376 Illinois Slofe Lahoialonj of Kalural Hislori/. 



to deposit its mass of eggs. At such times it frequently finds 

 its way into cellars. The eggs are generally attached to sticks 

 or dead vegetation and are surrounded by a translucent gela- 

 tine. A little later the larvae are abundant in these pools, 

 feeding, as I find by dissection, largely on Crustacea, which the 

 gill-rakers on the branchial arches enable them to collect. The 

 stomachs are sometimes packed with Daphnia puhx. The 

 smallest examples examined — al)Out an inch long — had eaten 

 nothing l)ut animal food. They lose their branchia? and leave 

 the water l)efore the close of the summer, and many of the 

 pools in which they breed are soon after dried up. They are 

 not often seen afterwards until the fall rains set in, when 

 they again appear in cellars and under porches, evidently 

 searching for a place to pass the winter. Their movements on 

 land are very clumsy and their migrations to and from the 

 water seem to be performed at night. A large specimen kept 

 in an aquarium cast its skin tivlce between the 12th and 18th 

 of October. I am disposed to believe, from facts given below, 

 that this species remains, where conditions are suitable, or, per- 

 haps, unsuital)le, to its complete development, longer than one 

 season in the larva state, and may even breed in that state ^ 

 As l)efore stated, the young ordinarily leaves the water when in 

 the neighl)orhood of four inches long, and specimens four and 

 a half inches long have all the essential adult characters. 

 There are now in the Laboratory collection, hovvever, two spec- 

 imens measuring seven and three fourths and eight and three 

 eighths inches respectively, which retain in a remarkable de- 

 gree the larval characters, and had doulitless Init recently left 

 the water when they were captured. The tongue of these 

 examples is very small, elongate, and occupies no more than 

 half the space l)etween the rami of the mandil)le. The man- 

 dibular rami are less arched and less widely divergent than in 

 adults; the palatine teeth are more strongly arched forwards; 

 the rudiments of 1)ranchitB persist; the digits are strongly de- 

 pressed; and the membranous expansion is still present for a 

 short distance on the tail above, and sharp grooves indicate its 

 recent resorl)tion at other points. The colors also are those of 

 a recently transformed larva. One of these examples proves 

 to be a male with well-developed sexual organs. Unfortu- 



