102 
The Ohio Naturalist. 
[Vol. VII, No. 5, 
As it was necessary to draw samples of air from different 
parts of the box at intervals during the tests, six holes were 
made, one at one end of the box, two on the side of the box near the 
top and bottom, and three in the end opposite the jar; the first 
near the middle a and top of the box, the second half way from the 
top to the bottom of this end but at one side of the center, and 
the third near the lower corner. Sections of gas pipe about three 
inches long were threaded and screwed into the holes, so that 
about two inches extended outside the box. 
The apparatus for drawing the samples to be tested was 
made by using sections of one-inch glass tubing eighteen inches 
long. To one end was attached a rubber tube, and a pinch cock 
was used as a cut off. Another rubber tube provided with a 
pinch cock was attached to the pipe. The glass tube was then 
filled with distilled water and attached to the rubber tube 
leading from the hole, and after opening the upper cock the 
water was gradually allowed to dribble from the lower rubber 
tube by partially releasing the other cock. After one foot of 
water had run from the glass tube, both valves were closed. By 
uncoupling the glass tube from the upper rubber tube and 
placing the thumb over the end, the gas caught was easily mixed 
with the water by shaking. The water solution was then placed 
in small bottles, which were tightly corked, and after a set of 
experiments was made samples were taken to the laboratory, 
placed in test tubes and treated with a solution of silver nitrate. 
The presence of hydro-cyanic acid gas was indicated by a 
white precipitate. In case chlorides were present a similar 
preciptiate might be secured, hence the distilled water was 
repeatedly tested before using in order to assure accuracy of 
results. • 
No attempt at quantitative analysis was possible, in the 
limited time that could be devoted to the tests, but the precip¬ 
itates secured were rated as “trace,” “very slight,” “slight,” 
“good,” “very good,” and “heavy,” in order to form a basis for 
determining results. 
The writer is aware that the method followed can be improved 
and trusts that someone may have sufficient time to devote to 
this important work, in order to secure a more extended knowl¬ 
edge of the behavior of this gas in fumigating under various 
conditions. 
Bv experiment it was found that the results secured by using 
a glass tube eighteen inches long were not reliable, as only a 
small amount of air from the box was obtained. Two tubes 
were then coupled together so that about two feet of water could 
be drawn out. This resulted in a fair sample being drawn for 
each test. 
