250 BOTANICAL GAZETTE [APRIL 
openings were made of the same cross-section as the chamber. 
The distal end of each cork was fitted tightly into the opening of 
the bottles (4 and B). On the upper side of the plaster case was 
hollowed out a groove (d@), and perforations were made through 
the floor of this groove in such a way that the roots of the seed- 
lings (e) projecting through these holes would be suspended in 
a chamber of gas when 4 and B were filled with gas. The groove 
made a good support for the cotyledons and for the cotton which 
was wrapped around them to keep them moist. The plaster was 
entirely covered with a thin layer of paraffin. In this way an 
air-tight passage was made from one bottle to another. To 
insure a moist condition of the gases which filled chambers A and 
&, several sheets of damp filter paper were placed in the bottles, 
and the cotton around the seedlings was moistened frequently. 
Besides this, the connecting chamber was soaked in water to 
infiltrate the plaster before inserting the seedlings. 
In such a contrivance, judging from a preliminary experiment, 
it was thought that the gases diffused too rapidly through the 
chamber to be effective. However, by testing with a mixture of 
illuminating gas and air in one bottle and air in the other, and 
observing the contortions in the roots produced by the illumi- 
nating gas, it seemed evident that the gases must diffuse slowly 
enough to allow a considerable difference in the gases on the 
opposite sides of the roots. 
Seedlings of Zea Mays, Lupinus albus, Vicia sativa, and Pisum 
sativum were now supported in the chamber, and bottle A, filled 
with 100 per cent. CO, or with mixtures of air and CO, in vary- 
ing percentages, and bottle B filled with air. It was found that 
when a mixture of 20 per cent. CO, and 80 per cent. air on one 
side was opposed to air on the other, of 18 seedlings used only 
I was positive toward the air, the remaining being indefinite; 
while when 100 per cent. CO, was opposed to air, of 18 seed- 
lings only 2 were positive and 16 indefinite. In the former case 
the roots grew 5 to 10™™ during the experiment, while in the 
latter case only 0.5 to 1.5™™. 
Although probably there was not a very great difference in 
the composition of the gases on the opposite sides of the roots, 
