hatcher: diplodocus (marsh) 31 



three, four, or five vertebrae according to tiie individual conception as to which 

 should be considered as sacral vertebrae. If the sacrals are made to include all 

 those vertebrae that, though formerly belonging to the posterior dorsals or anterior 

 caudals, have laterally become so modified as to function as sacrals by affording 

 support to the ilia either by bearing true sacral ribs or by means of greatly ex- 

 panded transverse processes, or by both these methods, then the sacrum of Diplodocus 

 must be considered as composed of five vertebrce. These are usually firmly coossi- 

 fied by their centra, though the centra of the posterior and anterior of these five 

 vertebrae may occasionally remain free or only slightly coossified. The remaining 

 three median vertebrge are always coossified by their centra and usually have their 

 neural spines coalesced into one powerful spine, subequally expanded transversely 

 and antero-posteriorly. These three median vertebrae constitute the sacrum of Diplo- 

 docus as understood and interpreted by the late Professor Marsh, wlio has described 

 the sacrum as consisting of three vertebrae. Osborn, on the other hand, has in- 

 cluded among the sacrals the posterior of the five modified vertebrae, while exclud- 

 ing the anterior chiefly because of the absence in it of a true sacral rib springing 

 from the body of the centrum which he finds present on all the four succeeding 

 sacral vertebrae. He therefore considers the sacrum of Diplodocus as consisting of 

 four vertebrae, the tliree anterior of which he considers as having constituted the 

 primitive Dinosaur sacrum, while the fourth has been added posteriorly by the 

 modification of tlie anterior caudal. The two splendid sacra belonging with skele- 

 tons 84 and 94 in our collections are unusually complete and throw much light 

 upon the structui'e and development of this element in Dinosaurs. In each instance 

 the vertebrae are all firmly coossified with and give support to the ilia. In 84 the 

 right ilium alone is preserved, and this is united to all five of the vertebrae which 

 function as sacrals either by the means of true sacral ribs or the expanded diapo- 

 physial laminae or by both these elements. All are coossified by their centra, and 

 the three median have their neural spines coalesced. Thus in this skeleton it will 

 be seen that the conditions found to obtain in the sacral region are very similar to 

 those described by Osborn except that there is a rather greater modification of the 

 last dorsal in the direction of that which obtains in the true sacrals than was noticed 

 by the latter in his description. (Compare the description and figures given above 

 with those of Osborn.) In skeleton 94, however, there are noticeable certain other 

 more marked differences, which are Avorthy of especial notice as bearing directly 

 upon the nature of the primitive Dinosaurian sacrum. In this skeleton the sacrum 

 is present, with both ilia in position. The centra of the true sacrals are all coossified 

 as in other sacrum. The neural spines of sacrals one and two coalesce and are co- 



