123 MAUD D. HAVILAND 
may in some degree protect them from ovipositions by the 
females of their own and other families. 
When ready to emerge, the imago gnaws a hole somewhere on 
the dorsal side of the cocoon and creeps out. As Gatenby (9) 
has remarked, this hole differs from that made by Aphidius 
in having irregular edges, and is not necessarily placed in the 
dorso-posterior region of the aphid’s skin. 
he number of broods occurring in one year is not known, and 
probably depends on the number of species of Aphidius 
upon which the hyperparasites can live. Two broods were 
reared from Aphidius ervi in 1919; but the host did 
not appear in any numbers before July, and it is possible that 
earlier broods may have occurred with a different host. All the 
imagos of Lygocerus had emerged by the end of August, 
and there is no evidence to show whether the species over-winters 
us larva or pupa. 
In captivity the imagos generally live five or six days, but 
sometimes as long as ten. They were observed to feed on 
sugar and water, on honey-dew from the aphides, and on sap 
oozing from cut leaves, but they seemed to live as long, and to 
remain as vigorous, when no food was supplied. 
COMPARISON OF LARVAL CHARACTERS WITH THOSE 
OF OTHER SUB-F'AMILIES. 
The most complete comparative account of the larvae of 
entomophagous Hymenoptera is that of Sewat (26), who 
studied certain Ichneumonidae, Braconidae, and Chalcidae. 
Unfortunately he did not include the Proctotrypidae, and our 
knowledge of the larval morphology of this family, as already 
remarked, is very scanty. Seurat emphasized the importance 
of the tracheal system in determining the larvae of the different 
sroups, but, as Lichtenstein and Picard have recently pointed 
out (15), increased knowledge has somewhat modified this view. 
Some authorities have considered that the Proctotrypoidea 
are allied to the Chalcidoidea, but Ashmead (1) disputes this, and 
thinks them m every respect more nearly related to the Hymen- 
