274 Transactions of the South African Philosophical Society. 
week, the temperature departures, in particular, for the first three 
days being twice as great as those of Durban. But the maximum is 
reached a day earlier, and the succeeding fall is gradual. During 
the whole week the Kimberley pressures are slackening almost uni- 
formly, a faint and scarcely perceptible minimum commanding the 
averages on the same day as the deep minimum of the coast 
It appears, then, that for some days previous to a hot wind on the 
coast pressures and temperatures are both exceptionally high on the 
table-land. Such a pecuharity, occurring at a time when the surfaces 
of equal pressure are always greatly elevated over the interior, will 
obviously give an additional impetus to the (supposed) overflow of 
air common at this season. The paucity of the necessary meteoro- 
logical records makes it impossible to assert with any confidence what 
will be the final result at the lower levels of the increased overflow. 
We know enough of the winds of Kimberley to be able to predict 
that the overflow will probably take a spiral path. But it is not at 
all certain to what extent and how far the operation will be propa- 
gated. Manifestly the gradual slackening of the high pressures 
above does not communicate an immediate increase below. Possibly 
the enforced outflow sets up some sort of compression in the upper, 
or rather middle reaches of the atmosphere over the coast, making 
a still further increase of temperature there ; which again sets up a 
vortex motion, and an upcast, where otherwise we should have had 
augmented pressure. The littoral high temperatures may therefore 
arise from the compression, and friction, of the expanding air of the 
high land upon the elastic walls of this natural chimney. Of course 
the compression of itself, whatever its possible superior limit, would 
not be competent to develop the observed degree of heat. That is 
convincingly argued by the terms of the ‘‘ Characteristic Equation of 
a Perfect Gas,” pv = RT.* Lastly, in further confirmation of the 
possibility that the factors of a hot wind originate on the table-land, 
it may be added that the phenomenon is almost invariably accom- 
panied by clouds of the cirro-stratus type, and by a considerable 
falling off in the quantity of moisture present. 
The May depression does not greatly affect the temperatures while 
it lasts, albeit colder nights herald its breaking up and cold days 
succeed. There is apparently not much doubt of its really being an 
annual event, the mean dates of the double minimum departing prac- 
tically nothing from the 16th and 25th. There is a very similar 
depression at Adelaide (S. Aus.), but coming a little earlier, the dates 
of the double minimum being about the 11th and 17th. As will be 
seen from Fig. 2, the Durban maximum pressures are never very 
* Preston, Theory of Heat, p. 127; Greenhill, Hydrostatics, p. 281. 
