148 PROTOPLASM 



appear with relative frequency cut longitudinally or obliquely in 

 the sections, of Avhich, however, very little is to be made out in 

 Nansen's figures. 



With regard to the alleged composition of the ganglion cells, 

 i.e. of spongioplasma and nerve tubules, it may further be 

 pointed out that Nansen was quite unable to demonstrate such 

 a composition directly in by far the greatest number of cases. 

 The greater number of the ganglion cells investigated by him 

 consisted of a dense tangle of such nerve tubules, and the 

 reticulation of their protoplasm hence depended on the cross 

 sections of numerous thickly - packed tubules, together with 

 their delicate sheaths of spongioplasma. Since, however, as I 

 shall further explain when describing nerve fibres, the assumption 

 both of nerve tubules and of their spongioplasmic sheath is 

 untenable, there is left for ganglion cells of this kind only the 

 simple and most natural interpretation still remaining, namely, 

 that the reticulation of their substance does not depend upon 

 special relations peculiar to nerve cells, but is the ordinary 

 reticulate meshwork of the protoplasm, which merely exhibits 

 modifications that can be in great part explained from the 

 peculiarities in form possessed by these cells, and their being 

 drawn out into processes. It is far from my intention to deal 

 in detail here with the extended literature upon nerve cells. I 

 merely thought it necessary to go rather more specially into the 

 work mentioned, which in essential points approaches so near to 

 my results. 



Nerve Fibres 



If pieces of the ischiadic nerve of the frog be investigated 

 by teasing, after having been preserved in very various 

 reagents, such as picro-sulphuric acid, picro-sulphuric-osmic 

 acid, Miiller's fluid, or weak alcohol (45 per cent), the 

 following can be made out (with or without staining) as to 

 the minuter structure of the axis-cylinder as a whole. It 

 is always distinctly composed of longitudinal fibrils, which 

 are about 0'6 to 0'7 p apart. Since the thickness of the 

 axis-cylinder varies, the number of the fibrillse is naturally 

 subject to change. I have seen axis-cylinders of such fineness 

 that their width only contained four fibrillae. The apparent 

 fibrillse are always finely punctate, i.e. they exhibit at 

 fairly regular intervals, which are slightly farther apart than 



