Pressure and Temperature Results for the Great Plateau. 249 
and still more curiously, the mean date of the greatest maximum for 
Durban and Kimberley together is Jan. 10th—.e., half-way between 
Dec. 20th and Jan. 3lst—which is also, as it happens, the very 
mean date of greatest minimum—i.e., half-way between Jan. 4th 
and Jan. 16th. And here it is only fair to state that this neat result 
was only detected after Table 26 was fully complete, so that the 
dates were not ‘‘cooked’’ to bring it about. 
But perhaps the most useful application of the sine formula will 
be to the differences of mean minimum temperature, Durban minus 
Kimberley. Here it takes the shape— 
y = — (0-4) sin & + 12, 
the angles being reckoned from near the end of September ; nor 
does it differ anywhere by more than a few tenths of a degree from 
the observed differences. The value of so simple a connecting link, 
if only as an aid to memory, between the meteorological elements of 
the two places will be readily recognised. The calculated values, 
reckoned from just after the equinox, are given in Table 27. The 
corresponding calculated differences of mean and maximum tempera- 
ture having no special exactitude are not thought worth insertion. 
Lambert’s formula is commonly used when it is required to 
smooth out the irregularities of temperature or other meteorological 
curves to any great degree of exactitude. It is— 
T OL US oar 
a =p + p,cos= + g,sin— + p, cos — 
nN n n 
+ g,S8in ~— + p, cos — ee ae 2 
d n | 23 ny 1) B n 
in which a is the required temperature, or whatever it may be, and 
0) Dis Do, - - - Gis Gay Gz, . . Constants to be determined. - When a 
is the maximum, minimum, or mean temperature of any month, p 
is the mean yearly value, and p, approximately one-half the yearly 
R : 
range — corresponding, that is, to the T and ) used in the simple 
sine curve above. 
Putting n = 12 and solving Lambert’s equation by the method of 
least squares, we get the coefficients as in Table 28. The pressure 
numbers have also been added. But if 
Da Usa 
Gi= cos Va. &eC,, 
| 
