49 



in the form of roundish cells with a round nucleus, and a 

 more or less distinctly granulated protoplasm without any 

 limiting membrane. In the fibrillar bundles, on the other 

 hand, are situated stellate and caudate cells ^vhich precisely 

 resemble these in their nucleus and protoj^lasma. That these 

 two forms are really identical is shoAvn by the phenomena 

 of contractility and migration Avhich they exhibit. A round 

 cell may be seen to stretch out in the direction of a fibrous 

 partition towards which it slowly moves, and at length be- 

 coming fixed, there assume the spindle-shaped or stellate 

 form ; while, on the other hand, a spindle-shaped cell may 

 be seen to move into a mucous space and become round ; so 

 that the form of these cells depends entirely upon the nature 

 of their temporary resting-place. Quite different from these 

 are the fixed cells of the cord which VirchoAv has described 

 and figured (in the ' Cellular Pathology') as branched con- 

 nective tissue cells. These when examined fresh are ramified 

 bodies pale or granulated^ the ^prolongations of Avhich form 

 numerous anastomoses ; sometimes tapering off into mere 

 threads, sometimes remaining uniformly wide. The system 

 thus formed is provided with nuclei which require special 

 mention. They are more frequently contained in the fixed 

 corpuscles than in the prolongations, though the larger of 

 these may jDOssess several. Many at least of these nuclei 

 were evidently situated in the walls of the bodies, looking as 

 if they adhered to the outside. The whole system thus 

 formed is found to traverse the unfibrillated mucous space, 

 no less than the fibrillated portions. It evidently corre- 

 sponds to the connective tissue network of Virchow; only 

 that it is in the author's view not solid, but a hollow system 

 of channels. Where the outline looks dark it is filled up with 

 granular protoplasma ; and this is owing to the presence within 

 the cavity of the above-mentioned variable cells, which are 

 caudate or stellate when lodged in the angular intersections 

 of the hollow network, and become round when the channels 

 in which they are contained are susceptible of easy dilata- 

 tion, as when passing through the mucous alveolar spaces. 

 The identity of this network with the system of channels 

 demonstrated by injection and silver impregnation is shown 

 first by the unequal dilatation of both, and especially by the 

 appearance of partly injected portions, where spaces into 

 which the blue injection had found its way were plainly 

 continuous with the stellate corpuscles and thread-like pro- 

 cesses of the connective tissue network (see fig. 2). 



The last and most difficult question is the nature of the 

 walls of the plasmatic channels. The notion that they have 



VOL. X. NEW SER. D 



