88 THE ANCESTRY OF VERTEBRATES 



changed . his opinion. That part of a dorsal nerve may 

 acquire a certain independence, may be wholly split off, 

 as it were, from the main stem, is shown by the acusticus 

 which is very generally considered as belonging to the 

 facialis. That such a splitting occurred in the trigeminus 

 may be easily accounted for by the fact that it innervates 

 - the prostomium as well as its own segment, and by the 

 strong and special development of the first pair of visceral 

 archs (the jaws) and the first pair of gill-slits (the mouth). 

 Both the trigeminus and the facialis, the first two seg- 

 mental nerves, as I am inclined to consider them, send strong 

 sensory branches (rami ophthalmia trigemini et facialis, 

 ramus buccalis facialis) into the prostomium which, having 

 no ganglion of its own, must be innervated from the 

 anterior segments of the soma as far as its dermal sense-organs 

 . are concerned. This evidently has caused the strong devel- 

 opment of these first two nerves and their ganglia, especi- 

 ally of the trigeminus. In i4/72pA/oxus also we find (cf. fig. 26) 

 that the first segmental nerve, in front of the first somite, 

 has a double nature, 

 r Oculomotorius. — Finally we have the oculomotorius as 

 the ventral root belonging to the praemandibular segment 

 and innervating the eye-muscles that arise from it. If it 

 can be doubted whether the eye-muscles can be derived 

 directly from segmental trunk muscles, then it also 

 seems questionable whether the oculomotorius, trochlearis 

 and abducens can be directly compared to the ventral 

 roots of spinal nerves. In going from the trunk to the head 

 in an embryo we see the ventral roots disappear in the 

 same way as the myotoms when we approach the auditory 

 vesicle. A comparison with Petromyzon and Amphioxus, 

 however, renders it probable that the myotomes and ventral 

 roots of this region have atrophied only secondarily, 

 and that at least the M. rectus externus with the N. abducens 

 may be homologized to a regular myotome with its ventral 

 root. Oculomotorius and trochlearis, however, afford more 

 difficulties. I need only recall here the anomaly of the 

 dorsal origin of the trochlearis, while the oculomotorius 

 springs from the mid-brain which lies in front of the 

 isthmus and accordingly does not belong to the epichordal 

 part of the neural tube, the former stomodaeum. Hatschek 

 (1892, p. 158), in Petromyzon, recognizes only the abducens 

 as a segmental ventral nerve and derives the oculomotorius 



