PKESIDP-XTIAr. ADDRKSS SKCTIOX A. 27 



wind is accepted as without doubt correct l)v Xil> Mlkolni 

 and Koppen, as far, at least, as fast moving cyclones are con- 

 cerned. It seems to me, however, that in the application of the 

 normal wind there lies a real practical dit^culty. The figures 

 agreed upon as representing the normal wind for a certain 

 gradient are far from representing the exact value of the normal 

 wind. This value has not been derived from theory, but from 

 the discussion of a vast amount of statistics. But the discussion 

 of the statistics shows that tliis value can only be taken as a 

 general guide, and cannot be taken as the basis of very exact 

 determinations. For we get different values and sometimes 

 vastly ditferent values, for the normal wind in different quad- 

 rants of a cyclone, for different seasons, and very dift'erent 

 values also for the high seas, the coastal regions, and the main- 

 land. From the figures given by Loomis I suspect that we will 

 also find very different values for the normal wind for a given 

 gradient in cyclones and anticyclones. It seems, therefore. 

 that the judgment of the normal wind is largely an affair of 

 experience, and that the application of these rules will take time. 

 It is particularly difficult to apply them where investigations 

 bearing on the normal wind have not been made. Hitherto such 

 investigations have been mainly confined to Europe. Clement 

 Ley. Sprung, Koppen, Kassner, and others have analysed a vast 

 amount of data, but it cannot be said that the results they arrived 

 at are sufficiently uniform to allow a purely mechanical appli- 

 cation of Air. Guilbert's system. Time will prove whether this 

 system can do all that it claims, but whatever may be the ultimate 

 result, from what I have said it seems evident that it deserves 

 ■careful consideration. 



Whatever method be adopted, it is clear that weather fore- 

 casting, to be efficient, can be vmdertaken only by a central 

 organisation which can control observations over a large area. 

 Let us hope that the time is not far off when South Africa also 

 will have a Weather Bureau to centralise the observations taken 

 in as many stations as possible south of the Equator. The 

 most certain fact in South African weather conditions is the 

 general i)rogression of pressure .systems from west to east. At 

 the present moment the western half of South y\frica is really 

 a " dark continent " as regards meteorology, and we know 

 scarcely anything as to weather changes taking place there. The 

 South African weather problem will never be solved 

 with such an unknown region lying to our west. \\'e must be 

 able to connect up movements of the barometer at Swakopmund 

 and Port Xolloth with subsequent movements at Kimberley and 

 Bloemfontein, and to this end there must be a chain of stations 

 across the continent. This is, perhaps, according to the ex- 

 perience of the Transvaal forecasting department, the most 

 important need in South African meteorology' at the present 

 moment. 



