640 COMPARATIVE ANATOMY 



derm in front of the notochord. Consequently, Delsman stresses the 

 contrast between a prechordal archencephalon and an epichordal deuter- 

 encephalon. In early stages of vertebrate ontogenesis, the contrast 

 between the two regions is very evident. The line of demarcation between 

 the two is the plica ventralis which, according to Gegenbaur ('72), indi- 

 cates the limit between the segmented "vertebral" and the unsegmented 

 "prevertebral" part of the skull. Thus it follows, thinks Delsman, that 

 the brain vesicle of Amphioxus is to be compared with the deuteren- 

 cephalon of the vertebrate brain. 



The similarity of larval amphioxus and annelid is seen in the relations 

 of the mouth of the annelid and the neuropore of larval amphioxus. 

 Both openings lie just posterior to the preoral lobe and just in front of 

 the first mesodermal segment. 



Delsman regards the anterior endodermal diverticula of amphioxus 

 as the anteriormost pair of gill pouches, and assumes that the mesoderm 

 of the preoral lobe, like that in annelids, is formed by the forward exten- 

 sion of the first myotome. Thus the mesoderm of the preoral lobe is, 

 he concludes, secondary. 



The notochord of amphioxus ends primarily at the level of the neuro- 

 pore, while in vertebrates the forebrain lies anterior to the notochord. 

 Hence it follows, according to Delsman, that the forebrain of vertebrates 

 has been formed from the ectoderm of the preoral lobe of amphioxus. 



Delsman appears to think that comparison of annelid with vertebrate, 

 and more particularly the homology of the annelid mouth with the 

 neuropore of amphioxus, compels him to deny metamerism in the preoral 

 region and to assume that the mandibular segment with the prostomium 

 form not two segments but a single segment. Thus the anterior endo- 

 dermic diverticula of amphioxus become gill-pouches and not mesodermal 

 somites. Consequently, the series of three divisions which form the 

 "brain" of annelids have not the value of segmental ganglia and do not 

 represent as many segments. The trigeminus to him becomes a single 

 nerve, the first somatic, which belongs with the mandibular segment. 

 The definitive vertebrate mouth is assumed to be a fused pair of gill-slits. 



Objections to the Delsman Hypothesis. Since Delsman assumes that 

 amphioxus is intermediate between annelids and vertebrates, he is 

 confronted with the difficulty of explaining the absence of well-marked 

 spinal ganglia in amphioxus, in which they should be more evident than 

 in vertebrates. He attempts to meet this difficulty by assuming that 

 either the ganglia of amphioxus have fused with the neural tube, or have 

 been lost by degeneration, as the cerebral ganglia have been lost. 



Another difficulty arises out of the development of the neural tube. 

 From the standpoint of the Delsman hypothesis, we might expect the 

 neural tube to arise as a tubular invagination of the ectoderm, afterwards 



