II FLIES WITH AQUATIC LARVAE 181 



with small vegetable particles, like fine dust, which 

 make them much more apparent. The threads ex- 

 tend in all directions from leaf to leaf, and the larva 

 has access to a perfect labyrinth, along w r hich it can 

 travel to a fresh place by help of the current and 

 with the speed of lightning. I suppose that it grasps 

 the thread with its prothoracic claw r s, for when it 

 comes to rest it is always found holding on by them. 

 To recover its first position is not difficult if the net- 

 work of threads is intact, or if the larva has even a 

 single thread to grasp. Sometimes it hauls itself up 

 hand over hand, like a Leech or a Looping caterpillar, 

 applying its two suckers alternately to the thread. 

 It can also creep along its thread by means of the 

 prothoracic hooks only. I have often seen the body 

 swept into a position at right angles or nearly so 

 with the thread ; then the larva seems to be holding 

 fast by its head. When thus attached, it can travel 

 against the stream at a slow but quite appreciable 

 rate. 



Although the larva commonly slides along a thread 

 previously made, and easily seen to be an old one by 

 the small particles which cling to it, it can upon a 

 sudden emergency spin a new thread, like a Spider 

 or a Geometer larva. The new threads are perfectly 

 clean, and consequently invisible even on a white 

 ground so long as they are submerged. I got proof 

 of the formation of fresh-spun threads by dislodging 

 larvae which had attached themselves to leaves wiped 

 clean a minute before. But the clearest proof is got 

 by suddenly lifting out of the water a leaf with many 

 larvse upon it. One or two are pretty sure to let 



