344 Transactions of the Boyal Society of South Africa. 
ration of the winter and summer parts of colmxm (5) of Table 2 gives us 
the following result : — 
Eange of Monthly Dew-point Temperatuke compared with 
Monthly Eange of Temperature. 
Winter. 
Summer. 
A 
-3-4 
+ 1-9 
B 
3-3 
2-1 
C 
31 
2-2 
D 
3-1 
2-3 
E 
2-8 
2-7 
F 
2-8 
• 3-0 
G 
2-6 
2-8 
H 
2-6 
2-8 
I 
2-5 
2-7 
J 
2-5 
2-8 
K 
2-2 
2-7 
L 
1-7 
2-5 
M 
1-5 
2-4 
N 
1-6 
2-2 
giving the curious result that in the winter half-year the diurnal variation 
in the total moisture content of the lower air diminishes as the air 
becomes damper, whereas in the summer half-year the corresponding 
variation is less both when the sky is very clear and when it is very 
cloudy — an intermediate state of things with regard to the sky (and of 
course with regard to the range of temperature) giving the greatest range 
of dew-point from VIII. to XIV. o'clock. Summer and winter taken 
together obliterate this purely summer effect. The relative humidity has 
no similar analogue. 
The only other question discussed in this paper is the effect of a 
variation in the range of temperature upon the harmonic constants in 
the diurnal curves of temperature and barometric pressure. 
For any given month arranged as in Table 1, the first five lines A — E, 
and the last five J — N, are taken, and the mean diurnal curves of tem- 
perature and barometric pressure of each for the given month for the 
years which happen to fall in these respective two sets are computed. 
Thus, for example, the mean diurnal curves of temperature and baro- 
metric pressure for July for the years 1898, 1902, 1901, 1897, and 1906 
are computed for the greater range of temperature (which here averages 
32°-5) ; and the same two elements for the same month for the years 
1909, 1910, 1908, 1900, and 1899, are computed for the lesser range of 
temperature (which here averages 27°' 0). 
