40 MEMOIRS OF THE CARNEGIE MUSEUM 
genus within the group. Still it is easily conceivable that in any genus or species 
the presacral formula might vary, without increasing the total number of presacrals 
in the individual, according as the exact position in the presacral series at which 
the change from cervicals to dorsals took place, and I am inclined to the opinion 
that as a rule in any given genus of the Sauropoda where marked specialization has 
taken place, there has been a tendency to increase the number of cervicals at the 
expense of the dorsal series, due to the gradual shifting of the pectoral girdle from 
amore advanced to a more posterior position, by which process anterior dorsals 
Have been transformed into posterior cervicals and the cervical region considerably 
elongated at the expense of the dorsal. An extreme instance of this is to be seen in 
Diplodocus carnegii where the disparity in length in the neck and dorsum, probably 
due primarily to the increased number of cervicals and decreased number of dorsals, 
has been still further emphasized by the proportionate length of the individual 
vertebree in the two series. 
If, as does not seem improbable, the total number of presacrals was ever de- 
creased in any genus it would appear to have been more readily accomplished by 
the successive elimination of the less specialized, or at least more simple, anterior 
cervicals than by the disappearance of the extremely complicated dorsals. Nor 
does it seem probable or even possible that such a decrease in the number of presac- 
rals could have been brought about by the gradual shifting of the pelvic girdle to a 
more anterior position. Such anthypothesis presupposes the addition to the sacrum 
of successive posterior dorsals and the liberating of posterior sacrals as anterior cau- 
_dals, an hypothesis which to the present writer appears unworthy of serious con- 
sideration. As to the total number of caudals in the present genus we have nothing 
upon which to base anything like a reliable estimate. From the character of the pos- 
terior five or six of the series of nineteen anterior caudals in the type of H. priscus 
we may judge that while the tail was relatively short the number of caudals was 
considerable and probably not less than forty. ‘The increased number of caudals is 
indicated in the first place by the character of the chevron found attached to the 
thirteenth caudal and which, as already remarked, resembles in general form the 
chevron of a caudal occupying a more anterior position in Diplodocus. Moreover, the 
very gradual change which is seen to be taking place in the posterior caudals of the 
series preserved in the type of H. priscws indicates a very considerable number of 
posterior caudals as having intervened between the last of the series and the end of 
the tail. The extreme shortness of the centra in the caudals of Haplocanthosawrus 
may be considered as sufficient proof that the tail was proportionately rather short 
as compared with that of Diplodocus. 
