90 BRITISH FOSSIL REPTILES. 



siologically to tlie functions of the diverging appendage which is developed into a 

 potent locomotive member. This hmb appertains properly, as the pi-oportion contri- 

 buted by the ischium to the articular socket and the greater breadth of the pleura- 

 pophysis show, to the second sacral vertebra ; to which the ilium chiefly belongs. 



The first caudal vertebra, which presents a ball for articulating with a cup on the 

 back part of the last sacral, retains, nevertheless, the typical position of the ball on the 

 back part of the centrum ; it is thus biconvex, and the only vertebra of the series which 

 presents that structure. I have had this vertebra in three different species of extinct 

 Eocene Crocodilia. In the Crocodilus toliapicus, PI. 5, fig. 7 ; in the Croc, champso'ides, 

 PL 3v^, fig. 10; and in the Crocodilus Ilastinf/sia;, PL 1 D, fig. 7. 



The advantage of possessing such definite characters for a particular vertebra is, 

 that the homologous vertebra may be compared in different species, and may yield 

 such distinctive characters as will be hereafter pointed out in those of the three species 

 above cited. 



The first caudal vertebra, moreover, is distinguished from the rest by ha\'ing no 

 articular surfaces for the hfemapopliyses, which in the succeeding caudals form a 

 haemal arch, like the neurapophyses above, by articulating directly with the centrum. 

 The arch so formed has its base not applied over the middle of a single centrum, but 

 like the neural arch in tlie back of the tortoise and sacrum of the bird, across the 

 interspace between two centrums. The first hsemal arch of the tail belongs, however, to 

 the second caudal vertebra, but it is displaced a little backwards from its typical position. 



The detached centrum of a caudal vertebra, besides being more slender and com- 

 pressed, is distinguished from those of the before-described vertebrae by the two 

 articular surfaces at the posterior border of their under surface. The zygapophyses 

 become vertical as far as the sixteenth or seventeenth, beyond which the two posterior 

 zygapophyses coalesce in an oblique plane notched in the middle, which is received 

 into a wider notch at the fore pai't of the neural arch of the succeeding vertebra. The 

 sutures between the pleurapophyses and diapophyses are maintained during a long 

 period of the animal's growth, and demonstrate the share which these two elements 

 respectively take in the formation of the transverse process. So constituted, these 

 processes progressively decrease in length to the fifteenth or sixteenth caudal vertebra, 

 and then disappear. The neural spines progressively decrease in every dimension, 

 save length, which is rather increased as far as the twenty-second or twenty-third 

 vertebra, beyond which they begin again to shorten, and finally subside in the tei'minal 

 vertebrae of the tail. 



The caudal hcemapophyses coalesce at their lower or distal ends, from which a 

 spinous process is prolonged downwards and backwards ; this grows shorter towards 

 the end of the tail, but is compressed and somewhat expanded antero-posteriorly. 

 The hsemal arch so constituted has received the name of ' chevron bone.' 



A side view of the body of a middle caudal vertebra of the Crocodilus toUapicus is 



