A NK\V FAMILY OF HYPROIDEA. 1'27 



ccenosarcal tubes. In this they resemble the polypes in the group Hydrocorallina?. 

 The cavities of the polypes nnd of the meshwork of tubes are hence directly 

 continuous. The tubes are formed of the two primary layers with, if present 

 at all, an extremely thin layer of mesogla'a between them. The endoderm is always 

 unilaininar and distinct ; the ectoderm is composed of smaller cells, the nuclei, but 

 not the outlines, of which are clearly marked. It appears to be frequently more than 

 one cell in thickness, and to be irregular in outline, spirit specimens showing 

 projections (Fig. 10) which come in contact with the chitinous walls. Probably 

 during life there is very little space between the hard and soft parts. Within the 

 ccenosarcal tube are frequently seen large globular structures containing darkly 

 staining portions (Fig. 10. x.) 



The whole of the external layer of coenosarcal tubes is studded over 

 with an enormous number of small defensive polypes of very simple structure 

 (Figs. 3, 5, 15). One corresponds to each of the cylindrical structures which rise up 

 from the outer thin peris arc layer, and consists of a stalk and a head. The former 

 is directly continuous with the ccenosarc tube, and consists of an internal axis, the 

 nature of which is difficult to determine. It is solid, and contains no prolongation of 

 the tubular canal of the ccenosarc, but appears to be continuous with the endoderm, 

 though I cannot as yet make out the cellular structure. It is very much shrunken 

 in spirit specimens. The outer layer is thin and continuous with the ectoderm of 

 the coeuosarc (Fig. 15). In spirit-preserved specimens it has the appearance of a 

 thin layer of granular protoplasm with nuclei, the diameter of which is greater than 

 the thickness of the protoplasmic layer, in which the outline of cells cannot be 

 distinguished. No trace of muscle elements can be distinguished. 



The head consists of a little mass of nematocysts, with the remnants of the cells, 

 which have given rise to them, and the nuclei of which can be clearly seen. Each 

 nematocyst is fusiform in shape, and the whole form a little group lying close to the 

 open end of the cylindrical tube, through which the threads can doubtless be ejected. 

 None of the nematocysts have the threads put out, and their exact form cannot be 

 therefore described. 



In relative size and structure these defensive polypes resemble more than 

 anything else the machopolypes of the Plumulariidas, and have no resemblance to the 

 defensive polypes of Hydractinia or Podocoryne, whilst none have as yet been 

 described in the Ceratelladae. 



In figure 3 is represented a diagrammatic restoration of a branch of the colony 

 only the soft parts being drawn. The upper surface is supposed to have been cut 

 away, and the network of cronosarcal tubes is shown branching and anastomosing. 

 The connection of the alimentary polypes with the tubes is drawn, and the small 



