254 Transactions of the South African Philosophical Society. 
the monthly results published by the Meteorological Commission in 
Cape Town. | 
Theory and practice are contrasted in Table 31, by comparing the 
calculated monthly changes of temperature with those observed for 
Kimberley, Bloemfontein, and Aliwal, separately and together. The 
chief deviations occur Nov.—Dec., and June-July, the observed 
variation being in the one case 1°°9 greater, and in the other 1°7 
less. The other deviations are not important, and the gradients 
are remarkably similar. Annual hot and cold periods seem to 
account in the main for such differences as do occur. 
We gather from the close agreement between theory and observa- 
tion how quickly the table-land of South Africa becomes heated by 
the sun’s rays and how few of the rays it stores up. There is little 
resistance to or accumulation of heat at any time. Accepting for the 
present this responsiveness of the air to the action of the sun, we 
have now to face the curious and, at first sight, contradictory fact 
that the diurnal range of temperature at Kimberley is pretty well as 
great in winter as in summer, whereas, during the day at any rate, 
we should expect the winter range to be much less. For at the end 
of December and beginning of January the sun increases its altitude 
between VII. and Noon by fully 61 degrees of arc, but only by 
37 degrees between the same hours at the beginning of July. Now 
taking the mean daylight temperatures during December and January 
to be the mean daylight temperature of December 31st or January Ist, 
we should, if the cosine formula applies to this special case, expect a 
rise of temperature between VII. and Noon of some 42°. The actual 
rise from VII. to the maximum (which occurs about 1h. 20m. p.m.) 
is on the contrary only 19°. Again about July 1st the calculated rise 
would be 39°, whereas the rise to maximum (which occurs about 
2h, 35m. p.m.) is 27°-—assuming, of course, that the initial tempera- 
ture at sunrise is Independent of any previous solar heat. If we 
calculate new values of A and B in the cosine formula, by consider- 
ing only the rise of temperature to Noon, we shall get for the hourly 
rate of change— 
Winter y = 40°5S?cos Z + 36° 1 
Summer y = 31°4S?cosZ + 55°'8; 
or by considering the rise of temperature to maximum :— 
Winter y= 51°3S?cosZ + 36°°0 
Summer y = 33°68?cos Z + 55°0, 
In Table 32 will be found the result of the computation using the 
first of the two sets of coefficients :— 
