468 MAUD D. HAVILAND 
The rest of the internal structure demands no particular 
comment. 
The tracheal system becomes functional when the parasite 
leaves the host. The two main lateral trunks are united by an 
anterior and a posterior commissure. Dorsal and ventral 
lateral branches are given off in each segment 1-10. There 
are six pairs of open spiracles. The first 1s placed between 
segments 1 and 2, and the remainder on segments 3, 4, 5, 7, 
and 9. Of the considerable number of examples examined 
only two departed from this rule in possessing, in addition, 
a pair of spiracles on segment 8. 
PUPATION AND HMERGENCE. 
Pupation lasts from twenty-two to twenty-six days, and 
at the end of this time the Cynipid gnaws an irregular hole 
on the dorsal side of the cocoon and creeps out. In captivity 
the adults lived from three to eight days. They fed upon the 
sap oozing from cut leaves and upon the honeydew of the 
aphides. ‘They sometimes sipped the latter from the anus of 
the living animal, and were occasionally observed to scrape 
the dried sugar from empty skins with their mandibles. 
It is not known how many broods may be reared in the 
season, nor how far these Cynipid hyperparasites are specific 
for different Aphididae, but as far as it goes the evidence 
suggests that they have a considerable range of hosts. Thus 
the number of broods is probably determined by the number 
of Aphididae available. . 
Also at present there is no evidence as to how the parasites 
and hyperparasites of Aphides pass the winter. J have found 
living larvae of Aphidius salicis, Hal, in Aphis 
saliceti, Kalt., in cocoons collected in July, and opened in 
the laboratory in January. This suggests that a few may pass 
through the winter in this stage; but, although I paid parti- 
cular attention to this point, I could find no indication that 
Aphidius ervi had not all emerged by the end of August, 
for, of the considerable number of cocoons from different 
localities that were examined, all were empty. 
