NERVOUS SYSTEM OP ANTEDON ROSACEUS. 545 



the mutual relations of these several forms ; in the second place 

 it is of special interest in connection with the subject of the 

 present paper, as showing that the Asterids are in at any 

 rate one extremely important respect far more primitive than 

 the Crinoids. I propose to say a few words on both these 

 points. 



Starting with the Asterid nervous system it is easy to derive 

 from it, theoretically, the nervous systems of other groups. 

 The sinking down of the radial nerves in Ophiurids and Echi- 

 nids may possibly be connected with the development of the 

 protective calcareous plates on the ambulacral surface, while 

 the similar position they hold in Holothurids is probably due 

 to the descent of this group from mailed ancestors provided 

 with calcareous ambulacral plates, a line of descent for which 

 there is a considerable amount of evidence forthcoming. That 

 the radial nerves of Ophiurids, Echinids, and Holothurids are 

 really the same things as the radial thickenings of the nerve- 

 sheath in Asterids, in spite of their difference of position, is 

 practically proved by the identical relations of the branches of 

 these nerves or thickenings to the tube feet, which branches in 

 all cases alike form sheaths immediately beneath the epi- 

 dermis. The external plexus of Echinids may clearly be 

 viewed as a somewhat modified nerve- sheath ; and the internal 

 plexus of Romanes and Ewart, which is said to be connected 

 directly with the external plexus through the substance of the 

 test, may be explained as due to this nerve-sheath having com- 

 menced to shift inwards, just as the radial nerves have done, 

 but at present remaining entangled in the substance of the cal- 

 cified dermis. 



As regards the origin of the Crinoid nervous system, I think 

 that the Asterid again gives us an important clue, though 

 much yet remains to be explained. It is commonly assumed 

 that the subepithelial bands of the Crinoid are homologous 

 with the radial nerve-bands of an Asterid, and I think the 

 homology must be accepted when we consider how absolutely 

 identical the relations of these two structures are to what is 

 perhaps the most characteristic feature in an Echinoderm, i.e. 



