211 ^ 
TERTIARY VERTEBRATA OF THE FAY0M. 
postcro-external main cusp. The inner side of these comparatively small molars is 
considerably higher than the outer. 
The dimensions (in centimetres) of the mandible and teeth just described are: 
Length from anterior end of symphysis to end of molar series . 15'5 
Greatest width of symphysis 5-5 
Length of symphysis Q-T) 
Depth of mandibular ramus beneath ?n. 3 4-7 
The dimensions of the teeth are : — 
?«. 1 
7 / 1 . 2 
m. 3 
Length. 
1-6 
1-9 
0-0 
Width. 
1-3 
]-5 
1-4 
Vertebral Colimin. — On account of their massive character the ribs and vertebrse of 
JEosiren are among the commonest fossils in the Qasr-el-Sagha beds. For the most 
part, they only occur scattered in small groups or as single specimens, but occasionally 
a considerable number are found in association. The best of these sets is here noticed ; 
it includes fifteen more or less nearly perfect vertebrae associated with part of a 
skull and an vs innominatimi. 
'fhe atlas is wanting. The axis is greatly abraded and broken, but in the form of 
its odontoid process and articular surfaces for the atlas it was evidently closely 
similar to that of Ilalitherkmi. In the third ('?) cervical vertebra the centrum is 
rectangular in outline and much wider than high ; it is much less shortened antero- 
posteriorly than in later forms ; its ends are smooth and flat. The transverse 
processes are broad and long, and are perforated at their base by the vertebrarterial 
canal. Another cervical vertebra of the same series is very similar. The number of 
the cervicals is unknown. 
The thoracic and lumbar vertebrm are like those of llalitherium, but the hasmal 
ridge is very slightly developed, and in the thoracic region the zygapophyses are better 
developed and the interlocking of the vertebrae is closer. 
One of the vertebrae (FI. XX. fig. 4) in the series referred to presents some peculiar 
characters. It appears to be a first lumbar ; the anterior firce of the centrum narrows 
somewhat towards its neural border, the neural arch is much higher than in the other 
vertebrae, and the anterior zygapophyses are also considerably elevated. 'J'he transverse 
processes rise from the ventro-latcral borders of the centrum and curve backwards 
and downwards. The whole vertebra differs so far from the others tliat perhaps 
its association with them may be a matter of accident, and it may actually belong to 
some other animal. 
A sacral vertebra (PI. XX. fig. 6) is preserved in wliich the arch is comparatively 
low with a short neural spine ; the centrum is transversely oval, with a slight median 
