ON THE STRUCTURE OF CERATELLA FUSCA (GRAY). 13 



These gonophores arise directly from the ccenosarc and are not carried by modified 

 zooids or blastostylcs. 



Tliey may be very mimerous indeed especially on the medium-sized branches 

 wliere their number in one specimen far exceeds that of the hydroid zooids. Two 

 are represented as seen by reflected light in figure 6 and in all specimens examined 

 the gonophores are at the same stage of development. I have been unable to detect 

 any indication of the formation of reproductive elements in the hydrophyton. 



The gonophores much resemble in structure those figured by Weismann* in 

 Pcnnaria cavoliiia or Cladocoryne fioccosa. In essential structure the transverse section 

 of the gonophore of the former is identical with that of a similar section of a 

 Ceratella gonophore as shown in figure 8. The longitudinal section again of the 

 gonophore of Cladocoryne or of Pennaria as figured by Weismann agrees almost 

 precisely with that of Ceratella represented in figure 12. In the latter the 

 manubrium is well developed and surrounding this lie the reproductive elements 

 which in longitudinal section (Pig. 12) form a horseshoe-shaped mass and in transverse 

 section a ring. External to the latter is a thin layer of ectoderm (ect.') which comes 

 in contact with the external layer of ectoderm (ect.) at the part corresponding to the 

 mouth of the medusa bell where the former layer dips inwards. (Fig. 12, M.) This layer 

 must correspond to the sub-umbrella ectoderm of the medusa and the special point 

 mentioned indicates also the position at which, in development, the " glockenkern " 

 of Weismann grew in by proliferation of the ectoderm cells. I have been unable to 

 find any gonophore younger than the stage figured though many have been cut in 

 section, the manubrium being in every case well developed. 



Between the two layers of ectoderm lies the endoderm with four radial canals 

 seen clearly in transverse sections ( Pig. 8) whilst in longitudinal sections (Fig. 12) 

 the indication of a ring-shaped space around the distal end can always be detected. 



Each gonophore may l)e connected with more than one of the ccenosarc tubes 

 and its ectoderm is continuous with that which covers the colony externally. 



(3.) The Hydrophyton. — This may be divided into two parts (1) the external 

 layer common to the whole colony and (2) the branching network of tubes. 



The external layer is formed entirely of ectoderm cells (Figs. 8 and 12) and is 

 only one cell thick though it may come into contact with the ectoderm of the tubes 

 lying immediately beneath it. It is especially well marked in the younger Ijranches 

 and may be worn away to a greater or less extent in older ones though typically it 

 forms a covering for the whole of the colony (Figs. 9 and 13). It is directly 



* Die Eiitsteluing der gexualzellen bei den Hydromedusen. PI. XVII., Figs. 1-5 ami Fig. 7 ; PI. XVIII., Fig. 1. 



