ON THE STRUCTURE OF CEHATELLA FU8CA (GRAY). 15 



an irregnkur external layer of ectoderm (ect.) the former being continuous with that of 

 the xooids. The whole is covered by a uuilaminar ectoderm (E). A few thread cells 

 are present. Up the layer branch the tubes run more regularly than usual and three 

 or four may be traced for a considerable distance running parallel to its length but 

 giving off lateral branches. In sections both transverse and longitudinal of Ceratella 

 these connecting bars or webs crossed by ccenosarc tubes form a very characteristic 

 feature (Figs. 11 and 13. C). 



As stated above there is a strongly-marked difference between the endoderm and 

 ectoderm. The former (Fig. 3) is regular and takes stain somewhat more readily 

 than the latter, which is often very irregular and several cells thick, though most 

 often the outlines of cells cannot be recognised, and a structure resembling a 

 syncytium is formed. In the latter are found (1) nuclei, (2) thread cells, (3) bodies 

 surrounded by a clear space and staining evenly and deeply (Fig. 3A). The 

 thread cells are apparently confined to the ectoderm, though of this I cannot feel 

 absolutely sure, and are found in great abundance in the inner parts of the branch 

 whence they must migrate to the surface if they are to be of service to the colony. 

 It is a somewhat curious fact that they are as a rule present in far greater numbers 

 iu the ectoderm of the cosnosarc tubes than in the most external layer. Figure 3 

 represents a small portion of the latter on a part of a colony where the 

 gonophores were numerous and here thread cells were present in greater numbers 

 than elsewhere. Of the nature of the third-named structures it is difficult 

 to be certain but it is probable that they are ectoderm cells in which thread 

 cells are being formed. There is at all events a curious agreement in appearance 

 between them and the structures which Professor Moseley has described as developing 

 thread cells in Millepora.* He says " the thread cell appears to be developed out of 

 the nucleus of the ectodermal cell, the ectodermal cell becoming much enlarged and 

 forming a wide chamber, in which the process of development takes place. The 

 ovoid nucleus becomes enlarged together with the cell, but not at all in the same 

 proportion the cell always appearing as a wide cavity around it. The nucleus as it 

 enlarges has a rounded nucleolus developed at one end of it." The nucleolus has 

 large granules developed within it, whilst the nucleus becomes finely granular. In 

 the next stage one large coil of the thread appears in the nucleus." 



The earlier stage seen in Millepora when the nucleus with nucleolus at one 

 end of it lies in the cell which forms a clear cavity around it, corresponds exactly to 

 that represented in figure 3A.c. in the case of Ceratella. Though a complete series of 

 stages could not be obtained still those drawn in figure 3A. will serve to show that in 

 all probability Ceratella resembles Millepora closely in the formation of thread cells. 

 In figure 3A.rt. the cell is small and the nucleus but little larger than that of an 



* Loc. cit., p. 129. 



