are larger than the posterior ones. 
The dorsal vertebrae have developed strong diapophyses (shown 
clearly in three vertebrae of PI. XVII), which diminish in length towards 
the sacrum where they cannot be identified from the basal plate of the 
neural arch. The articulation surface of these diapophyses lies all along 
their front and outer margins. 
The single sacral vertebra (PI. XVII and XIX, fig. 2) differs slightly 
from the adjoining ones. The articulation surface of the hypocentrum 
for the rib is larger than in the others and the base of the neural arch 
forms strong diapophyses. 
The exact number of caudal vertebrae cannot be ascertained either. 
There is one caudal piece (PL XX) with 28 vertebrae, but the end is 
missing, and this can easily have had another ten. The proximal end of 
this piece is not in connection with the sacral vertebra and its first 
vertebra has distally completely coalesced chevron bones, In another 
4 
