56 MR W. E. AGAR ON THE DEVELOPMENT OF THE 



stage 36, for comparison with the Lepido'siren series. The neural arch a-b is 

 present, fused at its dorsal end to the occipital arch, and so enclosing the nerve a in a 

 foramen. In another larva of this stage fusion had not taken place, so that a comes 

 out through a notch between the two arches, the hinder one (a-b) merely forming a 

 protuberance on the base of the occipital arch. My material did not enable me to 

 trace the fate of this apparently vestigial arch. 



The occipital arch now touches the auditory capsule, enclosing the jugular foramen. 

 The way in which the junction between the arch and the capsule takes place in 

 Lepidosiren has already been described. The stages necessary for showing this process 

 were wanting in Protopterus. 



The foramen pro-oticum has been completed, and the original single opening has 

 been divided into two by chondrification of a strand of connective tissue separating 

 v. 2 from v. 3 , vii. lateralis. In Lepidosiren this division never takes place. 



The front end of the wall of the Gasserian recess and the orbito-temporal process 

 is produced into a long spine arching backwards towards the auditory capsule. 



The ant-orbital process is growing out from the dorsal margin of the trabecula. 

 This will be considered more fully later (p. 57). 



The dorsal margin of the internasal septum is widening out to form the roof of the 

 nasal capsule. 



The branchial arches are making their appearance. They develop successively 

 from in front backwards. At this stage the fifth arch has not yet appeared. The first 

 branchial arch at this stage corresponds to the second of Wiedersheim * and Bridge. 

 It is behind the first of a series of five clefts. From its relations to the aortic arches, as 

 well as from the fact that there is never a cleft in front of it, this first cleft must be 

 the hyobranchial one — homologous to the one which closes during larval life in 

 Lepidosiren, and the first arch at this stage homologous to the first in Lepidosiren, 

 while the subsequently appearing cartilage in front of it is no branchial arch. 

 (K. Furbringer suggests that it is derived from hyoidean rays.) 



In the next stage, figured 36+ (fig. 9, Plate II.) the chondrocranium has become 

 further developed. Complete fusion has taken place between the occipital plates and 

 Balkenplatte, and the basilar plate so formed encloses the notochord dorsally and 

 ventrally, except at its front end, which projects freely into the basicranial fontanelle. 

 The free end of the notochord ultimately disappears. The dorsal ends of the occipital 

 arches are growing inwards towards the middle line to form the supra-occipital 

 cartilage. 



The front end of the basicranial fontanelle is closed by the " cartilaginous basis 

 cranii." Between this and the internasal septum a vacuity is left. 



Shortly behind the level of the eye the ant-orbital process diverges from the 

 trabecula. This process is just recognisable in stage 35. Here we find a short, straight 

 cartilaginous process growing out from the dorsal edge of each trabecula, reaching about 



* Morphol. Studien. Jena, 1880. 





