466 BOTANICAL GAZETTE [DECEMBER 
pressure of 5.5 mm., there was still 66.6 per cent of germination. 
The closest estimation possible from the data at hand would 
indicate an oxygen minimum of about 2.5-3.5 mm. for the low- 
ers at 31°. 
The lowering of the minimum is somewhat greater in the lower 
than in the upper seeds. The oxygen minimum of the uppers is 
decreased from about 12 mm. to less than 7 mm., while for the 
lowers the minimum is decreased from 9.5 to about 3 mm. All 
of these results show that with the increase of temperature there is a 
decrease in the demand for free oxygen. The probable reason for 
this will be discussed later. The physiological difference of the 
embryos of the upper and lower seeds is shown clearly by these 
experiments, the embryo characters being just as strikingly differ- 
ent as the coat characters; and both sets of characters act together 
in securing the difference in delay of the two seeds. 
Temperatures fluctuating between 25 and 40° are apparently 
no more effective in producing germination than the constant high 
temperatures employed. Since the pressure in the germinators 
remains constant while the vapor pressure fluctuates with the 
temperature, it is evident that the oxygen pressure fluctuates also, 
and that its fluctuation is inversely as the temperature, rising as the 
temperature falls, falling as the temperature rises. CROCKER 
found such fluctuating temperatures more effective in producing 
germination than constant temperature of 35° in the upper seeds 
with testas intact. The fluctuation may render the testa more 
permeable to oxygen, but in view of the effect of temperature on 
oxygen demand, the inference cannot be made with certainty. 
A peculiar result was observed in all the control experiments at 
high temperatures. At normal room temperature, both the 
experimental seeds and the controls show less growth in the uppers 
than in the lowers; but at high temperatures, whether constant 
or fluctuating, this relation is reversed in the controls. This is 
noticed on comparison of lower and upper controls in table II. At 
the same time, both lowers and uppers of the experimental seeds 
show less growth in a fluctuating temperature of 25-40° than cor- 
responding lowers and uppers at constant high temperature of ae 
and equal pressures. The first two experiments in table II, com- 
