42 BULLETIN 647, U. S. DEPAE1MEXT OF AGRICULTURE. 



tlie tree from which ants were excluded and 10.4 per cent on the tree 

 to which they had access. 



On July 17 the banded tree still led in white-fly infestation, there 

 being 3.711 living young on this tree to 1.497 on the ant-patrolled 

 tree, and by the 13th of August white-fly eggs, larvae, and pupae were 

 too numerous on both plant- to count. This final heavy infestation of 

 both trees was expected, as it was improbable that complete white- 

 fly control in the height of its breeding period could be accomplished 

 by the ant. 



The relations of the ants to emerging white flies brought out in 

 the foregoing experiment led to similar observations on other trees 

 On June 12, at the height of a white-fly emergence period, two more 

 trees from which ants were excluded and two on which they were 

 present were inspected. On the first two trees there were 431 empty 

 pupa cases from which white flies had emerged, and 369 adult white 

 flies, or 85.6 per cent of all which had emerged, still remained upon 

 the trees. On the ant-invaded trees there were 600 empty pupa cases 

 with emergence slits, but only 36 of the white flies, or 6 per cent of 

 the emergence, remained upon the trees. 



These observations indicate that the principal direct effect of the 

 Argentine ant upon the citrus white fly in Louisiana is to destroy a 

 varying proportion of them, thus entitling this ant to be called a 

 white-fly enemy. 



RELATIONS WITH APHIDS. 



The relation of the Argentine ant to aphids has been observed 

 principally on the orange-infesting species, chief of which is Aphis 

 yossypii Glov. In Louisiana, however, certain observations have 

 been made upon the relations of this ant with aphids on loquat, 

 elder, privet, oak. cypress, and certain weeds. 



The Ant as a Peotector of Aphids. 



The orange aphid appears in considerable numbers, sometimes 

 very large numbers, on the newer growth early in the spring, often 

 increases throughout April and May, causing some of the leaves to 

 curl, and thereafter rapidly disappears, while a heavy parasitizatiou 

 is indicated by numerous dried skins punctured by the exit holes of 

 the parasites. This condition, which has long existed both in Cali- 

 fornia and in Louisiana, has not been altered materially even in 

 groves and trees overrun by the Argentine ant. In Louisiana it 

 occurs in scattered groups in January and February, often greatly 

 increases in March, and becomes numerous on tender leaves and some 

 of the blossoms in certain orchards during April and May. Even 

 where heavily attended by the ant, however, its natural enemies. 



