460 BOTANICAL GAZETTE [DECEMBER 
absorbent cotton on which the seeds lay. This method was dis- 
carded for the temperature and pressure experiments, as a more 
satisfactory arrangement consisted in immersing the lower end of 
the capillary tube in a flask of water, thus drawing the atmosphere 
through water before it entered the germinator. This method had 
the added advantage of showing at once whether the flow of air 
through the capillary tubes was being interfered with by dust 
particles. The air was kept moist in this way, but owing to the 
very rapid exchange of gas in the apparatus, the atmosphere was 
not saturated. The water which ran the aspirators had a fall of 
45 feet, and was furnished through a separate constant level system 
which gave the aspirators uniform power. 
The seeds were in all cases prepared for experimentation by 
soaking them in ice water, far below the minimum temperature 
for germination, for at least 12 hours, after which the testa was 
removed carefully, without injury to the seeds. ‘They were exposed 
to constant conditions for 10 days, and the elongation of the hypo- 
cotyl and the geotropic response was used. as the criterion of 
germination. 
Temperature series at various pressures were run at 31 C., 
and the influence of fluctuating temperature was determined by 
using a fluctuation of 25—40° C 
At the low pressures and high temperatures employed, the 
evaporation of water in the apparatus is very rapid. In determin- 
ing the actual oxygen pressures to which the seeds are subjected, 
it is necessary to know what volume of air is drawn through the 
apparatus in a given time, and what part of the gas pressure is 
due to water vapor. The atmosphere is not saturated, for if it 
were saturated at these reduced pressures, the water vapor pressure 
alone would be much higher than the total pressure used. The 
amount of air at normal pressure entering the apparatus is 3-5 liters 
per hour, a very rapid exchange, since it means 8-12 times that 
volume per hour within the germinators due to expansion under 
diminished pressure. Under such circumstances it is not surprising 
to find that the water vapor is removed so rapidly that the satura- 
tion point is not approached. It was clearly demonstrated, how- 
ever, that water was not a limiting factor in these experiments. 
