THE ARGENTINE ANT IN RELATION TO CITRUS GROVES. 41 
excretion in this secondary manner, especially when other food is 
scarce in the trees. If any large proportion of the white-fly excre- 
tion were taken b}^ the ants, however, the sooty-mold fungus would 
be by so much the less prevalent in ant-invaded trees ; this, however, 
is not the case. The ants do prevent to a very large extent the col- 
lection of excretory matter and the formation of sooty mold after 
mealybugs, even inducing such rapid excretion in certain young 
stages that the mealybug is unable to form the wooly covering, its 
body remaining almost naked and pink. 
Effect of the Ant on Abundance of the White Fly. 
An experiment to determine the effect of the ants on abundance of 
white flies was started on April 25, 1914, a young orange tree with 
S38 white-fly eggs being banded to exclude ants and a similar tree 
with 1,474 eggs kept accessible to ants for comparison. The per- 
centages of young stages of the white fly dying from unknown 
causes and the quantity of new growth on the two trees were noted 
at every examination to make sure that the difference in white-fly 
infestation was not due to varying food conditions. 
On May 13 there were on the tree from which ants were excluded 
949 sound and 113 dead larvae and pupa? and 189 unhatched eggs; 
on the tree that was accessible to ants there were 434 sound and 109 
dead larvae and pupae and 112 living eggs. Between May 13 and 
June 12 the nonbanded tree was merely kept under surveillance by 
scouting ants, but on June 12 white-fly emergence was at its height 
and the ants had formed a heavy trail into the tree, where they were 
capturing the emerging adults. Living white-fly larvae and pupae 
were comparatively scarce and about equal in number on both trees. 
The remarkable thing was, however, that on the ant-invaded tree 
there were 167 empty pupa cases from which white flies had emerged 
and only 8 of the adult white flies, whereas on that from which ants 
were excluded there were 151 empty pupa cases and 130 of the 
emerged adults. In other words, the nonbanded tree was swarming 
with ants, some of which were carrying adults, and only 4.51 per 
cent of the emerged adults remained on the tree, whereas on the tree 
from which ants were excluded almost the same amount of emergence 
had occurred and 86 per cent of all the emerged white flies were still 
on the leaves. 
From June 12 to about the middle of August the white flies in- 
t creased faster on the tree from which the ants were excluded than 
on the other. On July 1 there were 5,435 living young on the 
former and only 1,919 on the latter. The percentage of dead was 
practically the same on both, being 12.9 per cent of all young on 
