28 BULLETIN 647, U. S. DEPARTMENT OF AGRICULTURE. 
and 18.7 per cent of those, on ant -excluded trees, showing a slight 
tendency toward worse infestation under ant attendance. The 
marked tendency of meatybugs to establish themselves under the 
sepals of the young fruits and in similar situations to secure sheltered 
feeding places must be taken into account when considering the 
subject of the transfer of mealybugs by ants to establish new colonies. 
Between May 24 and July 24 mealybug infestation rapidly in- 
creased on the trees to which ants had access, while it decreased, 
with slight fluctuations, to an almost insignificant amount on those 
from which ants were barred. From July 24 to September 2 there 
was a slow reduction in the amount of mealybug infestation on the 
ant-traversed trees, an increasing number of mealybugs' remains indi- 
cating increased effectiveness of the natural enemies, which had 
become more numerous following the food supply. On the trees 
from which the ants were barred the mealybug infestation in the 
same period, with minor fluctuations in which the highest point 
was slight infestation of 7 per cent of the fruit, was maintained at a 
negligible amount. 
The most important early activity of mealybug predators occurred 
on the very small fruits, these insects occurring with mealybugs 
under the sepals as soon as the mealybugs arrived there and pre- 
venting the growth of infestations from these spots. From July 24 
to the close of the experiment, September 12, the number of preda- 
tory enemies, again following the available supply of food insects, 
was greater on the trees traversed by the ants than on those from 
which ants were barred, there being from five to eight times as many 
on the former as on the latter trees at the times examined. The prin- 
cipal enemies of the mealybugs occurring on these trees were Coc- 
cinellidae, Hemerobiidae, Chrysopidae, Pyralidae, and Syrphidae. 
During this latter period of the experiment, following the decrease 
in percentage of infested fruits on the ant-traversed trees, the in- 
creasing effectiveness of the mealybug enemies was manifested in the 
occurrence of an increasing number of fruits which had been rid 
of mealybugs, their previous infestation being indicated by bits ot 
cottony secretion, sooty mold, etc. 
Experiments III and IV. 
Two other experiments conducted at Alhambra, Cal., with nursery 
trees and potted seedling orange trees brought out very similar 
results. The nursery trees, owing to too late transplanting, failed 
to thrive and did not become very heavily infested with mealybugs, 
but showed less plainly but quite as certainly the results of ant 
attendance in increasing these insects. 
In the experiment on potted orange seedlings, 6 of the young 
plants were infested artificially with mealybugs, and on May 17, 
