_ years. 
CEREAL EXPERIMENTS AT THE WILLISTON SUBSTATION. 7 
TEMPERATURE. 
The daily variations in temperature at Williston are recorded 
automatically by a thermozraph. These records are checked by the 
use of maximum and 
MAY JUNE JULY . AUG. “SEPT. 
minimum thermome- SPSS SsSe lege geggas 
™~ 
ters. The mean, max- Sawa ee ag 
a_i 
imum, and minimum 
temperatures, by 
months, during the 
erowing season for the 
years 1909 to 1914, 
inclusive, ‘are shown 
in. Table IV. 
Table IV shows that 
the highest monthly 
mean temperature is 
for July. This is the 
only month that a 
minimum lower than 
32° I’. has not occurred 
in some one of the six 
o 2ECO7 
|_ 9 Cae 
The frost-free period = og | | | || aEVCc EE 
is shown in graphic 
form in figure 3, cover- 
ing a period of 33 years 
S82) to, 19I4).0 2 it 
shows that the average 
date of the last killing pemerars</ Sipee siti Leer eas 
frost in the spring is ee ) 
May 18 and that the 
average date of the first AVERAGE 
frost in the autumn is ‘ 
September 14. The Fig. 3.—Diagram showing the length of the frost-free period and the 
shortest period be- average dates of the last killing frost in spring and the first in 
tween fros ts, as shown aa = ane period from 1882 to 1914, inclusive, at the 
by the graph, was 80 
days (in 1883) and the longest period was 155 days (in 1905). The 
average frost-free period for the 33 years is 119 days. 
Previous to 1893, the temperature records were kept at Fort Buford 
and since that date at Williston. 
