474 IV'EW YORK STATE MUSEUM 



The larvae live at the bottom of streams of rapid water in the 

 swiftest parts, under stones. They readily feed on soft bodied 

 caddis worms, Sialis larvae, very young Chauliodes larvae, 

 younger members of their own species; and doubtless have a 

 wide range of food habits. I have succeeded in getting hungry 

 larvae to eat bits of fresh beef by placing them in a traA' of 

 water in a photographic dark room. They do not feed well when 

 exposed to bright light; and they seem to prefer live food. la 

 the dark room they will sometimes eat large dipterous larvae. 



They spend the winter some distance below the bed of th'." 

 stream buried in the sand and gravel. 



The larvae usually crawl when they care to move about in 

 the water, but thoy can swim backward readily and sometimes 

 are found to swim forward. 



Nothing very definite is known as to the number of molts or 

 as to the length of the larval period. The same problems are 

 here involved as those stated on a preceding page in speaking 

 of Chauliodes larvae. I have kept larvae of Corydalis over 

 winter in running water in dark cells made of flowerpots. Out 

 of 28 which were kept alive in the cells for nine months, only 

 two were found to molt, and these do not warrant the drawing 

 of any general conclusion. In September 1899 I took from one 

 locality in Fall creek, Ithaca, over 100 larvae and measured 

 their heads. The range of sizes was so gradual that it gave no 

 clue to the probable number of years represented in the lot. 



Though the larvae naturally live in beds of streams till full 

 grown, they are capable of living out of water in moist soil for 

 an indefinite length of time. I thus kept them in a breeding 

 cage in a greenhouse for over four months, at the end of which 

 they were accidentally killed. The larval spiracles are doubt- 

 less open and functional, at least when the insect is out of water, 

 and may be used for breathing from air which collects under 

 stones in running water. 



The mature larvae leave the water in May or June and pupate 

 in cavities under fiat stones near the stream. xVt times they 

 crawl for many rods and even up high banks before selecting 



