184 ESSENTIALS OF ZOOLOGY 



development. The first of the series is the premandibular 

 cavity and is independent, but the cavities on each side are 

 connected by a canal, and this canal comes into close union 

 with the hypophysis and the anterior end of the notochord. 

 From their position in front of the alimentary canal and their 

 attachment to the hypophysis, it is supposed that they are 

 homologous with the head cavities of Amphioxus. The rest 

 are serially repeated in continuation with the other divisions 

 of the coelom, which extend between the gill slits and expand 

 below into the pericardium or the forward part of the splanch- 

 nocoel. They are independent above, therefore, but are con- 

 tinuous below. They are called in succession the mandibular, 

 hyoid, and first, and so on, branchial cavities. The premandi- 

 bular yields the oculomotor muscles ; the mandibular, the 

 superior oblique muscle ; and the hyoid, the external rectus. 

 The remaining cavities are associated with the development 

 of the visceral muscles of the arches. For some reason a 

 desire has been exhibited to look upon the first three cavities 

 as head somites ; but without regarding them as somites, it 

 could be argued that the mesoderm which gives rise to the 

 muscle of the eye is derived from the anterior myotome 

 mesoderm of the head. But there is no difficulty in regarding 

 such muscles as being splanchnocoel in origin, for after all one 

 of them at least is related to a nerve (IV) which cannot be 

 regarded as a ventral root, but more probably as the motor 

 survivor of a mixed dorsal root. 



With these exceptions the cranial nerves are (1) purely 

 sensory, I, II, and VIII; (2) mixed, V, VII, IX, and X, 

 and like the dorsal nerves of Amphioxus they are situated 

 between the somites and pass outside the somites. These 

 visceral nerves of the head are derived from the neural crest 

 and establish a second union with the skin just above the 

 visceral cleft, forming what is called in each case an epibranchial 

 ganglion. The neural crest is continued behind the Xth and 

 survives in higher Craniates as the accessory or Xlth nerve, 

 which is motor. This crest also gives off internally small 

 transient dorsal roots homologous with the spinal dorsal 

 roots and corresponding to the hypoglossal roots below them. 

 We have now good reasons for knowing that the spinal dorsal 



