SEA-AJSTEMONES : THEIR WEAPONS. 413 



j^jth of an inch in diameter. I slightly toucheJ the 

 animal, and it in an instant enlarged the apcrtui'e 

 to 2^oth of an inch. In a specimen of S. hellu, less 

 than half-grown, I found the cinclides numerous, and 

 sufficiently easy of detection, but rather less detined 

 than in cUanthus or nivea. They occurred at about 

 every fourth intersept, three intersepts being blind 

 for each perforate one, and about three or four in 

 linear series, but not quite regularly in either of these 

 respects. In this case they were about ■^\\\i of an 

 inch in transverse diameter — a large size ; and I 

 measured one which was even 5^^^^ ^^ ^^^ inch. By 

 bringing the animal before the window, I could dis- 

 cern the light through the tiny orifices with my naked 

 eye. 



From several good observations, and especially 

 from one on a cinclis, widely opened, that happened to 

 be close to the edge of the parapet of a dianthus, I 

 perceived that the passage is not absolutely open, at 

 least in ordinary, but that an excessively thin film lies 

 across it. By delicate focussing, I have detected re- 

 peatedly, in difterent degrees of expansion, and even 

 at the widest, the granulations of a membrane of exces- 

 sive tenuity, and one or two scattered cni'dce, across 

 the bright interval. On another occasion, in the case 

 of a cinclis at the edge of the parapet, a position sin- 

 gularly lavourable for observation, I saw that this 

 subtle film was gradually pushed out until it assumed 

 the form of a hemispherical bladder, in which state it 

 remained as long as I looked at it. At the same time 

 tiic outline of the cinclis itself was sharp and clear, 

 when brought into focus further in. The film, what- 

 ever it be, is superficial, and does not appear to be a 



