77 
Comparison of Tables 8, 9, and 12 shows that the time required for 
germicidal action is materially increased by a reduction in tempera- 
ture, even though it is not below 60° F. 
This is more noticeable with B. subtilis than with B. coli communis. 
At a lower temperature the killing power of the formaldehyde be- 
comes so limited that only a few experiments were performed at a tem- 
perature lower than 60° F. With a temperature of 40° F. and humid- 
ity of 95 per cent, B. coli communis was killed within an hour and a 
half; but, considering the method of exposing, the organism, and the 
fact that as high humidity as that is rarely observed at that tempera- 
ture except under artificial conditions the results can not be consid- 
ered satisfactory. 
DISINFECTION OF SLEEPING CARS WITH FORMALDEHYDE 
GAS. 
These experiments were conducted for the purpose of determining 
the practical value and efficiency of the formalin-permanganate method 
of liberating formaldehyde gas in the disinfection of sleeping cars. 
For comparison, experiments were also done with other well-known 
methods of evolving the gas for disinfection purposes. The work was 
done in the cars of the Pullman Company at their local yards. We are 
indebted to the officials of this conipan}^ for the use of the cars and 
their cooperation and assistance in carrying out this work. The cars 
were selected at random, regardless of when they had been occupied; 
usually they had been used the previous night. Often the cars were 
shifted about in the yard during the experiment. Although the size 
of the different cars varied slightly, for our purposes their air space was 
approximately 4,500 cubic feet. 
Usually 2,000 c. c. of formalin, containing 35.66 per cent by vol- 
ume of formaldehyde, were used for an experiment. Before starting 
all berths were let down, inside doors opened, and windows and venti- 
lators closed. When practicable, the lamp ventilators were closed on 
the inside with cotton waste ; otherwise they were closed on top of the 
car by wrapping cloths around them. No pasting of cracks was 
done. 
All organisms used, except B. tuberculosis , were grown, prepared, and 
exposed by the Wilson method, as described on page 17. For B. 
tuberculosis , tuberculous sputum, showing on examination many 
tubercule bacilli, was used. Exposures of this were made by the 
Wilson method and on small pieces of carpet, using in each case a drop 
or two of the sputum. For convenience and safety, the pieces of car- 
pet were carried in a Petri dish and exposed in the car by simply re- 
moving the top of the dish. After preparing the exposures they were 
dried in the incubator at a temperature of 37° C. for 1 hour, and, as the 
laboratory is some dis tanc e from the car yards, about another hour 
