~The Winds of Kimberley. 83. 
ponent curve, on the other hand, passes through its maximum and 
minimum points shortly after the two hours of mean temperature ; 
but apart from this there is no obvious connection between it and. 
either the temperature or the pressure curve. The long-drawn-out 
maximum of the winds prevailing at night, and the shorter but 
sharper maximum of the day winds, previously alluded to, here 
receive at least a partial explanation from the fact that the diurnal 
temperature is below its mean value for a longer period than it is 
above. | Rie 
Fig. 4 displays in one curve, and on the same scale, the resultants 
of the component curves of Fig. 3. In this figure northerly com- 
ponents are measured vertically upwards or downwards, according 
as they are positive or negative, easterly components being measured, 
horizontally to the right or left in the same way. The resultant 
wind for any hour is supposed to be moving inwards from the number 
denoting the hour to the origin of the co-ordinates. Thus, during 
the hour ending 9 a.m, the mean resultant direction of motion is 
almost due south, 7.c., the resultant wind is northerly. The curve is 
very nearly an ellipse, having its major axis on the line joining the 
hours ending at noon and midnight. It is a curious and possibly 
significant fact that if the times of maximum, minimum, and mean 
temperature be marked upon the curve, the line joining the times of 
maximum and minimum will intersect the line joming the times of 
mean temperature almost at the origin.* 
The angular velocity of the resultant varies considerably, being 
greatest for an hour or two after sunrise, and for about the same 
interval after sunset. For each quarter of a day the mean angle 
described is approximately—_ 
Brom midmieit to Gants.) .cclo ue) | OOF 
ee OF OEM GOUMOOM syst yer ocaataacaueetoe descr: 140° 
ee MOOI) Ould es wae Mea ilsckaca casas 70° 
Py On De tl yhOrUAIGINO WG: ce cress cet ea naa. 114° 
the east and west vibration showing itself by far the more influential. 
This point may be investigated in a rougher, if not less interesting 
way by tabulating the mean actual movement of the vane direct from 
the automatic registers for each quarter of a day. The method 
followed here was to consider the exact points (as nearly as they 
could be ascertained) at which the pencil crossed the respective 
* To prevent confusion, these lines are not drawn. Should the reader wish to 
insert them, it must be remembered that the horary numbers marked round the 
curve stand for the hour ending at the time indicated. Thus, 8.30 a.m., for 
example, should be taken as the mean time of the point IX. 
