Vol. 56.] DICRANOZYGOMA LEPTOSCELTJS. 651 



neural canal, are seen to be in close contact one with the other ; 

 but in the middle and lower part of the back there are interspaces 

 between the adjacent centra, which I regard as indicating probably 

 the existence of intercentra in that part of the vertebral column. 

 A similar condition is seen in Gynognathus and Pareiasaurus, when 

 intercentra are developed on the visceral surfaces of the vertebrae. 

 The interspace between the centra appears to be least at the base 

 of the neural canal, and to augment laterally, permitting a large 

 amount of lateral curvature of the body. Four vertebrae extend 

 over 3| inches. 



The neural arches are at first small. The neural spine is small, 

 vertical, strong, wedge-shaped, sharp in front, flattened behind, 

 terminating downward in a close-set V-shaped pair of post-zyga- 

 pophyses. The transverse processes are given off at the base of the 

 neural spine and directed outward, slightly upward, and backward, 

 rounded from front to back, where the measurement is \ inch, and 

 expanded at the extremity, which is not unlike the corresponding 

 surface in an Ornithosaur. The articular face for the rib looks 

 downward and outward on its inferior surface. The transverse 

 measurement over the processes in an early dorsal exceeds 1 inch. 



Only two early dorsal vertebrae have the neural spines preserved. 

 They are followed by eight from which the neural spine and roof 

 to the neural canal is lost. Thus there is a marked contrast 

 between the aspect of the early dorsal and lumbar vertebrae, in which 

 the upper part of the neural arch is preserved, due to the circum- 

 stance that the transverse processes rapidly become shorter, and 

 disappear by ascending the side of the neural arch and becoming an 

 oblique, sharp, lateral ridge which extends upward and backward, 

 fully | inch long, terminating in a prominent rounded tubercle (ex- 

 ternal to the post-zygapophyses) which has no relation to the rib. 

 These processes form practically the oblique sides of the neural 

 arch, and become merged in the neurapophyses, with the oblique 

 facets of the zygapophysis between them in front and behind. The 

 pre-zygapophysial facet has become much larger than in the early 

 vertebrae, but still looks inward and upward. The neural spine, 

 instead of being elevated and free, has become flattened, and forms 

 the median ridge of the neural arch, which has become nearly hori- 

 zontal, with the two flattened sides sloping outward from it. It 

 terminates backward in a median point, and the lateral areas are 

 widened a little at the posterior angles by the post-zygapophyses. 

 The last lumbar vertebra has this superior part of the neural arch 

 shorter, and the median ridge is more elevated. 



The sacrum is entirely hidden beneath the iliac bones. I 

 suppose that there may have been four sacral vertebrae, because 

 the length (3 inches) corresponds to the length of the four lumbar 

 vertebrae. There cannot have been fewer; it is improbable that 

 there are more. The last sacral rib was slender, and less than 

 1 inch long; the first is more than 1 inch long, and not stout. 

 The last shows an expansion at the external extremity, which 

 articulated with the ilium. 



