88 FOSSIL REPTILIA OF THE 



The centrum is more or less antero-posteriorly compressed (PL XXII, figs. 1 to 7, 9, 

 14), with concave terminal articular surfaces (ib., fig. 6) not intercommunicating ; on each 

 side of the shallow myelonal canal (PI. XXII, figs. 2 and 4, m ) is the deeper, usually 

 triangular, articular surface (np) for the neurapophyses ( n ). These, in each vertebra, converge, 

 and, save in the atlas (PI. XXI, fig. 1, n ), coalesce at their summits with each other and 

 with the neural spine (ib. figs. 2, 3, 4, n , ns)- In most Fishes the neural arch coalesces 

 with the centrum, as in Cetaceans ; its separate state is a saurian, chiefly crocodilian, 

 modification ; it is such in the Ichthyosaurs, and adds to the power of inflecting the spine 

 vertically, as in the specimen (PI. XXIX, fig. 2). 1 



Most of the neurapophyses interlock by means of coadapted zygapophyses (PI. XXI, 

 fig. 6, z,z'). The hsemapophyses are developed beneath the abdominal ribs (ib., 

 fig. 2, h,h',) and beneath the bodies of most of the caudal vertebrae (ib., figs. 4, 5, h) ■ 

 they are always distinct from their centrum (c), and do not coalesce below with each other, 

 or with a haemal spine. The hypapophyses remain detached in the first two or three 

 vertebrae (ib. fig. 1, hy), and have advanced to the interspace between their own and the 

 antecedent centrum. That of the atlas (PI. XXIII, fig. 5, fya) is wedged between it and 

 the basioccipital ( ) ; that of the axis (ib., hyx) between it and the atlas, and so on (ib., 

 hy, 3). Conybeare, who first noticed this structure, describes it as follows : — " We have 

 only seen the inferior piece or body (if it can be so called) of the atlas ; and the odontoid 

 process (which in all reptiles forms a distinct piece) of the axis ; they very nearly resemble 

 those of the turtle." 3 



In the trunk the centrum of the atlas (PL XXIII, figs. 2 — 5, c a) is the most modi- 

 fied of that series of vertebral elements. Its fore surface (ib., fig. 2) presents at its upper 

 two thirds a concavity (c a), occupying the medial two fourths of its transverse extent, 

 the cavity gradually changing to convexity (b) in the lateral fourth. Beneath this 

 smooth concavo-convex articular surface is a rough, flat, triangular surface {t), inclining 

 from its upper base backward. The upper joint-surface ( c a), is for the basioccipital 

 (PL XXIII, figs. 1, 2, fig. 5, 0, in dotted outline), the lower one {t), is for the hypapophysis 

 (Fig. 5, hya). The hind surface of the atlantal centrum (PL XXIII, fig. 4) is flat, and with 



1 More extreme and abrupt vertical flexures, shown in two specimens in the British Museum, may be 

 posthumous, due to disturbance of the decomposing carcase prior to final burial in the Liassic mud, in and 

 with which the skeleton subsequently became petrified. 



2 'Trans. Geol. Soc.,' vol. v, 1821, p. 574. The homology of the "odontoid process" as the 

 "centrum of the atlas," and that of the anthropotomical " body of the atlas " with the hypapophysial part 

 of that vertebra, is shown in my " Description of the Atlas, Axis, and Subvertebral Wedge-bones in the 

 Plesiosaurus," 'Annals and Magazine of Natural History,' vol. xx, 1847, p. 217, figs. 1 — 6. 



In 1835 Sir P. de M. Grey Egerton communicated to the Geological Society his discovery of not 

 only Conybeare's " inferior piece of the atlas," but the homotypal parts of the two succeeding vertebrae 

 (' Proceedings of the Society,' vol. ii, No. 41, p. 192), and subsequently gave a detailed description, with 

 figures, of these parts under the name of " subvertebral wedge-bones ; " ' Trans. Geol. Soc.,' 2nd series, 

 vol. v, 1836, p. 187,pl.xiv. 



