CHAP. XVIII. BESEKVOIRS FOE WATER. 439 



the cavity beneath the bowed antennae, and then force 

 their way through the little slit-like orifice between 

 the valve and collar into the bladders filled with 

 water, I cannot conjecture. 



Tubers. These organs, one of which is represented 

 in a previous figure (fig. 26) of the natural size, 

 deserve a few remarks. Twenty were found on the 

 rhizomes of a single plant, but they cannot be strictly 

 counted ; for, besides the twenty, there were all pos- 

 sible gradations between a short length of a rhizome 

 just perceptibly swollen and one so much swollen that 

 it might be doubtfully called a tuber. When well 

 developed, they are oval and symmetrical, more so 

 than appears in the figure. The largest which I 

 saw was 1 inch (25'4 mm.) in length and '45 inch 

 (11-43 mm.) in breadth. They commonly lie near 

 the surface, but some are buried at the depth of 

 2 inches. The buried ones are dirty white, but those 

 partly exposed to the light become greenish from the 

 development, of chlorophyll in their superficial cells. 

 They terminate in a rhizome, but this sometimes 

 decays and drops off. They do not contain any air, 

 and they sink in water; their surfaces are covered 

 with the usual papillse. The bundle of vessels which 

 runs up each rhizome, as soon as it enters the tuber, 

 separates into three distinct bundles, which reunite 

 at the opposite end. A rather thick slice of a tuber is 

 almost as transparent as glass, and is seen to consist 

 of large angular cells, full of water and not containing 

 starch or any other solid matter. Some slices were 

 left in alcohol for several days, but only a few 

 extremely minute granules of matter were precipitated 

 on the walls of the cells ; and these were much smaller 

 and fewer than those precipitated on the cell-walls of 

 the rhizomes and bladders. We may therefore con- 

 29 



