7 o8 
Journal of Agricultural Research 
Vol. XXIII, No. 9 
The applications were made according to schedule for a period of six 
weeks, except for slight variations on account of unfavorable weather. 
The sprays were applied with a knapsack sprayer; the dusts, with a 
hand-operated dusting machine of the fan blower type, suspended by 
straps from the shoulder. The applications were very thoroughly made, 
from both sides, so it is 
very certain that the 
spikes were well cov¬ 
ered with the fungi¬ 
cides. On Saturday of 
each week, before the 
regular treatment was 
applied, a count was 
made of all diseased 
and disease-free spikes. 
The count included im¬ 
mature spikes that 
were bursting through 
the sheath, but none 
younger than that. In 
this way an accurate 
comparison was made 
of the effectiveness of 
the different treat¬ 
ments. At the same 
time all the diseased 
spikes were cut and 
allowed to fall to the 
ground, thus providing 
rows of mo Id-free 
plants for the follow¬ 
ing week. 
In connection with 
the control experiment 
a test was conducted 
according to the 
method described by 
Winston and Fulton 
(24) to determine the 
persistence on the 
plant of the copper 
preparations used. 
Briefly, the particular 
methods used were as 
follows: At the time the 
counts were being 
made each Saturday 
liberal samples of 
spikes from the Bor¬ 
deaux dust and Bordeaux spray-treated plots were collected at 
random in paper bags. The samples included about an equal number 
of diseased and disease-free spikes. Out of each lot 200 gm. were 
weighed out and washed thoroughly in 500 cc. distilled water into 
which 1 cc. of nitric acid had been poured. Any copper sulphate 
Fig. 4. —Results of spraying experiment from August 2 to September 
13» 1919. The graphs show the percentage of mold developing on 
castor-bean plants, week by week, for each of the fungicides used; 
and, for the purpose of comparison, the weekly rainfall over the same 
period. 
