42 Dr. G. Baur on Meiolania. 
formed by the connexion of the neuroids with the first inter- 
centrum (atlas-ring of birds and mammals). Such a ring 
does not exist in the Pleurodira, so far as I know. If the 
atlas of a Pleurodiran is macerated the intercentrum, if not 
coossified with the centrum, becomes free; it does not remain 
in connexion with the neuroids; these generally are united 
with the centrum or may become free, as in the Podocnemi- 
didee (Podocnemis Dumeriliana and expansa) and some Chely- 
didee (E/seya according to Vaillant). But if we macerate the 
atlas of a Cryptodiran we never find, so far as I know, that 
the neuroids of the atlas remain in connexion with the cen- 
trum; the neuroids of the atlas are connected with the 
intercentrum, or all parts may become separated, .as in the 
Chelydride, Chelonide. ‘This I wanted to state, and there- 
fore I used the words “it is the centrum which supports the 
neuroids of the atlas’? in the Pleurodira. I excluded the 
Podocnemidide, stating that the intercentrum also takes some 
part. 
If we compare Mr. Boulenger’s figures with these remarks 
there seems to be considerable contradiction. 
In the specimen of Erymnochelys madagascariensis before 
me the atlas is missing ; but for comparison I give an exact 
figure of the first two cervicals of Podocnemis Dumeriliana. 
Podocnemis and Erymnochelys are allied forms, and Hrymno- 
chelys is even not different from Podocnemis according to Mr. 
Boulenger. We may expect therefore that the first two 
cervicals show in both similar conditions. As anybody will 
see, my figures of Lod. Dumeriliana are entirely ditferent 
from those of Erymnochelys given by Mr. Boulenger. The 
neuroids are supported by the atlas centre, the intercentrum 
is connected a little with the neuroids. The intercentrum is 
very small compared with the centrum and projects below it. 
In Chelymys victorie we find a similar condition to that in 
Pod. Dumeriliana ; but the intercentrum simply touches the 
neuroids. ‘The neuroids are more connected with the centrum 
than in Pod. Dumeriliana. This condition is very much more 
expressed in the followmg genera of the Chelydidee— Chelys, 
Hydromedusa, Chelodina, Hydraspis—and in the Sternothe- 
ride ; in all these the neuroids have entirely gone on to the 
centrum, with which they become ossified. 
In the Cryptodira the conditions are entirely different. 
The most generalized arrangement we find in the Cheloniide 
and Chelydridg, in which the atlas-elements are only loosely 
connected; but even in these forms the centrum does not 
support the neuroids. The connexion between these elements 
is on a vertical plane, there is no support; the neuroids are 
