20 
Proceedings of the Royal Society of Edinburgh. [Sess. 
Table YIII — continued. 
Date. 
(1) 
M 0 . 
(2) 
m 4 . 
(3) 
R 0 - a ' 
(4) 
Estimated 
“c* Relative 
w Humidity. 
6 p.m. 
Forecast of 
3 Surface 
Minimum. 
Observed 
Yi Surface 
Minimum. 
Difference of 
"oo Columns 
" (6) and (7). 
Weather. 
(9) 
1919. 
April 8/9 
18-0 
10-8 
'44 
per cent. 
90 
+ -3° 
+ 2-0° 
-17° 
Clear, windy 
„ 12/13 
16-6 
9-6 
'42 
90 
Frost 
+ 1-0 
55 
55 
June 8/9 
26*7 
19-2 
•42 
70 
+ 6-7° 
+ 7-8 
-IT 
55 
55 
» 9/10 
31*1 
20-0 
*45 
85 
+ 7-0 
+ 7*9 
- -9 
55 
55 
„ 13/14 
26-7 
17-8 
•42 
80 
+ 5-5 
+ 6-0 
- -5 
55 
55 
„ 19/20 
23-3 
17-5 
•36 
75 
+ 5-6 
+ 6-6 
-1-0 
55 
55 
„ 21/22 
34-0 
21-4 
•40 
75 
+ 7-0 
+ 8-6 
-1*6 
55 
55 
„ 28/29 
30-6 
20T 
•34 
75 
+ 6*3 
+ 7-0 
— *7 
55 
55 
May 1/2 
12-8 
93 
•38 
90 
Frost 
+ 0*4 
Clear most of 
„ 2/3 
14-4 
9-4 
•42 
90 
55 
+ 0*5 
night, no wind 
55 
June 6/7 
26-6 
19-6 
•40 
85 
+ 8-5° 
+ 9-5 
-.1*0 
55 
„ 30/1 
133 
12-4 
•40 
90 
+ 4-2 
+ 5-0 
- *8 
55 
Average difference (9 occasions), — 1 -0°. 
of frost ; a very low forecasted minimum would certainly imply a con- 
siderable intensity of frost, but it would not be safe to assume that the 
surface would fall below 0° C. if the forecasted minimum was only a few 
degrees below that temperature. Thus in Table VIII, where the forecasted 
and observed minima are contrasted, I have not given the actual figures 
of the forecasted minimum when it was below 0° C., but have simply 
stated that frost was expected. The comparison is for calm clear nights 
between April 9 and October 16, 1919; before and after these dates the 
4-in. depth temperatures were so low as to assure the occurrence of frost 
every calm clear night. 
YIII. Conclusions. 
It is customary in most text-books and articles dealing with the cooling 
by radiation at night to ignore the temperature and upward conduction 
of the soil entirely, or at least to state that it is of little account. 
This may have some justification when dealing with the cooling over 
a large area probably covered with vegetation, but it seems to have no 
justification when we are considering the minimum temperature at night 
on plots of cultivated land on which the vegetation covers only a small 
fraction of the whole. 
It is interesting to note that Professor H. J. Cox, in his exhaustive 
study of temperature conditions in the cranberry marshes of Wisconsin, 
