46 THE VOYAGE OF H.M.S. CHALLENGER. 



manner, and that the alimentary canal has acquired its adult relations by its partial 

 passage into the ventral protuberance of the body. 



Whilst it appears to me easy to compare Cephalodiscus and Balanoglossus, it does 

 not seem to me impossible that the former may have affinities in other directions as 

 weU. For, imagine that the ventral elongation of the body would be conveniently post- 

 poned until after the end of a free larval life ; the stalk might then be invagiuated into 

 the body of the larva in preparation for its evagination when metamorphosis should take 

 place, as actually occurs in Actinotrocha. After the metamorphosis of the latter, the 

 alimentary canal has the same dorsal flexure as in Cephalodiscus, and this explanation of 

 the metamorphosis of Phoronis is in accordance with the suggestions of previous observers. 



The following considerations may perhaps indicate some affinity between Cephalodiscus 

 and Phoronis ^ : — 



1. The archenteron of Phoronis is developed by a well-marked invagination, whilst 

 part of the mesoblast (vide Caldwell) is formed by a process of (modified) archenteric 

 pouching (as in Balanoglossus). 



2. The prse-oral lobe is large in Actinotrocha, and is provided with a body-cavity 

 which is completely shut ofi", by means of a septum, from the body-cavity of the trunk. 

 The post-oral region is prolonged iuto tentacles, which, although differing in a striking 

 manner from the tentacles of Cephalodiscus, may still have some connection with these 

 structures, or with the operculum of the saaie genus. 



3. The " foot " of Phoronis has precisely the same relations as the stalk of Cephalo- 

 discus. 



4. The nervous system of Phoronis occurs outside the basement-membrane. The 

 ganglion of the prse-oral lobe of Actinotrocha is comparable with a portion of the nervous 

 system of the Hemichordata, whilst the post-oral uerve-ring of Actinotrocha (following 

 the line of the bases of the tentacles) may not impossibly be the homologue of the nerve- 

 ring which passes round the posterior border of the collar in Balanoglossus. If this were 

 the case, the lophophores of Actinotrocha and of Phoronis might be regarded as develop- 

 ments of the collar-region. 



5. Phoronis possesses a complete ventral mesentery, the dbrsal mesentery, however 

 (persistent in Cephalodiscus), having disappeared in the adult animal. The ovaries and 

 oviducts of Cephalodiscus are supported by lateral mesenteries which are apparently 

 arranged in the same manner as the lateral mesenteries in Caldwell's diagram "B."^ 

 The oviducts of Cephalodiscus do not, however, open into the body-cavity, and it is 

 possible that the collar-pores, rather than the oviducts, may be the homologues of the 

 nephridia of Phoronis. 



* Cf. W. H. Caldwell, Prelim. Note on the Structure, Development, and Affinities of Phoronis, Proc. Roy. Soc, 

 No. 222, 1882 ; and Blastopore, Mesoderm, and Metameric Segmentation, Quart. Jotirn Micr. Sci., vol. xxv., 1885. 

 « Proc. Roy. Sac, 1882. 



