235 



live at first like the other upon Polygonum and Setaria, but is most com- 

 monly transferred to corn, either in the same field or by flying to a dis- 

 tance. It is beyond all possible question a fact that the ants burrow the 

 hills of corn industriously in advance of the appearance of these lice, 

 and when they themselves have none in their i^ossession with which to 

 stock their burrows. That they will eagerly seize and convey to their 

 cornfield formicaries root lice exposed to them we have repeatedly dem- 

 onstrated by experiment. They seem, however, to be not wholly depend- 

 ent upon this louse for food, since in the early spring, before the root lice 

 make their start, the Lasius often captures small larvae and various 

 soft-bodied insects, which it kills and carries to its nest. 



The third generation may appear from May 15 to 20. It is more gen- 

 erally winged than the second, so far as is indicated by our rather scanty 

 observation. 



The fourth and fifth generations were brought out late in May and 

 early in June in the single experiment which we carried to that length, 

 but the subsequent history of the louse has not been followed through 

 the season in detail. It is only certain that successive broods appear 

 throughout the summer until fall, breeding continuously- upon the roots, 

 and both winged and wingless viviparous females occur in variable pro- 

 portions, seemingly determined m part by the condition of the plant 

 upon which they feed, the winged lice being rapidly evolved as the corn 

 plant suffers from the attack. The midsummer generations become, 

 consequently, widely scattered, and the lice may almost disappear in 

 fields where earlier in the season they were excessively abundant. Colo- 

 nies started in old cornfields which have been planted to some other 

 crop thus abandon them, after living for a time on smartweed, pigeon 

 grass, and the like, and resort to the growing corn; but in midsummer 

 the roots of many other plants become infested — purslane, Panicum, 

 Setaria, and possibly squash, although our attempts to breed on corn 

 root lice from this last plant were quite unsuccessful. 



Even ragweed will support these lice at least temporarily, as we 

 proved in 1889 by transferring half-grown young of the second genera- 

 tion from smartweed roots to ragweed, where they lived and fed until 

 they acquired wings, five days later. In autumn we have found the 

 last viviparous generation and the oviparous female following, on purs- 

 lane, dock {Rumex crispa), fleabane {Erigeron canadense), black mustard, 

 sorrel (Oxalis stricta). and the common plantain [Plantago major)^ not 

 to mention two other plants not recognized or determined at the time. 



The bisexual generation of root lice makes its appearance in corn- 

 fields as early as October 1, and continues there throughout the month, 

 pairing and depositing eggs. Our only observation of the sexes in cop- 

 ula was dated October 21. We have not found the oviparous female 

 anywhere in the earth except in the burrows of ants, and there, doubt- 

 less, the eggs are laid. Certainly they are collected at that season in 

 the chambers of the ants' nests, and carried through the winter there 



