428 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



In the regenerating arm the nerve fibers at first lie directly upon the endo- 

 thelium of the dorsal canal, and it is some time before the lateral ingrowth of con- 

 nective tissue and the development of the calcareous substance of the ossicles forces 

 them away from the dorsal canal and into their normal position. 



Subsequent growth of the dorsal nerve cord proceeds normally. It is ensheathed 

 by wandering cells, and ganglion cells can be made out along its course. At four 

 points the latter occur in considerable numbers, and in these regions the cord 

 thickens and becomes drawn out into branches which lead to the ends of the dorsal 

 ligaments, and which, in combination with the ganglion cells previously described, 

 which at first are somewhat elongated and possess niunerous processes which later 

 unite with the nerve fibers, innervate the dorsal ligament. 



The rebuilding of the ventral portion of the mesodermal nervous system, which 

 runs on either side of the water tube, follows a course parallel to that of the dorsal 

 portion of the same nervous system, but lags behind the regeneration of the other 

 brachial elements. 



The two ventral nerve cords are formed from the endothelium of the water 

 vessel by a proliferation of cells into the connective tissue, which results in a 

 thickening of the walls on either side, and from the elements of these thickened 

 walls ganglion cells are differentiated; in other words, they arise exactly as the 

 dorsal nerve cord. 



Reichensperger found that the destruction of the dorsal nerve in the stump of 

 a mutilated arm usually induces autotomy beyond the injured region; but if this 

 does not occur regeneration is retarded for about 10 days, suggesting tliat it is 

 necessary for the nerve to regenerate before a general rebuilding can take place. 

 From this he assumes that it is the dorsal nerve cord which plays the most im- 

 portant part in regeneration. It is possible, however, that the retardation of 

 regeneration under these conditions results rather from the effect on the wandering 

 cells than from the damage to the nerve itself. 



It is probable that in the larva? the development of the mesodermal nervous 

 system will be found to follow the lines indicated by Eeichensperger in regenerat- 

 ing pinnules and arm tips, for in regeneration the ectodermal nerves follow the 

 same course as in ontogeny. 



Since it has been demonstrated that in the young larva? the ectoderm so com- 

 pletely fuses with the mesoderm that no dividing line can be foimd between them, 

 the idea suggests itself that the so-called mesodermal nervous system in reality 

 owes its origin to cells prunarily ectodermal which have passed downward into the 

 mesodermal tissues. 



During the early stages of regeneration the cells in the connective tissue 

 become more and more noticeably differentiated, and the epithelial cells with their 

 supporting processes become more and more distinct from those of the underlying 

 cutis. Correlatively the evagination of the tentacles from the water tube begins, 

 which is accompanied by a parallel extension of the connective tissue and epi- 

 thelium of the arm on the ventral side. A process similar to that which has taken 

 place in the endothelium of the dorsal canal now occurs in the single-layered epi- 

 thelium of the ambulacral groove. Several layers of cells are formed which 



