REPORT ON CEPHALODISCUS DODECALOPHUS. 43 



the posterior limit of the central nervous system, it is intelligible (from the relations of 

 the base of the operculum just explained) that the operculum is seen, in the section, to 

 spring from a region of the body quite near to the dorsal suri'ace. It is, however, clear 

 that the operculum m merely a free fold of the posterior border of the collar-region, 

 containing a portion of the collar-cavity, and that it is therefore exactly comparable to the 

 operculum described by Bateson in Balanoglossus. 



On each side of the body is found a well-marked " collar-pore " (fig. 3, c.p.), consisting 

 of a very short canal whose walls are formed of narrow, closely arranged epithelial cells, 

 and opening on the one hand into the cavity of the collar, and on the other to the 

 exterior, the external openings of the pore being overhung by the base of the opercular 

 fold. Both in the structure and in the position of these canals, Cefhalodiscm resembles 

 Balanoglossus to an extent which is almost inconceivable, except on the hypothesis of 

 some genetic connection between the two genera. 



A further Balanoglossus-iea,tu.Te possessed by Cephalodiscus is the existence of a pair 

 of well-marked giU-slits (fig. 3, g.s.) opening to the exterior immediately behind the 

 collar-pores, and so ftir as I have been able to make out from an examination of the buds, 

 apparently developed as outgrowths of the " pharyngeal " region of the alimentary tract. 

 The relation of these slits to the collar-pores is precisely the same as that of the first pair 

 of gill-slits of Balanoglossus to the collar-pores of the latter. Unlike Balanoglossus, 

 Cephalodiscus possesses no more than a single pair of gill-slits, but it must be remembered 

 that the young Balanoglossus remams for some time in a similar condition [i.e., with 

 but a single pair of gill-slits), and that Bateson has assumed the existence of an ancestor 

 of Balanoglossus in which no metameric repetition of the gill-slits had taken place. 



Fig. 4 illustrates the relation of the structures in the proboscis-stalk, which — as in 

 Balanoglossus — is a constricted region by which the proboscis itself is connected with 

 the rest of the body. The completely separated collar-cavities are clearly visible, as well 

 as the unpaired proboscis-cavity. The notochord is a slender rod, possessing a fine 

 lumen, and is supported by the mesentery, which forms the division between the two 

 halves of the collar-cavity. I am not at present certain as to the existence or non- 

 existence of Bateson's " proboscis-gland " in this region of the body. 



The lophophoral arms are deeply grooved on their ventral surfaces, and these grooves 

 are continuous with shallower furrows {gr.), which pass along the ventro-lateral portions 

 of the collar, on either side of the proboscis-stalk, as far as the region of the mouth. If, 

 as can hardly be doubted, the tentacles are cQiated, it may be assumed that a current of 

 water passes in the living animal down these grooves into the mouth, into which the 

 current is directed by means of the opercular flap developed from the posterior border of 

 the collar. It is probable that the gill-slits are, in this case, of great importance to the 

 animal. The two lateral currents which have just been supposed to enter the mouth 

 would doubtless introduce large quantities of water into the phar)'ux. The water would 



