84 PROP. H. G. SEELEY ON^ 



the correspondence extends to the lateral compression of the cen- 

 trnm and neural arch, inclined articular face of the centrum, mode 

 of attachment of the ribs, the convex external surface of the neural 

 arch, almost total suppression of the neural spine, thin texture of 

 the bone, and internal chambers in the neural arch and centrum. 

 Yet this strong evidence of affinity does not, I think, amount to 

 generic identity. The following points seem to me important dif- 

 ferences : — Eirst, in this specimen the anterior part of the neural arch 

 is much higher, ascends as it extends backwards, shows a large 

 cavity above the neural canal, has no trace of a neural spine, 

 and much more extensive diapophysial attachment for the ribs. 

 Secondly, the chambers in the neural arch of Ccelurus are divided 

 by a vertical median septum ; while in this specimen there are two 

 important median chambers in addition to the lateral chambers. 

 I attach less importance to the differences in the lateral walls 

 of the neural canal, in the character of the forward extension of the 

 postzygapophysial ridge, and in the form and condition of the arti- 

 cular face of the centrum, or to the other minor differences which are 

 evident on comparison of the figures. But the sum of the differences 

 seems to define a genus, but so close to Coelurus that other parts 

 of the animal may be expected to conform to the general con- 

 struction of that type. 



The sacrum of Thecospondylus being only known from a cast 

 (Quart. Journ. Oeol. Soc. vol. xxxviii. pi. xix.), from which the 

 external film of bone has almost disappeared, is compared with some 

 disadvantage ; and the resemblances which can be detected are 

 limited to the thinness of the investing layer of bone, the smooth- 

 ness of its internal surfp.ce, the elongation and lateral compression 

 of the vertebrae, and a certain approximation in their general form. 

 The difference which is most remarkable is the absence from the cast 

 of Thecospo7idylus Horneri of indications of films of bone such as 

 might have been expected to divide the neural arch from the cen- 

 trum, or the sacral ribs from both, or other evidence of internal plates, 

 such as are seen in the present specimen. But since such structures 

 must have existed, their absence may be attributed to fossilization, 

 and is no evidence of organic difference. It is because these animals 

 are both British Wealden fossils that I attach importance to resem- 

 blances which show an approximation which is consistent with 

 generic identity, especially as there is no known character, direct or 

 inferential, which would justify theii' location in distinct genera. 

 If I am warranted in referring this vertebra to the genus Theeo- 

 spondylus, it follows that the identification makes the type of sacrum 

 known, which might be expected in such a genus as Coelurus, and 

 completes a knowledge of the vertebral column in the family which 

 that genus evidences. 



I should infer that Thecospondylus Daviesi had the head elon- 

 gated, with long slender jaws ; that the neck was over half a metre 

 long, that the back and sacrum together were less than a metre and 

 a third long ; so that the measurement from the tip of the snout to 

 the extremity of the tail may have been under three metres. The 



