AQUATIC INSECTS IN NEW YORK STATE 311 



for, no matter at what level sections are taken, the dorsal sur- 

 face is always found to be convex, as is represented on plate 28, 

 figure 4. 



In order to explain how the larva of Donacia obtains its supply 

 of air from the intercellular spaces of plants, I do not think it 

 is necessary to assume any extraordinary structures for the 

 caudal spines. The caudal spines are nothing more than pro- 

 jections of the body wall for rupturing the tissues of the plant; 

 and, when this is accomplished, the air, being so much lighter 

 than the surrounding water and having a strong tendency to fol- 

 low along anything that will carry it to a higher level, simply fol- 

 lows along the outer surface of the caudal spines to their base, 

 where it is taken up by the spiracles, while the two large longi- 

 tudinal trachea connecting with the spines take up the supply 

 of air and act as reservoirs for storing it between the air-taking 

 periods. 



When the larvae are ready to transform to pupae, they spin 

 a tough, brownish cocoon, which is attached to the scars on the 

 upper surface of the rhizome from which the leaf stalks have 

 been shed. The silk is spun from glands opening in the 

 mouth. The cocoons are not only water-tight but air-tight 

 and are of a homogeneous consistency throughout without any 

 indication of a thread-like structure. The bottom of the cocoon 

 where it is attached to the plant is much thinner and lighter in 

 color and is firmly glued to the surface of the plant [pl.22o]. 

 In the case of p a 1 m a t a the cocoons are much smaller than 

 the rhizome and always placed on its upper surface, distant 

 from each other; but in those species that pupate on the roots 

 of Sagittaria, Sparganium and Potamogeton, the cocoons are 

 much broader than the roots and are placed singly, usually near 

 the point of attachment of the roots to the base of the plant. 

 Such a condition is shown well by plate 23, figure h; while in 

 the case of Donacia cincticornis, which fastens its 

 cocoons to the larger roots of Nymphaea advena, they 

 are arranged close together in groups of five or six around the 

 apices of the roots. 



