Packard.] INSECTS OF THE POND AND STREAM. 137 
most of the Syrphidae are Aphis-eaters in the larval state. 
Of that immense family Muscidae, of which the house fly is 
a tj T pe, which are almost purely terrestrial, Ephydra (Fig. 
98, fly and pupa case) is aquatic. It may be said on the 
whole that the aquatic larvae of the Diptera were originally 
terrestrial insects, which have adopted an aquatic life and 
are exceptions to the rule. 
In the Neuroptera, however, there are whole families which 
are aquatic both in the larval and pupal state, having exter¬ 
nal gill-like appendages in both stages, and in the case of 
Fig. 99. 
Perla. 
Pteronarcys, which is closely allied to Perla (Fig. 99, from 
Figuier), they are retained in the winged fly. The Neurop¬ 
tera are pciv pxcgIIgiicg water insects, and yet it is doubtful 
even whether they have not originally assumed this excep¬ 
tional mode of life, and, while the earliest of all insects, were 
not at first terrestrial. This is speculation and guesswork, 
but facts seem to point to this conclusion. 
Now the changes in structure fitting the insect for a life 
in ponds and streams are to be found in the organs of loco¬ 
motion and the breathing apparatus, and in this essay I 
9 
