126 THE SALAMANDER 



SO close together that — according to Miss O'Neill — it is not possible 

 to distinguish between them. In this region also the saccus endo- 

 lymphaticus is continuous right across the dorsum of the brain and 

 communicates with the sac of the opposite side. ^ Thus the spatium 

 interdurale is also continuous, since it is caused by, and develops pari 

 passu with, the development of the saccus. Ventrally the spatium 

 interdurale and the saccus terminate lateral to the pituitary body. 

 The trigeminus nerve passes through the interdural space to the 

 exterior, while the anterior root of the vagus and the acustico-facialis 

 root leave the cranial cavity at the border between the spatium 

 endolymphaticum and the spatium perilymphaticum. 



(ii) Spatium interdurale perilymphaticum (O'Neill). This inter- 

 dural space resembles the former in that it is caused by a saccular out- 

 growth from the auditory organ, but it is much more limited in extent. 

 The saccus perilymphaticus lies within the spatium interdurale and is 

 connected with the ear by the ductus perilymphaticus^ which passes 

 through a special opening in the cranial wall — t]\e foramen rotundum 

 of Hasse, or better, the apertura ductus perilymphatici (O'Neill). 



Hasse thought that the ductus perilymphaticus was connected 

 with the inner envelope of the brain (Gehirnhulle)^ and that the peri- 

 lymph flowed into the cavum epicerebrale, that is, into the space 

 between the brain and the Gehirnhiille. Miss O'Neill, on the other 

 hand, is convinced that this is not the case, but that the perilymph is 

 confined to the saccus perilymphaticus within the interdural space. 

 In this she is supported by Harrison (1902). She thinks, however, 

 that there may be, here and there, some connexion between the peri- 

 lymph and the endolymph where the two sacs come into close contact. 



The spatium interdurale perilymphaticum lies ventral to the spa- 

 tium interdurale endolymphaticum which is laid over it like a saddle, 

 extending beyond it anteriorly and posteriorly. 



(iii) Spatium interdurale hypophyseos (O'Neill). As is well known, 

 the hypophysis cerebri or pituitary body develops as an evagination 

 from the roof of the mouth, and finally separates entirely therefrom, 

 becoming applied to the ventral surface of the brain at the infundi- 

 bulum. In doing this it passes through the periosteal layer of the 

 dura mater and pushes the neural layer in front of it, thus forming 

 and lying within an interdural space — the spatium interdurale hypo- 

 physeos. 



' Dempster (1930) states that while the two sacci endolymphatici do communicate in 

 some Urodeles they do not do so in Salamandra, where they abut very closely and are 

 separated only by an exceedingly thin membrane. He accepts Sterzi's view as to their 

 position, but does not mention Miss O'Neill's paper. 



