132 Report S.A.A. Advancement of Science. 



From this Table I have constructed the wind frequency curve 

 in polar co-ordinates (the vectors are the resultants of the third 

 column) the two other curves are, the first, the curve of the frequency 

 resultants, the next that of the movement resultants in rectangular 

 co-ordinates. 



The frequency resultants all remain between N.E. and E.S.E., 

 the general resultant being nearly due E. This brings well out the 

 fact that we are quite in the trade wind zone. The curve in polar 

 co-ordinates for the wind movement would give similar results, 

 throwing only the resultant about 10 degrees South of East, and 

 bringing the major axis of the ellipse due North. I did not construct 

 it, as the minimum shown at 4 hours on the third curve gives a loop 

 that cannot well be smoothed away. 



This result is very different from those got in Johannesburg and 

 in Cape Colony. In Johannesburg the general resultant is N.N.W., 

 in Kimberley it is W.N.W. It compares very much better with the 

 wind system North- East of us, as found in Boroma, near Tete, on 

 the Portuguese Zambesi* where the main wind is from E. to S. E. 



The frequency curve constructed as explained above supposes a 

 uniform velocity in the winds in all directions. This is obviously not 

 correct ; the two curves in rectangular co-ordinates show the differences 

 introduced by taking into account the variable velocity. 



The period of time used in the computation of the wind movement 

 is evidently too short to smooth out the irregularities introduced by 

 errors, and by the variations so capricious a variable as the wind 

 velocity. I am therefore unable to state whether the minimum at 4 

 hours corresponds to a physical fact or not. Mr. Sutton found also 

 two minima for the wind velocity at Kimberley, and Mr. Innes notes 

 that there was a loop in the curve he gives for the wind movement 

 at Johannesburg. This loop corresponds also to a minimum. 



The two curves, however, give a minimum in the afternoon and a 

 maximum at night, and the movement curve has a second maximum at 

 about 10 a.m. 



Both the wind variations and the peculiarities we find in the 

 movement curve seem to me to fall in fairly well with the explanation 

 Dechevrens gave for similar but more complicated phenomena at 

 Zikawei (Shanghai). He explains them by something like the 

 Monsoon theory, that is, by the flow of air from the colder to the 

 warmer points. This applied to Bulawayo would work out in the 

 following way. During the day the warmer points are successively 

 E.N. and W. and the wind would veer that way; during night, on 

 account of the wedge shape of South Africa the warmer mass would 

 be that of the ocean South and the wind would veer through South to 

 the East again by sunrise. The force of the wind would depend on 



* Cfr I) The winds of Kimberley. J. R. Sutton. Trans, of the S.A. 

 Phil. Soc, Vol. XI S.I. 



2) Transvaal Meteor. Department. Annual Reports. 1904-5- 



R. T. Innes. 



3) The Meteorology of South Africa. Ch. Stewart. 



4) Beobachtungen zu Boroma, 1891-2. Von L. Manyhart, 

 S. ]., in Publicat des Haynald observ. (Kalocsa) Heft VII. 



