1914] Life History of Hydromyza Confluens. 141 
THE PUPA. 
The detailed description of this stage is reserved for another 
paper which is in preparation. 
Position.—The position of the pupa in the gall is variable, 
the only constant feature being the fact that its long axis 
always lies almost or quite parallel to the long axis of the 
petiole. However, the position of the head and the caudal 
end is not at all constant since in some cases the head is up 
(towards the surface) and in others it is down (towards the root- 
stalk). In galls which are well advanced in development and 
contain pupz this point can easily be determined from the 
exterior without breaking into the gall by noting the position 
of the window, a feature to be described later. When the 
window occurs at the lower end of the gall it is positive proof 
that the head of the full-grown larva or the head of the pupa is 
down; if the window is at the upper part of the gall, the head 
of the larva or pupais up. Ina series of observations in which 
242 pupz were examined it was found that 130 occupied a 
position in which the head of each was up and 112 in which the 
head of each was down. Other statistics of the same kind 
showed a similar result, namely, that the majority of the 
pupz lie in the galls with the head towards the surface of the 
water. 
The Window.—The window mentioned in the preceding 
paragraph is a very interesting and unique provision for the 
emergence of the adult and, as stated by Needham, is con- 
structed by the larva immediately before pupation. It is 
circular in outline and only large enough to allow the passage 
of the emerging adult. In constructing this window the 
full-grown larva works towards the exterior of the petiole until 
it reaches the epidermis. Here it removes all of the surrounding 
tissue (exclusive of the epidermis) from a circular area which 
is destined to be the window so that the latter is composed 
only of epidermis. A circular incision, which extends around 
approximately two thirds of the circumference, is made along 
the periphery of this area. The remaining one-third is left 
intact and thus a circular lid, attached at one side, is produced. 
The attached portion always has a definite relation to the 
position of the pupa, namely, it is constantly on the side of 
