Feb. 3 ,1923 Use of Alternating Temperatures in Seed Germination 303 
tucky bluegrass, the favorable effect of the alternations is not referable 
to the effect of extreme or mean temperatures but is the result of the 
changes in temperature. In several other tests of celery seed with better 
control of the lower temperatures, germination was much poorer at 
20 0 and with the alternation 20° to 25°, in comparison with other alter¬ 
nations, than in the tests represented in Figure 4. No other tests were 
made at 15 0 or with the alternation 15 0 to 20°. 
The slight fluctuations in the temperature of the cool germinating 
chambers were no doubt partly responsible for the high percentages of 
germination at low temperatures; but they certainly were not sufficient 
to explain the results entirely if the seeds required as wide alternations 
at low temperatures as at high temperatures and if no other significant 
factor was involved. Additional tests made at another time indicate, 
however, that other factors may be involved. Ninety-six simultaneous 
tests of celery seed from a single unusually sensitive lot were made in 
different parts of a single chamber, which was heated from below by a 
gas burner during the several hours of each forenoon and cooled each 
afternoon by a stream of cold water in the top of the water jacket around 
the chamber. The upper and lower temperatures were controlled by 
thermo regulators, and the lower temperature was held for about 15 
hours of each day. The seeds in the more rapidly cooled parts of the 
chamber germinated much more rapidly and completely than those in 
other parts of the chamber, though the extreme upper and lower tem¬ 
peratures reached were very nearly the same in all parts of the chamber. 
The actual range in percentage of germination in different parts of the 
chamber was from o to 32 per cent in 7 days, from o to 58 per cent in 
9 days, from 11 to 81 per cent in 11 days, from 41 to 91 per cent in 14 
days, and from 53 to 95 per cent in 21 days. After 21 days the seeds 
which had occupied the least favorable positions and had germinated 
most poorly were put in the most favorable positions, with the result 
that their germination soon increased to equal that of the seeds originally 
in the more favorable positions. 
Similarly, when the celery seed was daily transferred between two 
chambers constantly maintained at different temperatures and the cool 
chamber was cooled by a block of ice above the water jacket, as was done 
in the tests represented in Figure 4, the seeds germinated somewhat 
better if placed in the top of the cool chamber than if placed lower in the 
chamber where the temperature changes were less abrupt. 
In the tests discussed in the two preceding paragraphs the tempera¬ 
tures in the positions in which the seeds germinated best were almost 
always somewhat lower than in the other parts of the chambers. The 
differences in temperature were not, however, commensurate with the 
corresponding differences in germination. In the more rapidly cooled 
and constantly somewhat cooler parts of the cool chambers convection 
currents of air must have been more lively than in other parts, and this 
may have led to a significantly better renewal of oxygen in immediate 
contact with the seeds, as suggested by Vanha (21). Also there was a 
constant tendency in the cooler parts of the chambers for water from the 
air which had become saturated in the warmer parts of the chambers to 
condense upon the surface of the seeds. This condensation water would 
presumably be saturated with oxygen, which would be immediately 
available for the use of the seeds unless kept out by seed coats relatively 
impermeable to oxygen. While there is no direct evidence that oxygen 
relations really are important here this possibility must be admitted. 
