290 Journal of Agricullural Research voi. in, no. 4 



bark was thick these burrows were quite generally transverse to the grain 

 of the wood. This condition, as well as the oblique course of some of the 

 smaller burrows, is well shown in Plate XXXIX, figure 3. 



At the close of the fourth instar the larva burrows out into the bark, 

 if it is thick enough, and constructs a cell in which it hibernates. Here 

 pupation takes place in the spring. These cells are found in the ridges of 

 the bark on the trunk and larger limbs of the tree and in the wood on 

 small, thin-barked trees and limbs. In constructing the cell, the larva 

 burrows out to within a few millimeters of the surface of the bark, with- 

 draws itself 2 or 3 mm., then turns about to one side and excavates around 

 the posterior portion of its body until an oblong cell has been constructed. 

 The portion of the burrow leading to the cell, as well as the short portion 

 between the cell and the bark, is plugged with the frass loosened in mak- 

 ing the excavation. It is evident that at least the portion of the frass 

 at the outer end of the cell was never ingested, for if this were the case it 

 would have passed out at the anus, which has been at the other end of the 

 cell all the time. When the cell is complete, it measures about 10 mm. in 

 length and about 2 or 3 mm. in width and contains the larva bent upon 

 itself, ready for hibernation (PI. XXXVIII, fig. 6). 



The work of*the larval life is well illustrated in Plate XXXIX, figure 4, 

 which is a reproduction of a photograph of a limb in which the entire 

 burrow is traced, from the shells of the eggs in a crevice of the bark to the 

 larv'a in its pupal cell. The burrow was carefully marked with india ink, 

 so that the black lines represent the exact width of the burrow in every 

 case, and at the places where the larva entered the wood to molt the holes 

 have been marked about with white ink for the sake of contrast. The 

 larva in this case burrowed into the wood for a short distance at first, 

 then returned to the cambium layer, where it burrowed about until it 

 reentered the wood to molt. From the point where the larva entered 

 the bark to the place it emerged from the wood after the first molt the 

 burrow measures 69 mm. in length and 270/1 in width. 



It will be noticed that in each succeeding instar the burrow is nuich 

 wider and longer, so that in the second instar the length is 103 mm. and 

 the width is goo/t; in the third instar it is 210 mm. long and 1.21 to 1.56 

 mm. in width; and in the fourth and last instar the length is 456 mm. 

 and the width is 1.96 to 2.15 mm. The total length of this particular bur- 

 row is 835 mm., or nearly 3 feet. Since the burrows made during the 

 early instars are so small that they are hard to find, it seems likely that 

 Chittenden's statement ' that the complete burrow is only 6 to 10 inches 

 in length was based on observations made on burrows which represented 

 the last instar. Even these could hardly have been complete, for burrow-s of 

 this instar nearly 2 feet long have been found. His statement that they are 

 for the most part transverse to the grain of the wood makes it seem even 

 more evident that those of the last instar were described, for, as Plate 



' Chittenden, F. H. Loc. cit. 



