Surron—Influence of Water- Vapour upon Nocturnal Radiation. 31 
differences between the initial temperatures and the temperatures 
of the dew-point ; and, moreover, these differences are almost as 
great for the lower initial temperatures as for the higher. For 
relative humidities of eighty per cent. and over, the fall of tempera- 
ture is pretty nearly the same whatever the initial temperature : the 
reason being, probably, that we are now near the beginning of 
condensation. 
Our observations, then, show a relationship between the 
nocturnal cooling of the air and the relative humidity, but not any 
relationship at all to the absolute humidity. But, as we have said, 
it is a dry air which, of itself, d priori because of its feeble radia- 
tive power, should cool more slowly than a dampair. We have 
then to establish a connecting link between these two facts. It is 
to be found, it seems, in the dependence of the cooling of the 
ground upon the hygrometric state of the air. During 1899 and 
1900 I took a series of observations of the temperatures just 
under the surface of a bare patch of red sand, comparing the falls 
of temperature observed on clear nights in the three hours 
8-11 p.m. with the absolute and relative humidities. The sand was 
originally taken from a depth of three or four feet, and contained 
little or no humus. It was not always in exactly the same condition, 
being of course damper during the clear nights immediately after 
rain. I am inclined to think, nevertheless, that this source of 
uncertainty was insignificant in its effects, and quite lost in the 
averages. Dew was never visible on the sand during the hours in 
question. The averages are :— 
or nt of yet mae = ee of oe th N dmaber 
8 p.m. Pompe ae Clear Nights. 8 p.m. ere Rely Clear Nights. 
Under 35° 54 78 Under 30 °/, 8°3 22 
35°-39°°9 59 388 30 °/,-39 ,, 7-7 61 
40°-44°°9 61 49 A049). 65 55 
45°-49°-9 5:8 38 50 ,,-59 4, 47 56 
50°-54°°9 54 ily/ 60 ,, -69 ,, 4°4 39 
55° and over 4-3 4 70 ,, & over 3°9 51 
