123 
1918-19.] The Cooling of the Soil at Night. 
(a) The relative humidity. 
( b ) The dryness of the surface layers. 
( [c ) The temperature of the underground layers. 
And the radiation balances the conduction, the latent heat, and the heat 
given up in cooling the surface layers; in other words, R = C + L + CSL, 
and if any three of these quantities are known, the fourth may be found. 
During the past winter I have made observations of the quantities 
R = Radiation, 
C = Conduction, 
L = Latent heat, 
CSL = Heat given up in cooling surface layers, 
ail in calories per square centimetre, on every available clear night, and the 
first part of this paper shows that the relation given above between these 
quantities does hold in practice as well as in theory. If the relation can 
be shown to hold good, then by observing R, C and L, or R and C only if 
the surface does not freeze, we can calculate the possible fall of the surface 
temperature and so find the minimum surface soil temperature. 
Since radiation is a surface phenomenon, we may reduce its effect by 
screening the surface, or by covering the soil with a layer of some poor 
conductor. The practical results of this are shown in Section VII. 
Evaporation requires wind and a dry atmosphere, with or without a 
clear sky. It is very local in its effect, for as soon as a dry layer is formed 
on the surface, evaporation practically ceases. 
Rain, sleet, or snow can only fall from a clouded sky ; in comparison to 
the time during which they act, these varied forms of precipitation are very 
effective in cooling the soil. A heavy fall of melting snow will reduce the 
temperature of the soil to 32° F. to a depth of several inches, and will do 
more in a few hours towards cooling the lower layers of the soil than several 
nights of rapid radiation. 
The practical results of screening the soil from evaporation and rain, 
sleet, and snow are given in Section VII. 
II. The Physical Constants of the Soil under Consideration, 
and the Notation employed. 
The soil in which the observations given in this paper have been made 
is a garden soil consisting of a layer of loam, rich in humus, of about 6 
in. depth, resting on a stony subsoil of quite different nature. The 
constants here given refer only to the surface layer of made soil, in which 
all the observations have been taken ; they are the mean results of many 
