OSTEOLOGY OF CARNIVOROUS DINOSAURS. 



99 



longer than high, whereas in the upper Cretaceous genera mentioned above the 

 anterior centra are as high as long. 



In passing backward from the sacrum there is a gradual shortening of the centra, 

 and they become more and more cylindrical in form, though all except the terminal 

 ones have flattened sides. The vertebral centra are deeply concave on both sides 

 and below, the anterior ones being shallowly grooved longitudinally on the ventral 

 side as shown in figure 2, plate 22. 



The neural spines are unusually high, relatively narrow antero-posteriorly, 

 with flattened sides and ending above in a slightly expanded rugose end. The 

 expansion is not only slight in transverse diameter, but expands pronouncedly fore 

 and aft. Proceeding posteriorly the shaft of the spinous processes narrows more 

 rapidly than they shorten in length, thus making them tall and slender. Their 

 point of origin also shifts from the middle 

 of the centrum to a point entirely above 

 the posterior half. They rapidly de- 

 crease in height on the median part 

 of the series, the last distinct spine 

 being on the thirty-first caudal. On 

 account of their crushed condition the 

 spines posterior to the twenty-eighth do 

 not show in the lateral view in plate 30. 

 The transverse processes on caudal 

 1 are directed backward at an acute 

 angle to the centrum, but do not unite 

 with the ilia, as apparently indicated 

 in figure 1, plate 21. These proc- 

 esses on the succeeding vertebrae are 

 long, broad, thin vertically, and termi- 

 nated by straight ends without any 

 especial expansion. On the second 

 caudal tney are directed backward fig. 57.— fifth caudal vertebra and chevron of cerato- 

 at an angle of 45°, but they gradually saurus nasicornis marsh, type. no. 4735, u.s.n.m. 1 



n . . . . nat. size. Viewed from the left side. (After Marsh.) 



assume a more direct outward position 



at the same time; diminishing in size, they end abruptly, the last one appearing 

 as a thin pointed process on the thirty-third vertebra, whereas in Gorgosaurus 

 libratus Lambe they end with the fourteenth caudal from the sacrum. In fact, 

 I know of no other carnivorous, or, for that matter, herbivorous dinosaur, where 

 the transverse processes persist so far back on the tail. On the anterior portion of 

 the caudal series the transverse processes have their origin on the neural process, 

 and above the neural canal, but passing backward they gradually shift their 

 position downward, so that somewhere between the twenty-second and the twenty- 

 sixth it has changed its position to the center of the side of the centrum. 



The prezygapophyses are relatively short throughout the anterior two-thirds of 

 the tail, but on the distal third they rapidly lengthen, extending well beyond the 

 end Of the centrum as two slender finger-like processes, which clasp between them 



