121 
1918-19.] The Cooling of the Soil at Night. 
its flowers and roots in temperatures differing by 18° F. The air, as it were, 
rests on a cushion of non-conducting grass, and fails to make intimate 
contact with the soil, and so gains little or no heat from it. 
No such large differences have existed during the past winter (1918-19) 
between the temperature of the air resting in close contact with open soil 
and the surface soil itself — in fact, the maximum difference only reached 
5° F. on two occasions, and the average over many clear nights was only 
2*4° F. Dr Aitken mentions a grass minimum 7° F. lower than the air 
minimum over open soil, and the following figures for nights during the 
past winter are from my own observations : — 
Table I. 
Date. 
Surface soil 
minimum. 
Air over 
open soil 
minimum. 
Air over 
ashes 
minimum. 
Grass 
minimum. 
October 1, 1918 . 
November 13, 1918 
February 10, 1919 
38-0° F. 
31*0° 
22-0° 
38-0° F. 
28-5° 
19-0° 
32-0° F. 
27*0° 
17-5° 
28-5° F. 
240° 
15-0° 
Thus the average difference was only 1-8° F. between the minima of 
soil and air over it, while the average difference between the soil minima 
and grass minima was as much as 7 '8° F. 
It would appear, therefore, that the grass minimum gives us a totally 
wrong impression of the fall in temperature of the air over open soil. In 
large fields, or even in gardens where lawns or ash or gravel paths do not 
cut up the cultivated plots, the air minimum will not differ greatly from 
the surface soil minimum, and the determination of the soil minimum will 
go a long way towards solving the question of whether there will be a 
frost or not. 
The three main causes of the cooling of the soil are radiation, evapora- 
tion, and the fall of cold rain, sleet, or snow. 
Radiation to be effective presupposes a clear sky and a dry atmosphere ; 
we should therefore expect there to be a relation between the rate of 
radiation and the relative humidity. The existence of such a relation is 
discussed in Section III (a). 
But on equally clear nights of the same average relative humidity, and 
even the same sunset temperature, we find very different surface soil 
minima — these being notably low when the soil surface is dry, and lower 
in a spring month when the underground layers have not warmed up after 
winter, than in an autumn month of the same length of night when the 
