OF COENOPSAMMIA FROM LIFU. 369 



and slightly backwards at first, and then straight to the somewhat pointed and slightly 

 projecting free end of the nematocyst. Generally the basal end of the nematocyst 

 lies in a finely granular mass of protoplasm with a nucleus either oval or rod-shaped. 

 Usually a filament, comparable to the granular peduncle described by Lendenfeld in 

 Hydra (19), can be distinguished running down to the nervous layer; the structureless 

 peduncle of Hamann (12) cannot be seen, but numerous processes of the structureless 

 membrane pass everywhere into the ectoderm. 



Although I have very carefully examined the uematocysts in the tentacles of 

 seven polyps, I have failed to find any of a structure differing from the above. The 

 whole nematocj-st is thrown out when the thread is extruded, and in thick sections 

 I have been able to follow every stage of the process. The end of the thread is 

 ejected first and then turn after turn of the spiral follows (Fig. 8). The thread itself 

 under ocular 4 and y'j oil immersion, appears for its whole length to be absolutely 

 homogeneous and structureless. 



I have considered it necessai'y to describe the structure and the appearances 

 on extrusion in some detail, as Mobius (20) appears to have confused neraatocysts 

 similar to these in the tentacles of Caryopkyllia with those found on the mesenterial 

 filaments of the same form. He appears indeed to have considered them to be young 

 stages in the development of the mesenterial uematocysts. Bourne (4) too, following 

 Mobius, gave a figure of a similar nematocyst, apparently not quite ripe, as a developing 

 nematocyst of the mesenterial foi-m. 



Although I have not been able to follow every stage in the process of formation 

 of these nematoc3-sts, yet it seems to be as follows. In the place of the ejected 

 nematocyst appears a homogeneous mass of protoplasm, extending almost from the surface 

 of the ectoderm to the structureless membrane, and having near its base a nucleus. 

 The central part of this acquires a definite membrane and subsequently forms the 

 nematocyst. No nucleus can in any stage be distinguished within this membrane, 

 but the basal nucleus is nearly always well marked. Its protoplasm next becomes 

 very finely granular (Fig. 9), and the fine granules then fuse together and aiTange 

 themselves in definite spiral lines close under the external membrane (Fig. 10). The 

 young nematocyst is now about one and a half times as long and thick as when ripe. 

 Lastly the gi-anules fuse, forming the thread, and the external membrane still further 

 contracts. The thread seems to be tightly coiled up under a very elastic extenial 

 membrane and to be extruded mainly by its own elasticity. Even in preserved speci- 

 mens, when the ripe uematocysts are broken or cut across, three or four coils ot the 

 thread will uncurl. The thread indeed resembles both in its development and appearance 

 the elastic fibres in the connective tissue of Vertebrates. 



Stomodoeum (Fig. 11). The ectoderm of the stomodoeum is rather thicker than 

 that of the mouth-disc, which gradually merges into it. It is composed of an elongated 

 columnar epithelium with crowded, rod-shaped nuclei, all staining homogeneously. Over 

 the structureless membrane is a thick, finely granular nervous layer— not distinctly seen 

 in the figure— with a few round nuclei and supporting fibres. A few nematocysts like 

 those of the mesenterial filaments but much .smaller are fo\ind ; tentacular nematocysts 

 are extremely rare. 



