THE NERVOUS SYSTEM OF ANTEDON ROSACEUS. 283 



negative result obtained when the stimulation was applied very shortly 

 after the operation is, I think, most certainly to be ascribed to the 

 shock of the operation, which, as already noticed, must always be 

 kept in mind as a disturbing element. 



If the communication is not effected by the subepithelial band, nor 

 by any of the soft parts of the ventral surface of the arm — all of which 

 were scraped away in the operation — it must take place either through 

 the integument of the dorsal and lateral surfaces of the arm, or 

 through the calcareous segments, or through the axial cords, for these 

 are the only parts left uninjured. To determine which of these is the 

 real path of communication the following crucial experiment was made. 



Experiment 13. — A large and vigorous specimen was taken, and a 

 quarter of an inch of one of the arms, about an inch from the base, 

 thoroughly scraped with a scalpel all round so as to remove the soft 

 parts as completely as possible. The pinnules of the affected part and 

 for a quarter of an inch on either side of the wound were cut away to 

 prevent any possibility of contact communication between the parts 

 on either side of the injury. The injured part was then painted all 

 round very freely with strong nitric acid, the operation being repeated 

 until fully half the thickness of the calcareous segments had been 

 dissolved away. The wound was then washed freely with sea water 

 and the animal returned to the tank. It fell at once to the bottom 

 on its side with the injured arm and the other one of the pair 

 stretched straight out horizontally, and the other arms rather strongly 

 extended. After a few minutes it began to move slowly, and in six 

 minutes had completely resumed the normal position. After half an 

 hour's interval the distal end of the injured arm was sharply nipped 

 with forceps, when strong active movements of all the arms at once 

 resulted, the animal moving rapidly away from the source of irritation. 



The above experiment, which was repeated several times, both on 

 entire and eviscerated specimens, proves conclusively that the com- 

 munication between the distal end of the irritated arm and the motor 

 mechanism of the arms is effected by the axial cord ; in other words, 

 that the axial cord plays the part of an afferent or sensory nerve, 

 conveying impulses centripetally. Furthermore, that it is the normal 

 path of communication of such impulses is, I think, evident from the 

 response to stimulation being as ready when it alone remains as in the 

 uninjured animal. It remains, however, to show whether it is the 

 only path of communication. To test this I attempted several times 



