12 BULLETIN 681, U. S. DEPARTMENT OF AGRICULTURE. 
just within the space previously occupied by the hopper. A very 
thick cloud of dust was blown into the room. Immediately an 
explosion, or rather a large flame, occurred, followed by a second and 
more extensive one, covering nearly all the floor area of the dusty 
side of the building and extending almost to the roof. The door and 
windows were open. Otherwise, some damage might have resulted. 
As it was, the arms of the man who handled the torch were burned. 
The results of these preliminary experiments (observed by J. K. 
Clement, B. TV. Dedrick, M. P. Helman, J. Weaver, and F. Kline), 
may be summarized as follows: 
1. More than 40 attempts with a total of from 3 to 5 pounds of 
foreign materials, such as nails, small pieces of stone and flint, and 
matches, fed along with various kinds of explosive grain products, 
failed to produce a single explosion. 
2. Two attempts with a naked flame were successful in producing 
explosions. 
EXPERIMENTS FROM APRIL 6, 1916, TO OCTOBER 3, 1917. 
SERIES. 1. 
Object.— The object of the test was to determine whether explosions 
could be obtained by foreign materials capable of emitting sparks 
when ground, or by an electric arc located near the disks of the 
attrition mill. 
Conditions. — The two removable iron doors of the mill were replaced 
by wooden doors, provided with peepholes covered with several 
layers of mica, and drilled for the insertion of iron carbon holders for 
the electric arc. The position of the arc could be shifted between the 
shaft and periphery of the disks, but it was soon found that its location 
did not seem to have any effect on the regularity of the explosions. 
As a rule, in this and succeeding tests, the arc was located nearer the 
shaft than the periphery. A rheostat and ammeter were provided for 
regulating the current (usually 15 to 20 amperes) through the arc. 
The hopper underneath the grinding disks was put in place, and its 
hand hole was left open so that the flash due to an explosion could be 
seen. 
Resides. — 
1. Sparks from foreign materials did not cause an explosion. 
2. The electric arc caused four explosions out of six attempts when 
large Iiandfuls of feed were used. 
3. Low rates of feeding did not produce an explosion with the 
electric arc. 
4. Oat 1 Lulls, ground or unground, did not ignite, except when 
mixed with the higldy inflammable elevator dust. 
•3. With elevator dust, when fed by the handful, either alone or 
with oat hulls, four explosions out of four attempts were obtained. 
