120 E. S. Lull — Cretaceous Armored Dinosaur j 



tributed throughout this niid-torso region of the body. 

 The nodular intercostals average about 80x55 mm. in 

 size, the greatest diameter lying parallel with the ribs. 

 The nodules are well rounded eminences, elliptical in sec- 

 tion, and rising some 15 ±: nnn. above the general level of 

 the scute. The nodular scutes thus form regular trans- 

 verse rows across the creature 's back. They seem to have 

 formed longitudinal rows as well, as in the crocodiles, and 

 at least nine such longitudinal rows existed on either side 

 of the midline, probably more in this region of the body 

 (see restoration, fig. 7). Toward the distal ends of the 

 ribs, which are, however, not complete, the scutes of the 

 subdermal layer become visible and one can view their 

 passage dorsally beneath the dermal scutes {sd). The 

 scutes give an impression of crowding in the longitudinal 

 axis of the body, as though the force that tended to 

 converge the ribs distally had also wrinkled the skin. It 

 would seem, therefore, as though, as in the alligator, the 

 nodular (or keeled) scutes are dorsal, the flanks and 

 belly being protected by smaller elements. The homology 

 is not, however, precise, for in the crocodile the subdermal 

 armor is lacking and the lateral and ventral protective 

 elements are entirely cutaneous. In Nodosaurus a cuta- 

 neous investiture was of course present, but seems to 

 have been in a larger degree supported by osseous dermal 

 and subdermal scutes. 



In the group of vertebrae described above, the proximal 

 ends of the ribs only are preserved. They bear compar- 

 able dermal scutes similarly arranged, with, however, no 

 trace of the subdermal elements. Over the vertebrae 

 themselves the nodular scutes are larger, somewhat 

 hexagonal, although irregularly so, averaging 80x86 mm. 

 in size, with the greatest diameter fore and aft. There 

 is no median row, single as in the turtle, but two, one on 

 either side of the spinous processes, form the neural 

 series (pi. I, fig. 1, np). 



Sacral armor. — On the sacrum the armor plates are 

 very obscurely defined, but are in the main hexagonal, 

 although very irregular in outline, the two neural rows 

 being continuous with those just mentioned. Whether 

 they w^ere separated during life by the now visible 

 summits of the spinous processes is not clear. The armor, 

 as has been said, did not form a complete shield as in 

 Polacantlius, but was confined to a longitudinal belt 

 overlapping only the sacral diapophyses and not the ilia. 



