256 THE CANADIAN ENTOMOLOGIST, 



age, the series from each lateral passage along either side, and each send- 

 ing otY a small branch toward the other, somewhat before reaching the 

 lateral margin, where they terminate. AVhether these tubules are open at 

 the tips or not I have not been able to determine. If so, they doubtless 

 act as a sieve through which the air is admitted to the lateral passages 

 which convey it to the main tracheal trunks. But if we consider them as 

 closed, as I am inclined, the whole structure is remarkably well adapted 

 to aerating the tracheae by osmosis, whether the pure air is secured from 

 the air cells of the plants or from the water. The wedge-shaped apex of 

 each appendage shuts down tightly on either side, thus making a solid 

 cylinder with which to pierce the plant. That it does so pierce the tissue 

 of the root while constructing the cocoon, and that the passage thus made 

 replenishes the air of the cocoon, there can be no doubt. But whether 

 the larva secures air from the intercellular spaces of the root by direct 

 communication or osmosis, or by osmosis from the water, the appendages 

 thus serving as tracheal gills, would seem to need demonstration, inas- 

 much as Dr. Schmidt-Schwedt observed only the points of these append- 

 ages inserted into the roots. 



However that may be, I feel certain that the appendages are truly a 

 highly specialized form of spiracle. I would hardly arrive at this con- 

 clusion had I not observed a very similar structure in the pupas of the 

 genera Octotoma and Odontota of the tribe Hispini. The larvre of these 

 species mine within leaves, and the pupce remain within the leaves. 

 Projecting caudad from either fifth abdominal spiracle — which is usually 

 the last in Chrysomelid pup?e — is found a stout, chitinous spine about 

 the length of a body segment. In the pupa of Octotoma plicatula the 

 fourth spiracle is expanded caudally about half as much as the fifth, and 

 the third is but slightly expanded, merely being produced to a point 

 caudally. But the gradation is complete, and it is easily seen that the 

 spine-like process of the fifth segment is but an outgrowth of the spiracle. 

 Each of these spiracles, 3 to 5, has the external opening surrounded by a 

 circular tube, also connecting with the trachea, and this circular tube is 

 merely drawn out to a point, so to speak, to form the process of the fifth 

 segment, the process gradually increasing in length and acuteness from the 

 second to the fifth abdominal spiracle. This appendage forms merely a 

 simple tube with the sides curled up and in to form an elongate cavity, in 

 which the lining surface is membranous and finely reticulated. Further 

 than this I was unable to observe any structure, as the projections are 



