PAL^OMASTODON. 145 



border is in the same line with the hypapophysial ridge on the centrum : on its 

 lower surface it bears an oval facet for articulation with the ventral bar of the atlas, 

 this facet being sharply defined and separated from the lateral articular surfaces, which 

 are oval in outline and make an angle of about 115° with the long axis of the odontoid 

 process. The centrum is very long compared with that of the later Proboscidean axis ; 

 it bears a median hypapophysial ridge, which deepens posteriorly in such a way that 

 the posterior articular surface is greatly deepened ventrally. There are small transverse 

 processes (t.p.) perforated at their base by a large vertebrarterial foramen. The neural 

 arch is very high, but the precise outline of the spine {n.sp.) cannot be determined, 

 owing to the imperfection of the specimens ; it can be seen, however, that it was high 

 and broad, and overhung the arch both before and behind. A smaller axis vertebra, in 

 which the neural spine is nearly complete, shows that it consisted of two plates meeting 

 in front at an acute angle and enclosing posteriorly a deep fossa for the insertion of 

 muscles or ligaments. This axis differs from the ordinary Proboscidean type in the 

 following points : — (1) the large size of the odontoid process, which in the later forms 

 is reduced to a mere knob with its articular surface continuous with that of the 

 centrum which extends beneath it ; (2) the much greater length of the centrum ; 

 (3) the less massive, but much higher, neural spine. On the other hand, it is 

 very similar to that of Mosritherium. The neck in Paloeomastodon was probably 

 considerably longer and more mobile than in the later Proboscidea. 



Fore Limb. — An incomplete left scapula collected in the same locality as the type 

 mandible is the only evidence of that bone yet discovered. It is very closely similar 

 to the scapula of Elephas, differing only in the rather slighter prominence of the 

 coracoid process and in widening out less rapidly posteriorly above the glenoid cavity, 

 the whole blade being probably rather less expanded above than in JElephas. The 

 glenoid cavity is an elongated oval, the prescapular fossa is extremely narrow, and 

 the process projecting backwards from the middle of the spine in Elephas here seems 

 to be situated a little nearer the glenoid end. 



In the case of the humerus (PI. XVI. figs. 2, 2 a) also the only specimen that 

 can be definitely determined as belonging to PaloBomastodon headnelli, was collected 

 near the type mandible. In its general character this bone is like that of the later 

 Proboscideans, but is somewhat stouter and broader in proportion to its length, and 

 approaches more nearly some specimens of the humerus of Mastodon with which 

 it has been compared. It is also very similar to the humerus figured by de Blainville 

 (' Osteographie,' Atlas, vol. v. pi. xiii., Elephas) as that of Tetrahelodon angustidens. 

 The deltoid crest {d.) is greatly developed and strongly bent over to the outer side 

 of the bone. The very prominent and weU-marked supinator ridge (sp.r.) is shorter 

 than in the Elephants, though still very large ; its flat outer surface is slightly 

 inclined to the long axis of the shaft. The inner trochlea is much larger than the 

 outer, and, in fact, seems to be larger proportionately than in the Elephants ; but in 



u 



