Dr. Asajiro Oka: 29 



from both ColcUa and Diplosoma,^ but agrees with Pyrosoma. As 

 stated before, the only other compound ascidian in which the 

 peribranchial wall is so imperfectly developed as to expose the 

 greater part of the branchial sac, is found among the family 

 Diplosomidae. 



In short, it seems highly probable that the new family is 

 more closely allied to some members (genus Colella) of the Dis- 

 tomidae than to any other group of the Ascidae compositae. 

 Though it is sufficiently characterised by the hollow cylindrical 

 form of the colony with a large centrally placed common cloaca, this 

 form might be regarded as a modification of the colony form 

 actually met with in certain genera. If, in future, transition forms 

 should happen to be discovered, it might of course become neces- 

 sary to unite the Cyathocormidae with one of the closely related 

 families. At present, however, it seems best to consider our form 

 as the type of a distinct family and place it, in a phylogenetic 

 classification of the Tunicata, somewhere in the neighbourhood of 

 the Distomidae, representing a special branch leading in the 

 direction of the Pyrosomidae. By the intervention of the present 

 family the Ascidiae Salpaeformes would be much more closely 

 connected with the rest of the Ascidiacea than was hitherto the case, 

 rendering it doubtful whether we are justified in separating 

 Pyrosoma from other compound ascidians and placing it along with 

 Salpa and Doliolum in a difïerent order, the Thaliacea. 



Publ, Mar. SOth, 1913. 



