The Barometer in South Africa. 



87 



for Bloemfontein, Bulawayo and Salisbury, from which three inland 

 stations we also receive daily wires. The empirical reduction is 

 derived in this way : — Firstly, we assume the barometer reading and 

 its altitude above sea-level are correct, and by means of the usual 

 tables reduce the reading to sea-level. Then, by a comparison with 

 neighbouring stations and isobars drawn having regard to the 

 prevailing winds, we deduce a correction to our first sea-level 

 reduction This is done from day to day and the constancy of the 

 correction will be a justification for its regular adoption. We then 

 attempt to draw isobars, but a terrible gap exists north of Cape Town 

 and west of Christiana, so that allowances have to be made and some 

 guesswork indulged in. Before coming to synoptic maps thus made 

 let me explain by a few simply drawn diagrams a first theory of 

 South African weather, not that I claim that it is correct, but it may 

 serve to fix our attention, until it is replaced by some other. We have 

 to account for the following facts : — 



1. The Cape Peninsula has its wet season in winter. 



2. The extreme south coast has rain at all seasons. 



3. The rainfall at Durban is a summer rain, but it has two 



maxima, one before the other after the New Year. 



4. The overlapping of these three periods of rainfall. 



In accordance with Australian and South American experience we 

 assume that a series of cyclones are traversing the southern ocean from 

 W to E, in summer time far south of Africa, in winter time just 

 skirting South Africa. In winter time there is a permanent 

 anticyclone over South Africa, which in a general way wards off the 

 cyclones so that they are forced to skirt the coast, and their tendency 

 to move away from the equator is intensified. The diagram shows the 

 ordinary course of a winter cyclone. 



