320 



of Herbst's corpusfles there were more nuclei than the normal 

 corpuscle presents. There also seemed to exist a slight increase in 

 the number of the capsule-cells of Grandry's corpuscles. 



While regenerating the nerve-fibers follow the old nei-ve-courses 

 (which have changed into strands of BüN(iNHR), and pass again into 

 the primary corpuscles. It seems, however, that all along also new 

 corpuscles, especially Grandry's corpuscles, are formed, in which 

 process sheath-cells (lemnoblasts) grow larger and become tactile 

 cells, as Heringa has established as to embryological development. As 

 soon as the nerve-fibers have reached the tactile cells of Grandry, 

 they branch out, grow sinuously round them, always embedded in 

 the protoplasm of the capsule-cells and at length force their way 

 between the tactile cells. Directly after this the neurofibrils begin to 

 branch, broadening reticulations appear, which gradually spread 

 between the tactile cells, first as a delicate retiform structure, after- 

 wards as a close-mesh network. In this way the whole interspace 

 between the two tactile cells is occupied again by a net-shaped 

 neurofibrillar nerve-plate. 



Two things strike us here as being remarkable: 



First of all that in the beginning of the process of regeneration 

 the nerve-fibers bend round the tactile cells in various convolutions 

 and ramifications, but that in the following stages (after 2 or 3 

 months) this process is less pronounced, so that gradually the normal 

 condition asserts itself in the sa>ne way as with the motor end- 

 plates ; secondly that neither the nerve-fibers themselves nor their 

 terminal branches and terminal broadening ever run freely, but 

 always remain enclosed in the protoplasm of the conducting cells 

 and the capsule-cells, and that directly when they are within reach 

 of the tactile cells, a peculiar network is formed around them, 

 inside the protoplasm of the tactile cells, which could also be demon- 

 strated, in complete distribution, in the normal corpuscles of Grandry ; 

 lastly that here the process of regeneration of the intraprotoplasmic 

 network shows itself first round the end-branches (end-reticulations 

 and end-knots) of the nerve-fibers and then appears to extend gra- 

 dually ovei- the whole extent of the flat tactile cells. The whole 

 regeneration-process takes two or three months. 



In the case of Herbst's corpuscles the in-growing nerve-fibers also 

 follow the old nerve-tracks. At their point of entrance into the core 

 of the corpuscle we see also here that the nerve-fiber not only 

 proceeds linearly into the protoplasm of the syncytially connected 

 celss of the core, but also that it throws out its branches and passes 

 with many convolutions through the protoplasm, so that the aspect 



