12 BULLETIN 796, IT. S. DEPARTMENT OF AGRICULTURE. 
cations of 1918 were made from two to three weeks later than was 
originally planned, and, while the results of these experiments were 
quite satisfactory so far as the larvicidal action was concerned, a con- 
siderable amount of the material remained about the trees at the time 
examinations were made. There is reason to believe that the dates 
originally adopted are more satisfactory. These dates are as follows : 
Michigan, Ohio, Connecticut September 1 
New Jersey, West Virgina, Maryland September 10 
North Carolina and the Ozarks September 25 
Georgia and Texas October 30 
In the study of Table VI it should be remembered that the exam- 
inations were made several weeks and in some cases several months 
after application. Dead larvae, of course, were largely decomposed 
after such a lapse of time. Hence no account was taken of anything 
but living larvse. Dead skins and partly decomposed bodies were 
always in evidence but no attempt was made to determine the actual 
infestation from their numbers. The check plat by which the relative 
infestation was judged in all cases was made up of consecutive trees 
in adjoining parallel rows. 
DOSAGE. 
Table VI sets forth with sufficient clearness the essential fea- 
tures of dosage efficiency. It will be seen that the doses applied as 
a single fall treatment gave results which agree very closely with 
their previously determined larvicidal value given in Table I. The 
trees treated in these experiments varied in age from 6 to 15 years 
and in size at the butts from one to several inches in diameter. 
It might be argued that the collar girth of the trees would be a bet- 
ter index to the dosage required than their age. There are certain 
facts, however, which make this undesirable. First, as will be shown 
later, injury is more dependent upon the age of the tree and the 
development of protective tissue than upon its size. Second, there 
has been a wide latitude in the size of trees treated successfully by 
a given dose. Under field conditions there is at all times a heavy 
surface and lateral loss of gas. It seems probable for that reason 
that there is a certain area of gas diffusion necessary before the 
vapor in the center of this area attains toxic concentration. There 
are, of course, limits in both directions in the application of this 
statement, very large trees undoubtedly requiring larger doses than 
small ones. In practice, however, the latitude in collar girth of 
trees which a given dose will treat successfully has been enough to 
cover the usual variation in trees between the ages of 6 and 15 years. 
It will be seen in Table VI that the one-fourth-ounce dose gave 
only partial control. Although only one experiment included the 
application of this amount, the results agreed fairly well with its 
