92 PLEISTOCENE MAMMALIA. 



the one is a good deal larger than that of ordinary lion, and consists of the centrum, 

 the neurapophysis with the root of the neural spine, the right prezygapophysis with post- 

 zygapophysis and the roots of the diapophyses : the other is more mutilated, and 

 does not differ in size from the corresponding vertebra in lion. 



The second dorsal in Felis is intermediate in form between the first and the remainder 

 of the typical dorsal series. The centrum (c) is elliptical in section, and the prezyga})ophyses 

 {az) are widely separated, as in the cervicals, but the postzygapophyses {^oz) are nearer 

 together under the neural spine, as in the other dorsals. We have one specimen of this 

 vertebra from Sandford Hill in the Taunton Museum. It is all but perfect, and we have 

 devoted an entire plate (PI. XV) to its illustration. 



We can detect no difference whatever between it and that of the lion or tiger, except 

 size and the length of the neural spine {ns) ; but the amount of inclination of the latter 

 is exactly the same as in the larger Eeles. In comparing several skeletons of these species, 

 we have not the neural spine of exactly the same form in any two ; in most the form is 

 slightly sigmoid, the bend backwards near the summit being very decided. 



The large mounted skeleton of the lion in the College of Surgeons has the neural 

 spine of this vertebra precisely of the form of that of our fossil. It is, however, shghtly 

 shorter. We have also met with vertebrae of the jaguar ^ and leopard which closely 

 resemble it in general appearance. 



The greater length of this spine would indicate greater carrying power in the Felis 

 spelcsa than in ordinary lion and tiger. There is also another slight difference ; at the 

 posterior base of the spine (PI. XV, fig. 1") there is a deep and somewhat large 

 hollow in the fossil. This occurs occasionally in all the dorsal vertebrae of carnivora, and 

 we have found precisely the same formation in the jaguar, and it is evidently only varietal 

 in importance. 



In comparing the vertebra with the corresponding one of the bear, we find that the 

 centrum of the latter is proportionally much wider and shorter. The diapophysial arti- 

 culation {d) for the tubercle of the rib is much larger, the posterior parapophysial articulation 

 (fig. \',pe) for the head of the rib is much flatter, larger, and more even with the epi- 

 physial articulation. The diapophyses (figs. 1, 1', 1", d) are much wider, and with the 

 prezygapophyses {az) form a much larger mass, furnished like the first dorsal, with meta- 

 pophyses, which do not appear in Eelis. The prezygapophyses are also flatter, and do 

 not turn upwards at any part towards the neural spine, as is the case in this vertebra ; and, 

 as in all the Eelis we have met with, the post-zygapophyses {pz) are also flatter, further apart, 

 and afford a less firm lock with the third vertebra in bear than in Eehs. The posterior edges 

 of the diapophyses are formed by sharp ridges between the pleurapophysial articulations and 

 the post-zygapophyses ; but this ridge in our specimen, as in all recent Feles, forms a short 

 concave curve, where as in bear the edge forms a long, straight, or rather convex ridge. 



^ See also plate of skeleton of jaguar, De Blainville, 'Ost. Felis,' pi. iii. 



