122 THE BERMUDA ISLANDS. 



teries for a short distance outwards from their point of attach- 

 ment to the stomatoditeum, and also the outer surface of the lat- 

 ter for a short distance above its inner opening. It looks as if 

 the ectoderm of the stomatodseum were reflected upwards, so as 

 to cover its endodermal surface and the adjacent surfaces of 

 the perfect mesenteries. Further down (PI. 11, fig. 4) the two 

 lateral processes of the mesenterial filaments disappear, the cen- 

 tral one alone persisting. It is evidently the " glandular streak " 

 of the filament. The cells which cover the surface of the mes- 

 entery for some distance outward from this towards the column- 

 wall are very peculiar. (PL 11, fig. 4di.) They form a layer 

 much thicker than that formed by the ordinary endodermal 

 cslls, and are loaded with green granules, closely packed 

 together so that to the naked eye the region occupied by this 

 layer is of that color. Foreign bodies of organic nature are 

 imbedded in the cells, sometimes being surrounded by a num- 

 ber of cells containing no granules, or occasionally imbedded 

 in the mesogloea. 



In unstained specimens, when the animal is laid open by a 

 longitudinal incision, this region of the mesenteries is very 

 distinct on account of its rich green color. When the loose 

 cells of the green area are scraped away with a scalpel and 

 examined, they are seen to be of a very irregular shape (sug- 

 gesting a power of amoeboid movement), and to contain nu- 

 merous green globules, much smaller than the zooxanthellse 

 darker in color, and homogeneous in structure. Amongst the 

 cells are numerous zooxanthellse, and there are also numerous 

 spherical refractive bodies, apparently of a fatty nature and 

 with a slightly-greenish tinge, as well as the foreign bodies 

 already mentioned as seen in the section, and very numerous 

 delicate acicular silicious spicules. 



The. occurrence of these spicules and organic foreign bodies 

 in the cells of this region is very strong evidence in favor of 

 the supposition that they have a digestive function. The 

 green globules may be the products of digestion. If this be 

 the case it is exceedingly interesting, as indicating a method of 



