260 BULLETIN 100, UNITED STATES NATIONAL MUSEUM. 



The probable derivation of the atlanticum group from the group 

 of smooth species, verticiUatum, hybridum, ellipticum, and opercu- 

 latum, is indicated by the fact that the young colonies of atlanticum 

 and its subspecies are for a time smooth ; that is, their test processes, 

 at least in some colonies, develop rather late. The zooids in young 

 atlanticum colonies are short and more nearly round, as they are in 

 the smooth species, excepting of course the considerably aberrant 

 operculatum. 



The subspecies triangulum resembles some of the Atlantic speci- 

 mens of atlanticum, notably certain colonies of the group interme- 

 dium, more nearly than it does the Pacific forms. Like the former 

 it has a considerably roughened test surface and its zooids are short 

 and angular. Because of these indications of nearer relationship to 

 the Atlantic forms, the line of derivation for the subspecies trian- 

 gulum, in the chart, is carried through the Atlantic Ocean into the 

 Indian and Pacific Oceans. 



ORIGIN AND RELATIONSHIPS OF THE FAMILY PYROSOMIDAE. 



The origin and relationships of Pyrosoma should receive a word 

 of discussion in any taxonomic paper upon the group. Pyrosoma, 

 in its adult anatomy, shows no clear indication of close relationship 

 to the Doliolidae or Salpidae, and classing these pelagic forms all 

 in one major group, as is sometimes done, is hardly justified. On 

 the other hand Coelocormus (Herdmann, 1886) and Cyathocormus 

 (Oka, 1913), in their structure, suggest a probable origin of the free 

 swimming Pyrosomidae from the Compound Ascidians. Indeed, if 

 a detached cylinder of Cyathocormus had happened to be discovered 

 before the whole attached colony was known, it would have been 

 classed with the Pyrosomidae, rather than with other Compound 

 Ascidians, on account both of the form of the colony and of the 

 structure of the zooid. The anatomical evidence in favor of this 

 relationship seems very convincing, but thorough studies of the 

 processes of budding and colony formation in Coelocormus, and 

 especially in Cyathocormus, are greatly needed before we can be 

 certain that Cyathocormus and Pyrosoma have reached their peculiar 

 form in similar ways and arc truly comparable in their structure. 



< 'yathocormus is regarded by Oka as most nearly related to the 

 Distomidae among the Compound Ascidians, and this conclusion 

 appears borne out by the comparative anatomy. 



The Pyrosomidae should, therefore, pending study of budding and 

 colony formation in Cyathocormus, be classed as a divergent group 

 of the Compound Ascidians. 1 



1 For fuller discussion of the relationships among the Tuniroia see Metcalf and Bell (1918). 



