V] 



OF COLLAR-CELLS 



253 



which reminds us of the cylindrical spiral of a Spirillum among 

 the bacteria. In Dinenympha we have a symmetrical figm-e, whose 

 two opposite surfaces each constitute a surface of constant mean 

 curvature ; it is evidently a figure of equilibrium under certain 

 special conditions of restraint. The cylindrical coil of the 

 Spirillum, on the other hand, is a surface of constant mean curva- 

 ture, and therefore of equilibrium, as truly, and in the same sense, 

 as the cylinder itself. 



\ 



Fig. 82. Dinenympha gracilis, Leidy. 



A very curious conformation is that of the vibratile "collar," 

 found in Codosiga and the other "Choanoflagellates," and which 

 we also meet with in the "collar-cells" which line the interior 

 cavities of a sponge. Such collar-cells are always very minute, 

 and the collar is constituted of a very delicate film, which 

 shews an undulatory or ripphng motion. It is a surface of 

 revolution, and as it maintains itself in equilibrium (though a 

 somewhat unstable and fluctuating one), it must be, under the 

 restricted circumstances of its case, a surface of minimal area. 

 But it is not so easy to see what these special circumstances are ; 

 and it is obvious that the collar, if left to itself, must at once 



