MORPHOLOGY OF EYE MUSCLE NERVES 33 



Kerr ('02) was led to doubt the hypothesis of the secondary 

 connection of somite and tube on the ground that in the earhest 

 stages, when somatic motor nerves are visible in Lepidosiren 

 the tube and the somite are in immediate contact with each other : 

 ''Lepidosiren thus affords a definite anatomical basis for the 

 view that the nervous bridge between nerve center and end- 

 organ exists from the beginning, and that the growth of the nerve 

 is a drawing out of this bridge as the end-organ is pushed away 

 by the development of the underlying mesenchyma." On the 

 basis of such evidence Kerr concluded that the Hensen hypothe- 

 sis is 'almost demonstrated.' 



This inference seems to be a non-sequitur from the evidence 

 presented, since Kerr has traced the anlagen of the somatic 

 motor nerves only to those stages when the protoplasmic con- 

 nections are already established. Neumayer ('06, page 54) has 

 already called attention to the fact that the stages described 

 by Kerr correspond to advanced stages of histogenesis. If tube 

 and somite in Lepidosiren be normally in contact in early em- 

 bryonic stages, it would seem to be a form little suited to the 

 requirements of an investigation of the primitive connections 

 of nerve and muscle. Squalus embryos, in which a space is 

 normally present, would seem much better objects of research. 

 The normal distance between tube and myotome in Squalus 

 is so small, however, that the theoretical objection that it would 

 be difficult to explain the growth toward the muscle as a tropism 

 is not likely to be suggested. If the direction of growth of the 

 neuraxon processes of the medullary neuroblasts were determined 

 chemotropically by secretions of the myotome cells there would 

 be little chance for the neuraxons to go astray through a mixing 

 or diffusion of specific secretions. In such a case the theoret- 

 ical necessity for predetermined paths emphasized so strongly 

 by Held ('09) does not appear very convincing. The difficulty 

 of explaining how the extended processes of Rohon-Beard cells 

 reach their area of distribution by a chemotropic response is 

 much greater, but Held can hardly expect that his speculations 

 (page 274) as to the stimulation of the neuroblast process through 

 predetermined plasmodesmatous paths will be regarded as pref- 



JOURNAL OF MORPHOLOGY, VOL. 25, NO. 1 



