1922] ATWOOD— FORMALDEHYDE 249 
relay, which at the desired temperature turned off or on the heat 
under the bath. This arrangement permitted control of the tem- 
perature within o.2°C. Careful checking of the temperature at 
different points in the bath indicated that stirring devices were 
not necessary, other than the convection currents from the bottom 
of the bath upward. Two chambers were used for water soaked 
wheat (controls), two for the treated wheat, and two blanks to 
permit checking against leakage. 
After setting up the apparatus completely and before making 
a determination, each of the six complete trains was tested by 
suction as to its ability to hold up a column of mercury 10 inches 
high without small leaks permitting the column to settle back again. 
During the tests a gentle stream of air freed from carbon dioxide 
was drawn through each outfit for the entire period of hours of the 
run. In order that the rate of aeration might be uniform in the 
various outfits, and ample to provide for several complete changes 
of the air in the respiratory chambers during the course of the 
experiments, the suction secured from a water pump was conducted 
to the various chambers through tubing, connected to manometers 
in such a way that after careful calibration of the separate manom- 
eters, the rate of air flow could instantly be determined by a 
glance at the height of the paraffin oil surface in the manometers. 
Gas meters of this type were developed in connection with the 
chemical investigations incident to the recent gas warfare work, 
and are described in detail by BENTON (11). For each of the six 
trains air was drawn respectively through 50 per cent potassium 
hydroxide, a U-tube of moist soda lime, and through a barium 
hydroxide indicator to detect any failure of previous absorbents to 
remove all carbon dioxide. Air entered the respiratory chamber 
at the top and was removed from the bottom under the wheat 
arranged as described. The air then containing the carbon dioxide 
released by the wheat was drawn immediately through a bead 
tower containing fourth normal barium hydroxide, out and over 
another barium hydroxide indicator before passing to the tube 
connected to the water pump. 
The amount of suction was regulated by ground glass stopcocks 
between the pump and the last indicator flask. With the six com- 
