THE LESSER GBAIN-BORER. 39 
to the first lot, giving a complete range of mill products, all badly 
infested with Rhizopertha dominica except the clean rice and bran. 
Fully 95 per cent of the Rhizoperthas were dead in all the bags 
except those having two or more between them and the outside of the 
piles. In the well-protected sacks about 25 per cent were living, 
which showed that the gas was beginning to penetrate deeply. 
Tribolium and Calandra resisted the gas much better than Rhizo- 
pertha, and a longer exposure and probably increased dosage would 
be necessary to destroy all the insects in the center. 
A space between two partitions near the center of the system of 
storage was boarded up on four sides and ceiled with planed boards 
without grooves or lap, but solid and suitable as a foundation for 
the three-ply heavy roofing tarred paper which formed the lining of 
the room. This was carefully put on with a good lap at each seam, 
all of which were fastened with a good application of roofing cement 
before being lapped and nailed, so that a very tight room was the 
result. The floor was of tongue-and-groove lumber and appeared 
very tight. Special care was taken with the door and frame, which 
were beveled and faced with strips from an old rubber belt. The 
door was held shut after charging by a heavy bar sliding into iron 
hooks. The door itself was of two thicknesses of good boards, with 
an uncut sheet of the roofing felt between. That the leakage was 
A^ery slight was proved by the faint odor of gas, noticeable only after 
fumigation had been in progress some time, and this came out almost 
entirely through the floor. 
The room was the most nearly gas-proof structure possible under 
the circumstances and better than many nursery fumigating houses 
which have come under the writer's observation. There was no 
leakage through the walls while the experiments were being carried 
on, and the loss of gas through the floor and around the door was 
slight. An objection, consisting in the absence of ventilators other 
than the door, was obviated by leaving the door open all night at the 
end of the experiments, as the material could not be immediately 
moved out. The room was not rectangular as to shape of the ground 
floor, but the sides averaged 14.25 feet, and the height of ceiling jvas 
13.50 feet, making the cubic contents 2,741 cubic feet. 
For generators two stoneware jars or "churns" holding 3 gallons 
each were used. 
Several wooden trestles, 2 feet high, with planks laid across them, 
were placed in the room on which to pile the sacks, so that the gas 
could have access to them from beneath as well as from all sides. 
As the room was in a dark part of the sheds, it was necessary to 
have an electric light on an extension hung in the room in order that 
the sacks could be properly arranged. 
