THE CHONDROCRANIUM OF LEPIDOSTEUS. 251 



-a marked ledge (Allis, '03). A perforation of the mesial wall 

 of this anterior prolongation of the groove would give rise to the 

 conditions described by Ridewood ('05) in certain of the Clnpeid?e, 

 this perforation of that waW being called by hiin the temporal 

 foramen. The foramen is, however, probably simply a fenestra, 

 for, in a single specimen of Cbjpea alosa that I have examined, I 

 can find no nerves or vessels traversing it. It opens directly into 

 the dorsal end of the mid-brain recess of the cianial cavity, and 

 it would seem as if it must be closed by membrane, but no such 

 membrane was evident in my specimen. The pre-epiotic fossa of 

 Bidewood's descriptions of the^e fishes is simply a depression 

 in the mesial v.'all of the deeper posterior portion of the entire 

 temporal groove. 



A modification of the temporal groove, as above described, 

 occurs when, as in certain of the Teleostei. the massa angularis 

 is excavated from its lateral surface, ventral to and in the hollow 

 of the semicircular canal, to form the subtempoi-al fossa of 

 Sagemehl's ('91) descriptions. The axis of this fossa lies trans- 

 versely to that of the tempoi-al groove, and its arched roof forms 

 a transverse ridge across the floor of the latter groove, at about 

 the middle of its length ; and when the temporal groove is roofed 

 by dermal bones, and so becomes a fossa, and the subtemporal 

 fossa is extensive, as in certain of the Barbidre, it practically 

 suppresses that part of the temporal fossa that lies dorsal and 

 anterior to it. The tempoi'al fossa then becomes a large but 

 shallow depression on the posterior surface of the cranium, from 

 which, according to Sageniehl ('91, p. 553), a simple cleft may 

 extend forward between external and internal plates of the 

 pterotic (Squamosum, Sagemehl) and represent the anterior 

 portion of the primitive fossa. This cleft is said by Sagemehl 

 to be in large part filled with fatty tissue which contains pigment- 

 cells and a few scattered muscle-fibres, the presence of the 

 muscle-fibres seeming to show that the groove of these fishes was 

 primarily more extensive, and that it has suffered leduction as a 

 result of the development of the subtemporal fossa. In Albula 

 (Hidewood, '04) this type of fossa is also found, but the anterior 

 portion of the fossa has not been completely pinched off and 

 suppressed, and the fossa further differs fi-om that in the Barbidfe 

 in that its anterior portion lies l)eneath the frontal bone, and in 

 that its floor is in part formed by the prootic. 



In Elops the temporal groove is similar to that in Albnla, but 

 the trvnik-muscles have here pushed forward in the direction of 

 the little recess described by me ('03, p. 51) at the antero-lateral 

 corner of the deeper posterior portion of the temporal gi'oove of 

 Scomher. The resulting anterior extension of the temporal fossa 

 of this fish thus lies lateral to the anterior semicircular canal, in 

 the angle between that canal and the anterior portion of the 

 lateral semicircular canal, and ventral, instead of dorsal, to the 

 investing bones on the dorsal surface of the cranium. Its 

 anterior end reaches the alisphenoid region, and is there bounded 



