14 ON THE STRUCTURE OF CERATELLA FUSCA (GRAY). 



continuous with the ectoderm of both the hydroid zooids and the gonophores, and 

 has in many respects a close resemblance to the superficial layer of ectoderm as 

 described and figured by Professor Moseley in Millepora.* In the latter the exact 

 relationship of the superficial ectoderm to the zooids could not be ascertained but 

 in Ceratella where the latter are not retracted into spaces within the skeleton its 

 direct connection with the ectoderm of the zooid can easily be seen in sections. 

 Figure 3 represents a small portion of the layer as seen under a high power. The 

 outlines of cells cannot be definitely distinguished in the specimen in question 

 though a somewhat light space with a fairly distinct outline surrounds the thread 

 cells. The inner ends of the cells are in contact with the ectoderm of the tube 

 beneath and the layer thins out just where it passes over the projecting point of a 

 portion of the skeleton. In younger branches (Fig. 14) the cells of the layer are much 

 more definite in form and outline being each cubical with a distinct nucleus whilst 

 comparatively very few thread-cells are present. 



This superficial layer is only known to exist in the Hydrocorallinse, the 



Hydractiniidae and, now, in the Ceratelladas. In the last mentioned the soft parts of 



only Ceratella fusca are known as yet but the skeleton of Dehitella is so closely 



similar to that of the former that we may with much probability infer that a close 



agreement exists in the nature of their soft parts. The coenosarc tubes form a 



richly branching network of tubes occupying all the spaces in the chitinous 



meshwork which forms the skeleton. In Figures 9 and 10 this is represented 



diagrammatically by the grey colour the former being a longitudinal and the latter 



a transverse section of a branch. Throughout the whole system the endoderm is 



never more than one cell thick whilst the ectoderm is very irregular. Very often the 



endoderm cannot be recognised or else it forms an indistinct layer which stains more 



darkly than the ectoderm and contains no space, a result probably due to the action 



of reagents. At other times (Figs. 3, 11, 12, and 14) a distinct tubular space 



can be distinguished. The number of tubes varies naturally according to the size of 



the branch. Figure 13 represents a longitudinal section through a portion of a small 



branch with a smaller offshoot which formed part of a specimen brought by 



Mr. Whitelegge from Lord Howe Island. The general appearance of a part of the 



same specimen viewed as a solid object is represented in Figure 4. This particular 



specimen has the branches much finer than those of the others and the skeleton and 



ccenosarc tubes somewhat more regular in arrangement, whilst the hydrophores 



are not very strongly developed. Up the small offshoot pass the main skeletal ribs 



united by cross bars which are thin and almost web-like and up the centre runs a 



single coenosarc tube (B) which is connected with at least three of those in the larger 



branch. This tube consists of an internal layer of endoderm (Fig. 13 a, end.) and 



* On the Structure of a Species of Millepora occurring at Tahiti, Society Islands. Phil. Trans. B.S. London, 

 1870, Vol. CLXVII., p. 117. 



