50 Annals Entomological Society of America [Vol. Ill, 



sent vcntrad. which divides almost immediately, the one branch 

 proceeding to the head while the other, bending posteriorly, sup- 

 plies the brain and sub-a'sophageal ganglion. This posterior 

 branch is F,, in Plate VI. Fig. 4. 



The two dorsal trunks are connected with the spiracles by 

 short branches and are joined to each other in segments 4-9 and 

 1 1 by cross tubes, that of segment 1 1 being quite stout while the 

 rest are slender and send out two re-dividing branches. These 

 cross- tubes usually lose their striated appearance at their middle, 

 ap]jcaring composed of two tubes fused together by a jjair of 

 spongc-Hke knobs. In segment 10, there is no fusion, the tubes 

 remaining separate. Fig. 6, Eio, PL VI, shows a condition observed 

 in two larvae in which the left branch is quite swollen basally. 

 which perhaps was the normal thickness of both tubes, through- 

 out before atrophy occurred and the tubes separated. The fact 

 thai in C. resinicola there are complete cross tubes in segments 

 4-1 1 would appear to strengthen this view. Furthermore it will 

 be noticed that in C. resinicoloides the cross tube in segment 1 1 is 

 quite stout and may substitute in a measure for the tubes in the 

 above segment. Ramifying tracheae are apparently not so 

 necessary in the last few segments. 



The transverse lateral tubes D., — Djo, Fig. 4, terminate in the 

 spiracles except in segment 3 where there is no spiracle, in which 

 case the tube disappears near the body surface. They are con- 

 nected with one another either by the lateral longitudinal system 

 through the short tube R3, R^, etc., as in segments 3-8, or they 

 may run directly through the former, as in segments 9 and 10. 

 This arrangement, however, is frequently asymmetrical. 



The lateral longitudinal trunks arise from the dorsal ones in 

 segment 2, Fig. 4, A., PI. VI, and terminate as free branches in 

 segment 10, C,i, Fig. 4, and a])parcntly correspond to the ventral 

 tubes as illustrated by Eckel (5), but differ widely from the latter 

 in that they do not join the dorsal trunk posteriorly. They send 

 out many branches F., F3. K^, K^, etc.. l)ut the arrangement of 

 these is not always constant. The primary function of the 

 more anterior of these branches is to aerate and support the ner- 

 vous system, that of the more posterior ones to supply the Mal- 

 pighian tubules. 



The ventral longitudinal system Fig. 5, PI. VI, arises from the 

 main dorsal one close to the anal spiracles in segment 1 1 , and 

 proceeding anteriorly commences ramif\-ing in the middle of 



