1894.] 



HTOID ARCH OP CERATODTTS. 



637 



In each of these cases the ventral process mentioned above is 

 segmented off as a separate cartilage. This may possibly be a 

 symplectic from its position, as suggested by Huxley (5), but, in 



Kg. 3. 



hrrt. 



Ceratodus forsteri. — The hyomandibular and adjacent parts. 

 Letters as before. 



view of the great variability exhibited by the hyomandibular, it is 

 doubtful whether any importance can be attached to its indepen- 

 dence. In the last two cases (fig. 3, e and /) the aperture for the 



