A STUDY OF GOMPSILURA CONCINNATA. 23 
ing the following fall and spring, and when the season began they 
were placed in feeding trays which were kept covered with fine mos- 
quito screening. These were fed until the hosts pupated, and al- 
though this experiment was repeated for two years, the trays being 
carefully examined at least once a week during the time of feeding, 
no Compsilura were secured. 
The results of these experiments substantiate the statement referred 
to in the foregoing bulletin, that it is impossible for Compsilura 
larve to hibernate in overwintering brown-tail moth: larve. Dis- 
- sections during the fall of 1915 showed conclusively that Compsilura 
attacks the young brown-tail moth larvee and will live through part 
of the first larval instar in this host, but that the small size of the 
host prevents the parasite larva from maturing sufficiently to pupate. 
SECONDARY PARASITISM. 
Secondary parasites of Compsilura play an important -part in the 
spread and effectiveness of the tachinid parasite. These secondaries 
attack the Compsilura maggot immediately following its emergence 
from the host, or the fresh puparia, and before it hashardened. From 
no puparia secured from beneath the surface of the soil have sec- 
ondary parasites emerged, only those found above ground being at- 
tacked. During the seasons of 1915 and 1916 1,164 Compsilura 
puparia were collected in various parts of New England over the 
entire area covered by this parasite, and from 10.31 per cent of them 
secondary parasites issued. 
SUPERPARASITISM. 
To ascertain the effectiveness of Compsilura, a series of experi- 
ments was conducted from 1912 to 1916, with gipsy-moth larve col- 
lected on the border towns of Compsilura dispersion for the years 
1909 to 1913, five towns being selected in which the parasites were first 
recovered in the five-year period mentioned. These border towns 
have furnished the host material each year for the last four years, 
and the collections yielded an average parasitism of 10.21 per cent. 
These results were secured by making collections of from 10 to 20 
ufth-stage gipsy-moth larvee and feeding them singly, either in small 
trays or in screened boxes, care being exercised to safeguard them 
‘rom any parasitism after reaching the laboratory. 
It was from these same experiments that the highest parasitism by 
Compsilura was gained, and in several cases three puparia were se- 
cured from a single host larva. Data were also obtained from the 
foregoing experiments on the effect of wilt on Compsilura parasitism. 
If the parasite larva is ready to molt into the last larval instar, al- 
though the death of the host occurs from wilt, it will not prevent 
