132 TEETIAET VEETEBRATA OF THE EATtTM. 



condyles. Its body is keeled ventrally and it unites with the basisphenoid at about 

 the level of the glenoid cavity for the mandible : the line of union with the basi- 

 sphenoid [bsp.) is marked by a fairly prominent transverse ridge. The lateral border of 

 the median limb is separated from the anterior border of the lateral portion by a 

 sharp notch, through which the hypoglossal nerve probably passed, there being, as 

 in the later Proboscidea, no condylar foramen. 



With the possible exception of a small portion of their inner ends, the whole of the 

 occipital condyles are formed by the exoccipitals {exo.). The condyles themselves are 

 much larger and project further behind the skull than in Elephas, in which their 

 articular surface seems to look mainly downwards and to project very little beyond 

 the rest of the exoccipital bones. Here, on the other hand, the condyles are so 

 sharply marked off from the rest of the bone as to be almost pedunculate. Their 

 articular surface extends far on to the dorsal surface, and they are very strongly convex 

 from above downwards — in fact, roughly speaking, they may be said to form about 

 two-thirds of the surface of a cylinder — and at the same time they are slightly convex 

 from side to side. From the form of the condyles, therefore, it would appear that the 

 range of movement allowed to the skull in an up-and-down direction must have been 

 very great, while that from side to side was comparatively restricted. 



The foramen magnum {f.m.) is oval in outline and does not look downwards so 

 much as in Elephas, mainly owing to the fact that the posterior border of the basi- 

 occipital is less deeply notched. Ventrally and laterally the surface of the exoccipitals 

 runs forwards in a gentle curve to join the post-tympanic flange (j)ty.) of the squamosal. 

 On the ventral surface near the union with the squamosal the exoccipital bears a blunt 

 prominence (pp.), which is the only representative of the paroccipital process, and is 

 therefore homologous with the thin plate-like ventral process of the bone described 

 in the skull of Moeritherium. The upper portion of the bone above the condyles and 

 the foramen magnum slopes forwards, and the two elements meet in the middle 

 line, separating the supraoccipital from the foramen magnum by a distance of about 

 5 centimetres. 



The precise boundaries of the supraoccipital [soc.) are indistinct. In the middle 

 line it, like the upper portion of the exoccipitals, ,is inclined forwards ; above it is 

 hollowed out by a great median fossa, the floor of which is greatly roughened and is 

 obscurely divided into halves by a slight median ridge. This fossa is one of the most 

 characteristic peculiarities of the skull of the later Elephants, but in Palmom.astodon 

 it is far more sharply defined, probably, for the most part, on account of the smaller 

 development of the cellular tissue of the rest of the bone. The sides of the fossa are 

 formed by broadly rounded vertical ridges. The superior and supero-lateral portions 

 of the bone occupying the position of the lambdcidal crest are unfortunately broken 

 away, but it can be seen that in conjunction with the neighbouring squamosals and 

 parietals the body of the bone is excavated by a series of large cellular cavities, 



