OF AQUATIC LARV^. 
87 
noticed that there are five small, apparently useless points on 
the ninth segment, and the natural inference will be that the 
two succeeding segments have been invaginated in Calofteryx, 
carrying the terminal plates with them, thus forming the enclosed 
sac with the breathing-plates contained in it. 
In the case of Libellula, the plates have gradually increased 
in number until they have attained their present profusion, the 
five points of Agrion having become modified into a guard for 
the valves, and a concentrator for the jet of water used in 
propulsion. 
It may, perhaps, be interesting to state here that, in casting 
its skin, the larva throws off* the external skin of this sac 
intact, all the numerous plates being distinctly recognizable in 
the cast-off* skin, and still occupying the same position at the 
lower end of the abdomen. 
This fact appears to corroborate the foregoing opinion that 
the sac with its hranchicd in jEschna, &c., represents the two 
terminal segments of Agrion. 
The Larva of the EpJiemericlm are in many respects similar 
to those of the LdbellulidcB ; their breathing branchim are arranged 
down the sides of the abdomen, two or four gills being present on 
each segment ; these are connected from the two main tracheae 
which run down either side of the body, by short tubes running 
directly into them, and then subdividing into the usual capillary 
tubes. These gills are constantly waving about in the water, 
absorbing oxygen, and giving out carbonic acid formed in the 
body. 
In these beautiful creatures the tracheal system may be 
followed into the most remote points of the body, owing tO 
their transparency, especially when young. Sir John Lubbock 
states, that Cloeon has not these gills when young, and that it 
breathes by osinosis through the skin itself until the fourth 
moult, when the gills appear ; but still there are no tracheae till 
