356 SOME DIFFICULTIES IN THE LIFE OF AQUATIC INSECTS. 
stream, while the open mouth is turned down stream, the pupa rests 
securely during the time of its transformation. 
When the pupa ease is first formed it is completely closed and egg- 
shaped, but when the insect has cast the larval skin one end of the 
case is knocked off, and the pupa now thrusts the fore part of its body 
into the current of water. The respiratory filaments, which project 
immediately behind the future head, just as in Chironomus, draw a suf- 
ficient supply of air from the continually changed water around. The 
rings of the abdomen are furnished with a number of projecting hooks, 
which are able to grasp such objects as fine threads. The interior of 
the cocoon is felted by a number of silken threads, and by means of 
these the pupa gets an additional grip of its case. If it is forcibly dis- 
lodged a number of the silken threads are drawn out from the felted 
lining of the case. The fly emerges into the running water, and I do 
not know how it manages to do so without being entangled in the ecur- 
rent of water and swept down stream. The pupa skin splits open just 
as it does in Chironomus, but remains attached to the cocoon. 
The larva of the gnat is perhaps more familiar to naturalists of all 
kinds than any other aquatic dipterous insect. The interesting 
description, and above all the admirable engravings of Swammerdam, 
now more than two hundred years old, are familiar to every student of 
nature. 
The larva, when at rest, floats at the surface of stagnant water. Its 
head, which is provided with vibratile organs suitable for sweeping 
minute particles into the mouth, is directed downwards, and when 
examined by a lens in a good light appears to be bordered below by a 
gleaming band. There are no thoracic limbs. The hind limbs, which 
were long and hooked in the burrowing Chironomus larva, and reduced 
to a hook-bearing sucker in Simulium, now disappear altogether. A 
new and peculiar organ is developed from the eighth segment of the 
abdomen. This is a cylindrical respiratory siphon, traversed by two 
large air tubes which are continued along the entire length of the body 
and supply every part with air. The larva ordinarily rests in such a 
position that the tip of the respiratory siphon is flush with the surface 
of the water, and thus suspended, it feeds incessantly, breathing unin- 
terruptedly at the same time. When disturbed it leaves the surface 
by the sculling action of its broad tail. Once below the surface it sinks 
slowly to the bottom by gravity alone, which shows that the body 
is denser than the water. We have therefore to explain how it is 
enabled to float at the surface when at rest. The larva does not will- 
ingly remain below for any length of time. It rises by a jerking move- 
ment, striking rapid blows with its tail, and advancing tail foremost. 
When it reaches the top it hangs as before, head downwards, and 
resumes its feeding operations. 
In order to explain how the larva hangs from the surface against 
gravity, I must trouble you with some account of the properties 
