SYSTEMATIC ACCOUNT 



227 



Bathymetrical distribution 



I take this opportunity of offering my thanks to Professor E. W. MacBride, F.R.S., 

 for getting the materials on his personal responsibility, and for the facilities he afforded 

 me for carrying on this investigation in the Huxley Laboratory under his guidance and 

 supervision ; to Dr Stanley Kemp for having entrusted me with the drawing up of the 

 report on Cephalodiscus; to Mr H. R. Hewer for suggesting names for the new species 

 and other technical advice; and to Sir Sidney Harmer, K.B.E., F.R.S., for the trouble 

 he took in revising the original manuscripts and for suggesting corrections and improve- 

 ments which make this paper acceptable for publication. 



SYSTEMATIC ACCOUNT 



DIAGNOSIS OF THE SPECIES OF CEPHALODISCUS 



The most important specific structure is the coenoecium. The classification into 

 sub-genera is based entirely on the nature of colony formation. Harmer (1905, p. 4) 

 considers the length of the body and stalk of the zooids as satisfactory specific characters, 

 but measurement of the zooids of some species does not support this. In C. hodgsoni 

 the length of the zooids varies from 2 to 32 mm. and that of the stalk from 4 to 10 mm. 

 In C. densus, though normal zooids measure 4 to 7 mm., I have obtained giant zooids 

 II to 14 mm. long, with a stalk which varies in length from 25 to 40 mm. The number 

 of arms can be used as a specific character, as this is constant for each species except 

 C. hodgsoni in which there may be 10, 11 or 12. The number of buds produced can to a 

 certain extent aid in describing the different species. In some species like C. fumosus 

 buds are produced profusely, while in others like C. hodgsoni the number of buds is 

 only three to five. I doubt whether the production of buds is confined to a definite 



