8 BULLETIN 796, U. S. DEPARTMENT OF AGRICULTURE. 
It was found necessary in this work to differentiate between what 
were apparently normally active larvae and those which, although 
slightly active, were visibly affected by the gas. The larva? there- 
fore were classified into dead, stupefied, and active. Whether 
stupefied larva? eventually recover or die seemed to be largely a ques- 
tion of the duration and toxicity of the fumigation, and the degree to 
which the larva? were affected. Stupefied larva? when placed in the 
open air sometimes recovered and sometimes died, but the determina- 
tion of the exact condition under which either took place was impos- 
sible. It was determined, however, that larva? apparently much af- 
fected may be revived. On the other hand, when the dosage was 
sufficient to provide for a continuation of the fumigation much 
beyond the time at which the examination was made it may be as- 
sumed with safety that the stupefied larva? eventually would have 
been killed. 
In Table I it will be seen that a 1-ounce dose, applied July 19, 
showed 9 days after application a small number of larva? stupe- 
fied. By the time of the next examination, however, 14 days after 
application and in the two succeeding examinations, the larva? were 
all dead. 
Very little can be added to these data by an extended discussion. 
The principal facts in regard to dosage are brought out in Table I. 
In the study of the table, however, it should be emphasized that the 
relative numbers of dead and living larva? in the last two examina- 
tions for each dose are somewhat misleading, due to the earl} 7 de- 
composition of many of the dead, especially the smaller ones. The 
1-ounce and one-half-ounce doses were more quickly and fully ef- 
fective in the July applications than in the August applications. 
The one-fourth-ounce dose applied August 17 was partially effective 
but its action was not sustained, and by the time of the last examina- 
tion, 39 days after application, the dose had been dissipated so com- 
pletely that young larva? which were hatching at the time were able 
to enter the trees. The one-half -ounce and the 1-ounce doses both 
gave very complete immunity from this current infestation of newly 
hatched larvae. The importance of this point will be emphasized later 
in connection with the time of application and the final control ob- 
tained from the use of this gas. 
It was observed that the larva? of the fungus-gnat, Mycetobia 
sp., which feed in the gum, apparently succumb to the action of the 
gas in about the same proportion as do the larva? of exitiosa. 
From the results of the experiments with p-dichlorobenzene given 
in Table I it will be seen that at the soil temperatures prevailing 
in late July and August a dose of 1 ounce per tree has a very effective 
larvicidal action. The vapor apparently acts rather slowly, how- 
