2 62 
TRANSFORMATIONS OF INSECTS. 
perfect insects either fly in the air or crawl on dry land. Immer- 
sion in water is fatal to them after a time. The larvae of many 
Ichneumons appear to exist surrounded by the juices of their 
victims, and they must respire either by means of the air passing 
through the air tubes of the insects they live in, or by the more 
or less aerated fluids which form their nourishment. But the 
majority of the hymenopterous larvae respire by spiracles, and 
cannot live under water. 
A discovery of Sir John Lubbock’s is, therefore, very inte- 
resting. He observed many specimens of a small hymenopterous 
insect swimming in pond water with their wings, and quite at 
their ease. The fly belonged to a genus the larvae of which are 
eminently parasitic in the eggs of some Lepidoptera ; and the 
accomplished naturalist just mentioned considers it highly pro- 
bable that the slow-flying swimming insect was in search of some 
aquatic victims for its progeny. He found a second species, 
which used its legs as oars, but did not swim with its wings ; 
and he noticed that both kinds of these Polynemce were fond of 
creeping along the sides of the vessel in which they were placed, 
or on the leaves and stems of aquatic plants. They frequently 
swam boldly out into open water. The history of the metamor- 
phosis of these aquatic Hymenoptera has yet to be written ; and 
it is evident that its comparison with that of the Hydrocampidce 
amongst the Lepidoptera will form a most interesting inquiry. 
The metamorphoses of the Hymenoptera differ from those of 
the Lepidoptera , in consequence of the larva remaining for a 
greater or less period inside a cocoon before transforming into a 
pupa. Instances of retrograde metamorphosis appear to occur 
in the genus Cynips, and the wingless condition of the female 
members of the species of Cynips terminalis , and the absence of 
wings in Cynips aptera, are usually stated as examples. But 
the metamorphosis can hardly be said to be retrograde in the 
strict sense of the term, for the perfect female of Cynips terminalis 
and the adults of both sexes of Cynips aptera are more fully 
developed than their larvae and nymphs, and possess additional 
structures. In no sense can they be said to be lower in the insect 
scale. 
The transformations of these insects are not then retrograde, 
