REPORT ON CEPHALODISCUS DODECALOPHUS. 45 



2. Existence of an unpaired body-cavity in the proboscis, and of paired cavities in 



the collar and in the trunk. 



3. Proboscis-pores (paired in Bcdcmoglossus hiqTfferi), opening into the body-cavity 



of the prss-oral lobe. 



4. Collar-pores in similar relation to the collar-cavity, their external apertures being 



overhung by an operculum developed from the collar. 



5. Gill-slits (one pair for a considerable period in the embryonic history of Balano- 



glossus) ; their relation (in Balanoglossus, the relation of the first pair) to the 

 operculum and to the external apertures of the collar-pores. 



6. Existence of a notochord as a diverticulum of the alimentary canal, growing 



forwards into the proboscis-stalk. 



7. Dorsal central nervous system, most richly developed in the collar, but 



extending on to the proboscis ; the fact that the nervous tissue lies in the 

 epidermis. 



Before leaving this subject, it will be well to refer briefly to the highly interesting 

 pelagic larva of Balanoglossus, discovered by Weldon in the Bahamas.^ A noteworthy 

 feature of this larva is the development of a series of tentacles arranged in six grooves 

 passing, equidistant from one another, in a longitudinal direction along the surface of 

 the prse-oral lobe. Although the tentacles are not in the same position as those of 

 Cephalodisciis, it is a suggestive fact that this larva aSbrds another case of the 

 development of tentacles in the anterior part of the body in Balanoglossus or its allies, 

 and it is at least possible that their appearance in the Tornaria may be due to a process 

 of reversion or atavism. 



It appears to me that whatever may be thought of any single similarity between the 

 two genera given in the above list, the cumulative evidence of the whole sequence of 

 resemblances points irresistibly to the conclusion that Cephalodiscus and Balanoglossus 

 are near allies, and I would propose to remove Cephalodiscus from its previous position 

 amongst the Polyzoa, and to place it definitely as a second genus in Bateson's group of 

 the Hemichordata. The character of the Vertebrate features of Cephalodiscus (notochord, 

 gill-slits, nervous system) appears to justify an approximation of this genus to Balano- 

 glossus in particular rather than to any other group of the Chordata. 



The most important diff'erence between Cephalodiscus and Balanoglossus appears to 

 me to consist in the relations of the dorsal and ventral surfaces in the two genera. The 

 difference is, however, a non-essential one. Whilst in Balanoglossus the elongation of 

 the embryo takes place in the line of its long axis, the ventral elongation of a similar 

 embryo in a line at right angles to its primitive long axis would give rise to the condi- 

 tion found in Cephalodiscus. We may suppose that the stalk has originated in this 



> Vitk Proc. Eoy. Soc, vol. xlii. p. 146, 1887. 



