308 TERTIAET VEETEBEATA OE THE FATtM, 



zygapophyses which they face. The posterior zygapophyses {p.z.) are strongly 

 prominent trihedral processes, the posterior face of which is nearly vertical, the lower 

 bears the oval articular surface, while the upper is continuous with the upper surface 

 of the arch ; their upper angles are continued as ridges to the base of the neural spine ; 

 on the posterior face there is a shallow pit (/'.). The zygantral facets form the 

 sides of the deep zygantrum {za.) in the posterior end of the arch ; they are oval in 

 outline and inclined to the surfaces of the posterior zygapophyses at a very acute 

 angle. The transverse processes {t.p.) are very short and stout ; at their extremities 

 they bear the vertically elongated rib-facets, which are convex in all directions above, 

 but concave from above downwards below, as in Python ; their lower angles are on 

 a level Avith tlie ventral border of the centrum. The trefoil- shaped neural canal is 

 very small, relatively much smaller than in the recent genus. 



A rib associated with the type vertebrae shows that, while the articular end is just 

 as in Python, the shaft is more compressed from before backwards. 



The posterior portion of a left ramus of a mandible (PI. XXVI. fig. 3) of a reptile 

 found close to the type vertebrae may belong to this species. The specimen includes 

 the angular and articular region, in front of which it is broken away. The articular 

 surface (art.) is convex from side to side, and concave from before backwards ; it is 

 wider in front than behind, where it terminates in a point ; its anterior border runs 

 inwards and forwards, so that the whole surface is somewhat oblique, much as in 

 Python. The angle (ang.) is thick and not very prominent; it differs from that of 

 the mandible of Python in being less sharply marked off from the region in front. 

 The inner plate of the jaw, which if complete would run up anteriorly to the coronoid 

 process, is broken away, but enough remains to show that probably it was separated 

 from, the outer plate by a deep channel as in Python. The apparent sculpturing of 

 the surface of some of the bones (see PI. XXVI. fig. 3 a) appears to be due merely 

 to the etching of the surface by drifting sand. If this mandible did actually belong to 

 Gigantophis garstini, then, judging from its size compared with any vertebrae found, it 

 would seem that the head in this animal was relatively much larger than in recent 

 snakes ; but more satisfactory material is necessary before any definite conclusion 

 can be arrived at. 



C. 10022. Associated series o£ about twenty vertebrae and portions of two ribs. Type specimen, 

 described and figured loc. cit. supra; also figured PI. XXVI. fig. 1. The dimensions 

 (in centimetres) of one of these vertebrse are : — 



Grealeat height from top of neural spine to end of bypapophysis. 5 app. 



„ width (between ends of transverse processes) . . . 5'7 



Width between outer angles of posterior zygapophyses . . . 5-5 



,, „ upper angles of zygospbeue 2-1 



,, of iirticular cup of centrum 2-3 



Height of articular cup of ceutrum 1'9 



