CRANIAL GANGLIA IN AMPHIOXUS. 119 



are doubtless fibers whose cells are situated in the root or trunk 

 of the nerve. As shown in Fig. 4, these cells are sometimes 

 present in the dorsal and ventral rami, and it seems probable 

 that they will always be found there. I have not been so fortu- 

 nate as to have any cells impregnated far out beneath the epider- 

 mis in the position indicated by Hatschek, although such cells 

 are readily seen in haematoxylin sections. (4) Irregularly pyra- 

 midal cells situated near the canal at a slightly more dorsal level 

 than the pigment cells. These cells are usually provided with a 

 single coarse process which runs to the surface of the cord where 

 it ends in a few thick branches or in broad plate-like expansions 

 against the limiting membrane. From some point in its course 

 this thick process (dendrite) gives off a fine fiber which enters 

 the dorsal root. Eight of these cells are shown in Fig. I. 



There can scarcely be any question of the homology of the 

 cells of the first three types described with the spinal ganglion 

 cells of vertebrates. The description confirms the account given 

 by Retzius of cells within the cord sending fibers into the dorsal 

 roots, but it shows that by far the larger number of such cells 

 are situated where Retzius distinctly dented the existence of any 

 nerve cells. They are the cells whose nuclei attracted the atten- 

 tion of Rohde and the earlier authors. The facts given by Ret- 

 zius together with the discovery (7, 8) that the giant cells in the 

 cord of teleosts are comparable with spinal ganglion cells have 

 been considered as evidence that the spinal ganglia in vertebrates 

 have been derived from the spinal cord. Now that the disposi- 

 tion of the ganglion cells in Auiphioxns is more fully known it 

 shows that this animal is not so different from vertebrates in this 

 regard as was supposed. In Ampliioxus the spinal and cranial 

 ganglia form for each nerve an almost continuous mass extend- 

 ing from the central canal of the cord out into the root of the 

 nerve to and beyond the division into dorsal and ventral rami. 

 Thus it may be said that part of the ganglion cells in Auiphioxus 

 occupy a place within the cord which has been regarded as the 

 hypothetically primitive position for vertebrates, while most of 

 them occupy a place in the nerve roots which approaches the 

 typical position for higher vertebrates. Auipliio.vns is, therefore, 

 not quite primitive in this matter but rather approaches typical 

 vertebrates. 



