106 MK. E. PHELPS ALLIS, JUN., ON 



vertical position. Because of this, and also because of tlie 

 relatively much greater development of these bars, they have 

 crowded the bars of the more posterior arches posteriorly to 

 such an extent that their dorsal ends lie in the auditory and 

 spinal regions. The sphenolateral cartilage of either side (ali- 

 sphenoid, Sewertzoa) lies, at this stage, in the horizontal plane 

 of the parachordals, lateral to the hollow of the plica encephali 

 ventralis. The nervi facialis, trigeminus, and ophthalmicus 

 profundus all run outward postero-dorsal to this cartilage, the 

 oculomotorius and opticus, and apparently also the trochlearis, 

 running outward autero-ventral to it. In slightly older stages 

 an ethmoidal cartilage is developed, which begins at a point 

 ventral to the lateral edge of the trabecula of its side, and 

 ventral to the eyeball, and runs dorso-anteriorly dorsal to the 

 nasal sac. 



When the cranial flexure later becomes reduced, there is first 

 formed, in Pristiurus, and hence probably in all the Selachii, a 

 pronounced pontial flexure, but this flexure also becomes later 

 reduced, and almost entirely disappears in the adult. During 

 these changes the trabeculte curve forward, or forward and 

 upward, following the curved ventral surface of the brain, and 

 this change in their direction and position affects the mandibular 

 branchial bars, which, retaining their perpendicular relations to 

 the trabeculte, swing downward and forward, and so acquire a 

 position vertically beneath the latter cartilages. These mandi- 

 bular bars had apparently already become attached to the hyal 

 bars by connective or ligamentous tissues, and the latter bars 

 and the bars of the branchial arches had become similarly 

 attached to each other both dorsal and ventral to the branchial 

 clefts. When the mandibular bars swung downwaTd and 

 forward the}'' accordingly pulled on the more posterior bars, but 

 as the dorsal ends of the hyal and branchial bars, and the ventral 

 ends of the latter bars, had become attached to tissues in the 

 region where they lay, they were relatively fixed in position, and 

 the middle elements, alone, of each bar could be moved forward, 

 the branchial clefts of course shifting forward with them. This 

 gave rise to the sigma form of branchial bar actually found in 

 the adults of these fishes. In the Teleostomi, where there is but 

 little cranial flexure when the neurocranium begins to chondrify, 

 the trabeculte are laid down in the line prolonged of the para- 

 chordals. The mandibular branchial bars were accordingly there 

 laid down approximately in a vertical position, and the bars of 

 the more posterior arches, not being pushed posteriorly by them, 

 retained their jjrimitive attachment in the cranial region. The 

 sigma form of arch was accordingly not impressed upon the bars 

 of the branchial arches of these fishes. 



The trabeculae, in the older embryos described by Sewertzoff", 

 have fused with each other in the median line anterior to the 

 pituitary body, and been prolonged, as a wide median plate, to 

 the nasa,l I'egion. There the plate contracts abruptly and is 



