sargext: the optic reflex apparatus of vertebrates. 167 



a direct anastomosis of adjoining tectal reflex cells, a dendrite passing 

 directly over from one into the body of another cell. The difficulty of 

 establishing such anastomoses even where they abundantly exist, leads 

 me to believe that between these cells they may occur frequently. 



Tlie tectal i-eflex cells lie among the coarse ependymal fibres (Figs. 18, 

 20, fhr. e^end.) which arise from cell-bodies iu direct contact with the 

 ventricle and radiate from the median roof of the mesencephalon. Each 

 cell is surrounded by a loose pericellular capsule of finer neuroglia fibres 

 (Fig. 19, cps.pi'd.). Other fine fibres, probably nervous, penetrate this 

 capsule and branching minutely end in anastomosing arborizations in con- 

 tact with the cell-body (Figure 19, a). According to Houser, these fibres 

 are from the stratum medullare profundum. This seems entirely prob- 

 able, though I have been unable to trace them to their origin. 



Turner and Hunter ('99, p. 12.3) have described such a network cover- 

 ing the cell-body of a nerve-cell, and they regard it as the terminal appa- 

 ratus of the axis-cylinder of another nerve-cell, and Meyer ('97, p. 175) 

 has described the basket-like ending of an axis-cylinder over the body 

 of a nerve-cell. Held ('97) has shown that iu the case of many of the 

 cells of the pons, cerebellum, and cord of the rabbit, it is possible to 

 trace a growing intimacy of union between terminals of axons and cell- 

 bodies. About the time of the birth of the animal there is simple pi'ox- 

 imity of a group of fibrils, the terminals come into contact with and 

 gradually fuse with the cell-body by a process that Held calls 'concres- 

 cence.' For some time the junction remains marked by a layer of more 

 highly refractive substance, but in some cases this too disappears, and 

 the fibrils, which may penetrate far into the cell-body, can be distin- 

 guished from it only by their slightly diff'erent texture and staining re- 

 actions. My preparations lead 2ne to believe that I have here found 

 an intermediate stage in the concrescence described by Held, as some of 

 the fibrils of the arborizations penetrate within the surface of the cell- 

 body. 



In size the tectal reflex cells of Raja greatly exceed any other cells in 

 the brain. In young specimens of E. erinacea 12 centimetres long they 

 have generall}' a diameter of from 25 to 10 micra. The elongated spin- 

 dle cells are much greater in length. The nucleus is large and clear, 10 

 to 12 micra in diameter, or nearly one-third the diameter of the cell. It 

 contains one or more spherical nucleoli, 2 to 3 micra in diameter, which 

 are usually eccentric in position, lying nearest that side of the nucleus 

 which is nearest to the boundary of the cell. 



In the adult the cells are much larger, the maximum diameter being 



