88 Transactions of the South African Philosophical Society. 
perature. . Judging by the diagram is it not highly probable that the 
secret underlying the double diurnal oscillation of the barometer is 
to be looked for behind the complicated variations introduced as. east 
components of wind-frequency and air-movement by the, simple 
application of the sun’s heat? At any rate this aspect of the’ pro- 
blem is worth further inquiry.* The greatest excess variation from 
the east is followed by the greatest excess variation from the west ; 
and, as it were, in response, the barometer also executes its greatest 
vibration. If it be not cause and effect, the coincidence is remark- 
able. However this may be, the diurnal northing and southing of 
the air, at any rate, both in duration and amount, is quite a secondary 
effect, indicating no more, perhaps, than an attempt to restore the 
equilibrium disturbed by the east and west exchanges. , 
From Tables 7 and 8 it is evident that the normal wind of Kim- 
berley, moves inwards, at any time, from a moving radiant. always 
situated some 30° or 40° in front of the point where a vertical circle 
through the sun cuts the horizon; the only departure being, as 
already noted, in the case of easterly winds, which, so far as the 
observations go, have one maximum somewhat earlier, and ‘thus 
contribute to the third maximum of the east component variation. 
These consecutive directions may all be included in the scheme of 
Fig. 8, wherein each instantaneous direction is tangential to a spiral 
circling outwards from the area of highest temperature and. inwards 
to the lowest. The simplest conception of the diagram is to imagine 
it, together with the hour numbers, an appendage of the sun encasing 
the earth (which rotates inside it), S and 8’ overlapping. H ‘is the 
area of greatest. heat nearly facing the sun, C that of greatest cold, 
KK’ the path described by Kimberley through the system with a 
velocity approaching 1,000 miles per hour. The arrow-heads show 
the direction of wind motion. 
The hypothesis demands that the diurnal wind system shall be an 
anti-convection current; but since the text-books frequently state 
that wind must blow from cold to warm areas this seems at first 
sight an impossibility. While it is easy to draw a similar spiral of 
opposite curvature which shall equally represent the hourly motions, 
and yet shall be, in a sense, a convection current, the one mechanical 
consideration fatal to such a construction is that a current running 
in to H, say from the south (H being south of the equator), must 
first curve to the left, and then gyrate clockwise round H. Now it 
is quite certain that if the sun could remain vertically over:one spot 
true convection currents must eventually be established, but it is not 
* See, howaver, Mr. H. EF. Blanford’s paper ‘‘On the Winds of Caleutta” j in the 
Indian Meteorological Memoirs, vol. i. p. 18. 
