Pressure and Temperature Results for the Great Plateau. 273 
point to this origin. They are purely a coast phenomenon, as the 
high temperature does not extend far inland. 
“These coast hot winds appear to be due to an intensely heated 
lower stratum of air from three to five thousand feet in thickness, 
being confined beneath a uniform heavier upper stratum of cold air, 
which, owing to its homogeneity, cannot be broken through by the 
lower lighter stratum of hot air. There is good reason to believe 
that this arises from the approach from the N.K. of a thick mass of 
cold air slowly sinking from the upper strata of the atmosphere. 
As this impinges against the mountains and highlands up-country 
it is deflected, and drives the lower-lying hot strata of air partly 
before it, but in great part back along the low-lying surface toward 
the sea, in the form of a N.W. hot wind. The thicker the stratum 
of hot air, the further inland extends the hot wind, but usually the 
less intense the heat. 
‘‘ Despite these hot winds the months of September and October 
were not abnormally hot, the intervening days being usually rather 
below the average temperature.” 
It is not to be denied that this interesting account, and the quoted 
temperatures, do prove the very high temperatures to be purely a 
coast phenomenon, but the assigned origin is not so certain, nor is it 
shown where the low-lying stratum of hot air gets its heat from. 
This is a problem of sufficient importance to justify rather more than 
a superficial notice. And there are certain material and related facts 
which may be of utility in assisting us, or some future investigator, 
in framing some sort of a solution. These facts reveal themselves 
very plainly in Table 41. In this Table the Durban pressures at 
9 a.m., the Kimberley pressures at 8 a.m., and the maximum shade 
temperatures at both places, are arranged, as departures from the 
normal daily means of Tables 1-24, fcr the day of a hot wind and 
for three days before and after. There were six occasions to con- 
sider in the ten years 1890-9, three being in 1896. The average 
condition follows the separate details. 
We notice at once that the rule at Durban is for the temperatures 
to rise gradually for two or three days before the final rapid develop- 
ment of heat. The pressures begin the period well above the normal 
—nearly a third of an inch in one instance, nearly a tenth of an 
inch on an average—the curve being almost the inverse of the 
temperature curve, and agreeing moderately well with the formula 
proposed above as connecting the annual curves of the same two 
elements at Kimberley. But now we have the conspicuous and 
noteworthy fact that at Kimberley the maximum temperatures, and 
the pressures, are both greatly above the normals throughout the 
