442 Transactions of the South African Philosophical Society. 



The fourth prolonged drought at Durban, from February, 1899, to 

 May, 1900, was accompanied and followed by a prolonged drought 

 in Mauritius from November, 1899, to December, 1900. 



As a large proportion of the summer rainfall at Mauritius is 

 usually of cyclonic origin, the above results suggested that the 

 summers following winter droughts at Durban would be characterised 

 by the absence of cyclones in the neighbourhood of Mauritius, 

 though it would be difficult to give a satisfactory explanation of such 

 a connection. Upon examining the Mauritius records, however, it 

 was found that the mean number of cyclones within the 20° square 

 of which Mauritius is the centre was 3- 3 in the summers following 

 winter droughts at Durban, while the mean for all summers 

 since 1873 was 3*8. As the numbers of which 3*3 is the mean 

 varied between and 6, the small difference (3 -8-3-3) does not 

 appear to have any real significance, and may possibly have changed 

 sign had a larger or smaller area been considered, or a different 

 number of years used. It appears, therefore, that winter droughts 

 at Durban are not necessarily followed by an absence of cyclones in 

 the neighbourhood of Mauritius, and hence that their effect on the 

 non-cyclonic rain of the following summer in Mauritius is even 

 greater than indicated by the figures given in this paper, and that 

 after such droughts Mauritius is dependent on cyclones to a greater 

 extent than usual for the following summer rainfall. 



It remains to be seen whether further observations from stations 

 along the east coast of South Africa will bring to light a still closer 

 connection between the rainfall over this region and in Mauritius. 

 Though no claim is made that an accurate monthly forecast of 

 summer droughts in Mauritius can be made from a study of the 

 rainfall along the east coast of South Africa during the preceding 

 winter, yet it appears from the figures given above that one, if not 

 the principal determining factor has been discovered. As other 

 factors are discovered and the broader effects of solar influences 

 become better understood, it is to be hoped that seasonal forecasts 

 of a fair degree of accuracy will be possible. 



