216 BOTANICAL GAZETTE [MARCH 
Material and method 
Three lots of seeds were used: one gathered in late summer 
or early fall, 1915, at Pullman, Washington; and the other two 
at Gary, Indiana, one lot in the late summer or early fall of 
1914, and the other a year later. The experiments extended from 
March 15 to August 1, 1916. Thus after-ripening had proceeded 
for several months in two lots, and for one year and several months 
in the third lot. In some of the experiments, remaining coat 
effects were further eliminated by grinding some of the seeds with 
sand for four, six, or nine weeks, or treating them with concentrated 
H,SO, for two or three minutes and then washing them with run- 
ning tap water for two minutes. The optimum length of time for 
treatment with H,SO, was determined by experiment to be two 
minutes for Indiana seeds, and three minutes for Washington 
seeds. Along with these treated seeds were run seeds with coats 
not treated. These seeds, treated and untreated, were then put 
in lots of one hundred in Petri dishes lined with moist absorbent 
cotton. The Petri dishes were placed in a cardboard box lined 
with opaque black paper, and the box then put in a refrigerator for 
forty-eight hours, where the temperature was below the minimum 
necessary for germination, to allow time for the seeds to soak 
before they were tested for temperature effects on rate of germina- 
tion. At the end of this time the box containing the Petri dishes 
was kept at the desired constant temperature. The temperatures 
tested ranged from 9° to 42°C. Temperatures below room temper- 
ature were obtained in baths cooled with running water, and at 
room temperature and above, in a constant temperature incubator 
in which the variation was less than one degree. At one time only 
did the temperature vary more than this, and then it was in the 
water-cooled bath; this variation is indicated on the curves 1, 2, 
and 3, and in all the tables thus, (8°—10° C.). 
The percentage germination at various intervals of time for 
each temperature tried was noted for all seeds in the box. The 
results were plotted as curves; one set of curves for each type of 
seeds, that is, one set of curves for Indiana seeds collected in 1915 
and treated with H,SO, for two minutes, another set for Indiana 
seeds collected in 1915 and ground with sand nine weeks, another 
