140 BRITISH FOSSIL REPTILES. 



by the superadded zygosphenal (zs, fig. 6) and zygantral {aa, fig. 7) articulations, but 

 by the solidity of the zygosphene, by the size and form of the centrum, by those of its 

 articular cup (c, fig. 6) and ball (c, fig. 7), and of its hypapophysis {/i) ; and also by 

 the size and prominence of the diapophysis (d). The largest vertebrse (e. g. PL 2, 

 figs. 5-8, and PI. 3, figs. 16, 26, 27, 28) probably from about the middle of the body, 

 as compared with the vertebrae from the same part of the skeleton of a Fython Sebee, 

 twenty feet in length, are longer in proportion to their breadth, and the cup and ball 

 of the centrum are larger ; the hypapophysis (//) is more produced, and there is a 

 second smaller hypapophysis close to the anterior part of the under surface of the 

 centrum, which in most of the large vertebrae is connected by a ridge with the hinder 

 and normal hypapophysis ; but in a few vertebrae is not so connected. The articular 

 cup and ball are less obliquely placed upon the extremities of the centrum, being 

 nearly vertical (compare fig. 5 and fig. I, c). The rim of the cup is sharply defined, 

 and is more produced from between the bases of the diapophyses ; a deeper and 

 narrower chink intervening than in the Python. The transverse diameter of the cup 

 {c, fig. 6) is greater than that of the zygosphene (ib., zs) — a proportion which I have 

 not found in the vertebrae of any existing genus of Serpent, in which the base of the 

 zygosphene always equals at least the parallel diameter of the articular cup. The 

 articular part of the diapophysis is more produced outwards and less extended 

 vertically in JPalmopUs than in Python, and" it is uniformly convex ; a ridge is 

 continued from its upper end obliquely forwards to, but not beyond, the apex of the 

 anterior zygapophysis [z), forming the angle between the lateral and anterior surfaces, 

 whilst the horizontal articular facet forms the third surface of that three-sided conical 

 process. In the Python the non-articular part of the same zygapophysis is convex, 

 and the process is much more extended outwardly ; the proportions of the zygapophysis 

 in the Palaophis more resemble those in the Coluber and Hydrus, but differ from these, 

 as also from Naja and Crotalus, in the non-extension of the diapophysial point beyond 

 the articular surface. 



A ridge or horizontal rising of the bone extends from the anterior to the posterior 

 zygapophysis, but is more or less blunted or subsides midway, and is by no means so 

 produced outwards as in Python ; in this respect more resembling that in Coluber and 

 Hydrus. Below the middle of this ridge, on a level with the upper surface of the 

 centrum, there is a short, nearly parallel rising in Palceophis (fig. 5). The zygosphene 

 (fig. 6, zs) is slightly excavated anteriorly, and shows no trace of the tubercle which 

 characterises the middle of that surface in the Python (fig. 2) ; it is also broader in 

 proportion to its height. But perhaps the most characteristic feature of the vertebra 

 of the Pal(Bophis is the peculiar production of the posterior border of the neurapophysis 

 into an angle («, fig. 5) directed upwards, outwards, and backwards, and this is 

 common to all the species ; there is no trace of this process in the Hydrus (fig. 25), 

 and the nearest approach to it which I have hitherto met with among existing 



