STRUCTURE AND RELATIONSHIPS OF OPISTHOCCELIAN DINOSAURS. 185 
closely with the type of MW. grandis so far as it will admit of determi- 
nation. The Morosauroid sacra thus appear to fall into two groups. 
The first, as indicated by the type of M. lentus.and No. 5384 of this 
Museum, has the dorso-sacral free. The second, as represented by 
the type of JZ. grandis and the Kansas University specimen, has the 
caudo-sacral free. A more important difference is that in the sacra of 
the former type the transverse axis passes between sacrals 11 and Iv, 
while in the latter it passes through sacral m1 as in Apatosaurus and 
Diplodocus. If this feature proves constant it would indicate generic 
differences. 
From the above it will be observed that the sacra of Afpatosaurus 
and Diplodocus, so far as can be traced from the young animal to the 
adult, develop along similar lines. In Apatosaurus sacrals 11, 11, and 
Iv coalesce first as is shown by the two young specimens (Figs. 5 and 
10), by the closer union of all the elements of these same vertebre in 
the adult specimens (Figs. 7, 8, and 9), and by the noticeable tendency 
of articulating parts to atrophy from disuse. Sacral 1 usually unites 
next, and v last, as shown by Figs. 6, 7, and 8, although this order 
may sometimes be reversed, as is shown by Fig. 9. Coalition may be 
carried even farther, as appears by the union of Caudal 1 in Fig. 7 and 
presacral 1in Fig. 9. In Dzplodocus the same order is noticeable with 
the exception pointed out by Mr. Hatcher (Fig. 6), in which sacral 1 
and the spine of 1v remain free in a specimen apparently adult, while 
sacral v is codssified ‘by its centrum. However, this is a variation 
similar in kind to that noted in Afpatosaurus. 
The above deductions carry out, for these two genera, the theory 
advanced by Osborn and supported by Williston, namely: That sacrals 
II, 111, and Iv represent the primitive sacrum. But how far back in the 
- history of the group these conditions may have held, remains a ques- 
tion. If we assume that the Opisthoccelia are derived from a bipedal 
ancestry, which were in turn derived from a crawling ancestor, we 
may explain the condition as follows: 
It has been observed that the transverse axis of the pelvis passes 
through the ribs of sacral 11. This would be the natural point for the 
pelvis to attach to the vertebral column in a terrestrial animal where 
the thrust of the femuris upward. The fact that sacral 111 has its 
centrum, ribs, diapophyses, zygapophyses, and neural spine most 
reduced indicates that it formed the primitive sacral center. Such an 
attachment doubtless served the primitive crawling reptilia, but as 
specialization in terrestrial habits progressed, and there was need of a 
more rigid pelvic structure, sacrals 11 and Iv probably became con- 
joined, one after the other. In this way the sacrum developed by alter- 
