456 A. B. MACALLUM. 



it is usually seen are uncoloured^ or nearly so. The fibrils of the 

 intercellular order are generally so delicate that it is difficult 

 to arrest the removal of the colour of the overlying or under- 

 lying cells with potassic cyanide at a point where the distinct- 

 ness of the fibrils is retained. I had therefore no success in 

 trying to determine definitely the common occurrence of an 

 intercellular network. 



The simple intracellular nerve-twigs always termi- 

 nate in the neighbourhood of the nucleus, either 

 singly or after branching, each terminal point being 

 a delicate bead. The unbranched twig may end on the 

 side of the nucleus facing the point where the twig penetrates 

 the cell, or after curving around the nucleus on the opposite 

 or on one of the lateral sides. When a twig branches two 

 or more of the branchlets may terminate in the positions 

 mentioned. In many cases it is possible to trace a twig 

 for a certain distance after it enters the cell, but not to 

 determine how it ends, this being usually due to the deep 

 colour which the nucleus and the dense cytoplasma immediately 

 about it take from the gold. On the other hand, the cytoplasma 

 may acquire such a deep stain relatively that the determination 

 of the presence of intracellular nerve-twigs is almost impossible. 

 Frequently also these have but a pink or red tint while the 

 fibrils from which they arise are deep violet. 



All the forms of intracellular nerve terminations to be found 

 in the liver of Necturus are not so simple as that just de- 

 scribed. A form which I have several times met with having 

 the intracellular twig branching dendritically is represented in 

 fig. 4. A complicated mode of ending, a very common one 

 in good preparations, appears to belong to the perivascular net- 

 work. I am inclined, from reasons which 1 shall state further 

 on, to regard it as a general one for the hepatic cells of 

 Necturus. Fig. 5 shows how widely it diff'ers from the other 

 modes. Here a branch from the perivascular network pene- 

 trates a cell and becomes continuous with the cell reticulum in 

 such a way as to leave the impression at first that this reticulum 

 belongs to the nerve- twig rather than to the cell itself, this 



