386 - Transactions of the Boyal Society of South Africa. 
These facts suggest that next to the blanketing " effect of the clouds 
the relative humidity more or less modified by the velocity of the wind is 
alone of importance in determining the radiation-temperature gradient. 
This point we now proceed to emphasise. 
Table 6 shows for each percentage of relative humidity, in rising 
grades of 5 per cent, each, under clear skies : — 
Col. 1. The grade of relative humidity ; 
2. The mean relative humidity of the grade ; 
3. The number of observations in the grade ; 
4. The corresponding mean differences of temperature between the 
two thermometers ; 
5. The corresponding values of the dew-point temperature ; and 
6. The corresponding mean velocity of the wind for the hour ending 
at the time of observation. 
TABLE 6. 
Summary of Mean Eesults under Clear Skies. 
Relative Humidity. 
No. 
D. 
Dew-point. 
Wind. 
Grade. 
Mean. 
Times. 
F. 
P. 
M. per H. 
% 
% 
o 
Under 30 
26 
24 
8-3 
37-9 
4-6 
30-34 
32 
49 
7-3 
37-9 
4-8 
35-39 
37 
58 
7-8 
36-6 
3-6 
40-44 
42 
66 
7-2 
330 
3-9 
45-49 
47 
78 
6-9 
33-5 
4-5 
50-54 
52 
110 
7-0 
35-1 
4-3 
55-59 . 
57 
111 
7-1 
35-2 
3-6 
60-64 
62 
103 
7-5 
34-4 
3-0 
65-69 
67 
98 
7-6 
35-1 
2-6 
70-74 
72 
76 
7-1 
36-1 
2-2 
75-79 
77 
50 
5-8 
38-3 
2-7 
80-89 
84 
88 
5-5 
39-1 
1-9 
90-100 
92 
39 
5-9 
40-5 
1-5 
Evidently the dew-point effect is slight at the best, if it can be said to 
have any existence at all. On the whole the temperature differences D 
fall as the relative humidity increases. But there is an important break 
in the regularity of the fall, and D shows a rise, instead of a fall, all the 
way from a mean relative humidity of 47 per cent, to 67 per cent. A 
reference to the column of wind velocities furnishes the explanation, 
