JNTERNAL ANATOMY OF INSECTS. 47 
and becomes ventral. Generally there is a pair of spi- 
racles to each segment, and in those insects that have a 
hypochondriack joint a there is often a spiracle in it. The 
last segment of the abdomen is always without these 
orifices, as is the basal one in Velia, Ranatra, and some 
other bugs. A singular anomaly distinguishes the Li- 
bellulina : they appear to have no abdominal spiracles b , 
yet I have seen the abdomen of Libellula depressa when 
reposing, contract and dilate alternately, from whence it 
follows that this part is concerned in respiration. Spren- 
gel says that the larvae in this tribe have seven or nine 
on each side c , and Reaumur speaks of them as disco- 
verable in the pupa d . I have carefully examined the 
pupa-skin of most of the genera of Libellidina, under a 
powerful magnifier, but have not succeeded in discover- 
ing any thing like these organs in the abdomen. The 
Ephemera and probably the other Neuroptera have ab- 
dominal spiracles e . M. Latreille observed one on each 
side of the base of the scale on the footstalk of the ab- 
domen in ants f . Generally the abdominal spiracles may 
be described as planted in the crust of the insect ; but in 
many cases their station is in the membranous folds, 
which I have therefore named the puhnonarium, that 
sometimes separate the dorsal from the ventral segments : 
these folds allow of a considerable distention of the ab- 
domen, which is probably necessary when all the air- 
vessels are full. In a gravid Ichneumon I once saw it en- 
larged to more than twice its natural size by means of 
this membrane, through which the eggs were distinctly 
a Vol. III. p. 708. b Sprengel, Comment. 3. 
Ibid. " vi. 398. 
e De Geer, ii. 635. f Fourmis, 22. 
