A STUDY IN HYPERPARASITISM Pt 
age that would have occurred during the winter from other causes. 
Furthermore, no cocoons having exit holes (and there were many from 
which secondaries had already issued) were collected. Strictly, these 
should be taken into account in computing the proportion of cocoons 
carrying over living Apanteles larvae, and also the proportion yield- 
ing hyperparasites. On considering these various phases of the sub- 
ject 1t appears that in all probability not more than one-half of 1 per 
cent of the cocoons formed in July produce adult Apanteles the fol- 
lowing spring. It is possible that the proportion is frequently even 
ess. 
The table indicates that the percentage of cocoons producing 
neither hyperparasites nor primaries steadily increases as the season 
progresses. As already suggested, there are probably two principal 
causes for the failure of cocoons to yield either Apanteles or secondaries, 
namely, excessive hyperparasitism, and the feeding of the adult sec- 
ondaries upon the Apanteles or hyperparasitic larvae within the co- 
coons. The proportion of such cocoons, as would normally be expected, ~ 
mcereases week after week during the late summer and early fall when 
the hyperparasites are most active and most abundant. That the 
enormous mortality suffered by the primary parasite, in the case of 
the collections, is ascribable particularly to hyperparasites, and not 
in any appreciable degree to low winter temperatures, is indicated by 
the fact that more than 95 per cent of the cocoons produced at the 
laboratory and held under the same conditions as those obtained 
from the field, produced Apanteles adults in the spring. 
The data on the issuance of Apanteles from the many thousands of 
cocoons collected in July in connection with the regular reproduction 
work on this species at the laboratory are interesting. Figures are 
at hand for the collections of 1920 to 1923, inclusive. In 1920 the | 
~ number of cocoons collected was 11,971; in 1921, 12,081; and in 
1923, 29,499. These produced, respectively, 958, 2,650, and 1,290 
Apanteles adults. In all these cases the cocoons had been in the 
field from one to four weeks before the collections were made. As 
a result of being exposed to hyperparasites for this period, only 9 
per cent of the total number of cocoons taken in those years yielded 
adults of the primary species. In 1922 the cocoons were collected 
very early, as soon as possible after they were formed; consequently 
15,866 cocoons of the 38,855 collected that year produced Apanteles. 
Unfortunately, complete figures on the numbers yielding hyperpara- 
sites are not available. 
In a review of the above discussion the following points are parti- 
cularly impressive: (1) The first-generation Apanteles cocoons are 
relatively lightly parasitized; (2) during their long period of exposure 
the Apanteles of the second generation have a steadily diminishing 
chance of becoming adults, and probably not more than one-half of 
1 per cent of the cocoons formed in July produce adults of the pri- 
mary species the following spring; (3) an extremely high proportion, 
probably close to 50 per cent, of the total number of second-generation 
cocoons formed will yield neither primaries nor secondaries. 
It is quite apparent that A. melanoscelus would find it difficult 
to survive, or at least would be unable to maintain its effectiveness 
aS an enemy of the gipsy moth, if it did not have two generations 
annually. On the other hand, if not severely checked during the 
second generation, the species would multiply excessively. The 
