New York AGRICULTURAL EXPERIMENT STATION. 129 
This was found to show on analysis,— 
PCa he ses Sum tk whe See ee CN 24.91 per ct. 
PMMA SOG eset ce nets che as 61.57 per ct. 
RE Ziad NEY ES Le ci ese ae ae 0.74 per ct. 
A 20-pound lot made up at another time contained 0.62 per 
ct. of free alkali. . 
No, 2. Another mixture was made containing twice as much 
caustic soda and less fish-oil. This contained,— 
es eet oak Meiccy s wteas Asiajeas sToimee ere 23.36 per ct. 
POSIT MMSE Pear che 00 «arate «seats Loa Sears. AC.A9- per-et. 
CC met eA: Soh re he elas Meese ee hea 5 ¥ 11.22 per ct. 
EXPERIMENTS IN USING HOME-MADE SOAPS IN SPRAYING. 
Experiment 1.—Action on aphis of willow.—Soap No. 1 was 
used in solutions of three different strengths (one pound of soap 
in two, five, and seven gallons of water) upon willow leaves 
infected with the willow aphis (Lachnus salicicola Uhler). The 
insects were completely destroyed by each of the three solutions. 
The success of this home-made soap in solution of at least one 
pound of soap in seven gallons of water was thus shown in 
destroying aphis. 
Another point, however, demanded attention, the effect of free 
alkali in a soap solution upon the foliage itself. To test this a 
series of additional experiments was carried on. 
Experiments testing action of home-made soap containing differ- 
ent amounts of free alkali upon foliage.—In the following set of 
experiments, the solutions used contained one pound of soap in 
seven gallons of water. Soap No. 2, containing 11.22 per ct. of 
free alkali, was used without any addition of alkali. Soap No. L 
was used in its normal condition, containing 0.75 per ct. of free 
alkali, and also with added quantities of caustic soda, making 
the amount of free alkali in the soap 1, 2, 5, 10, 20 and 50 per 
ct. of the soap. 
(1) June 22, apple, pear, plum and currant leaves were sprayed 
with solutions of soap containing 0.75, 1, 2, 5 and 10 per ct. 
of free ‘alkali. A fine rain was falling while the trees were being 
| sprayed and the results of the experiment must be interpreted 
9 
