172 
C. A. LUDWIG 
valerianic, and carbonic acids, creosote, ammonium chloride, ammonia, 
pyridine, picoline, lutidine, collidine, parvoline, coridine, and rubidine. 
Kissling (2) reported that the strongly poisonous materials (to man) 
are carbon monoxide, hydrogen sulphide, hydrocyanic acid, picoline 
bases and nicotine. To this list of compounds in tobacco smoke 
Thoms (6) added a phenol boiling at 190-200°, furfural (in small 
amounts), and a substance boiling at 200-260° containing sulphur 
and nitrogen and no terpenes. Crocker and Knight (i, p. 346) have 
called attention to the presence of ethylene and correlated its presence 
with the effect of smoke on some phanerogams. 
For exact data as to the cultures used and the methods of conduct- 
ing the experiments reported below the reader is referred to the writer's 
former paper (3) in which the influence of illuminating gas and its 
constituents is discussed. For the purpose of filling the culture cham- 
ber with smoke the tubulature of the bell jar was fitted with a two-hole 
rubber stopper carrying two glass tubes such that one extended very 
little below the stopper while the other extended well toward the 
bottom of the chamber. The short tube was then connected to an 
aspirator and the other to a cob pipe. The pipe was filled with tobacco, 
''Prince Albert" brand, the aspirator started and the tobacco lighted. 
When the chamber became filled with a white, opaque smoke cloud, 
the aspirator was stopped and the tubes plugged. The air in the 
chamber soon became clear, but the upper surface of everything within, 
and to some extent the vertical surfaces also, became stained brown. 
The material producing the brown stain did not extend into the test 
tubes, and consequently not to the agar, because it was clearly limited 
to the surface of the cotton plugs. No reactions, therefore, can be 
laid to these products except as they may have been somewhat volatile 
and therefore capable of diffusing into the tubes. In some of the tests 
the smoke was passed through one or two wash bottles containing 
water. In these cases small amounts of a brown oily substance con- 
densed and floated on the surface of the water, and it took longer to 
produce the opaque cloud in the chamber. 
The development of the bacteria in the smoke was rather variable, 
perhaps owing to an unavoidable lack of uniformity of the conditions 
in the different trials. There was a strong tendency for the colony 
development to begin at the bottom of the slant and progress upward 
after a preliminary period of no growth. No reason can be given for 
the very pronounced nature of this tendency in smoke. It was ob- 
