INTERNAL ANATOMY OF INSECTS. 59 
emit it*. This brisk motion probably disengages it from 
the water. In many species, when in repose, they are 
laid upon the back of the animal '", but in others they are 
not c . 
The larvae of the Agrionidce appear to respire like 
those of the Ephemera, &c. by means of long foliaceous 
laminae or false gills filled with air-vessels ; but instead 
of being ventral, they proceed from the anus. They are 
three in number, one dorsal and two lateral, perpendi- 
cular to the horizon, of a lanceolate shape, beautifully 
veined, with a longitudinal middle nervure, from which 
others diverge towards the margin, which are probably 
bronchia. They are used by the animal, which swims 
like a fish, as fms, but it does not appear to imbibe the 
water like the other Libelhdhia?, nor to propel itself by 
ejecting it, — a circumstance which furnishes an additional 
argument for the more received opinion, that this action 
in them is for the purpose of respiration as much as for 
motion d . 
The larvae and pupae of the Libellulince, receive the 
water and air that they respire by a large anal aperture, 
which is closed at the will of the animal by five hard, 
moveable, triangular, concavo-convex pieces, all very 
acute and fringed with hairs. These pieces are placed 
so that there is one above, which is the largest of all ; 
one on each side, which are the smallest, and two below; 
when these are closed they form together a conical point 6 . 
Sometimes only three of these pieces are conspicuous f : 
a Reaum. vi. 465. » Ibid. t. xlii./. 4, 5. De Gecr ii. 023. 
c Ibid. 648. t. xvii./. 11, 12. 
11 Vol. III. p. 154. De Geer ii. 697—. t. xxi./. 4, 5, 12. 
c De Geer Ibid. 606—. /. xix./. 6. 
' Reaum. vi. 393. /. xxxvi./. 8, 9. t, t. 
