414 E. V. COWDRY 



it is during this stage that they become oriented end to end and 

 fuse to form neurofibrils, and finally, that the fibrils change for 

 a second time their chemical composition so that they acquire a 

 marked affinity for collagen staining dyes. 



The work of Firket ('11) on the formation of epidermal fibrils 

 in the cells of the beak and feathers of chick embryos, and that 

 of Arnold ('12) on the origin of fibrils in malignant tumors is 

 likewise open to criticism owing to their failure to demonstrate 

 transitions between mitochondria and the fibrils, and on ac- 

 count of their tacit supposition that structures which stain by 

 the same method of technique are of similar chemical composi- 

 tion (Bensley '10). 



3. The origin of neurofibrils 



In attacking the problem of the origin of neurofibrils it is 

 necessary to bear in mind all the possible factors, which, by any 

 stretch of the imagination, may be involved. These should be 

 carefully studied and eluninated one b}^ one from consideration; 

 but in so doing, it is absolutely essential to remember that the 

 cell maintains its existence by virtue of the interaction of these 

 same factors, and that sluv attempt to dissociate and analyse 

 them is artificial. The nucleus, chromidial substance, canalicular 

 apparatus and the ground substance ought therefore to be con- 

 sidered in addition to mitochondria. 



I have already criticised the theory of the mitochondrial origin 

 of neurofibrils and this discussion has been carried sufficiently 

 far to show that the mitochondria should not be selected as being 

 the sole agents in the formation of neurofibrils. It is a \'ery 

 different thing to conclude that neurofibrils are differentiated by 

 an actual transformation of mitochondria than it is to enter- 

 tain the possibihty that mitochondria may, in some way, be 

 associated with the histogenesis of neurofibrils; for I am not a 

 great believer in the feasibility of attempts to dissociate cyto- 

 plasmic activities. 



Changes in the nucleus prior to and contemporaneous with the 

 de\'elopment of neurofibrils have been described by Gerini ('08, 



