The Cycle Year 1905 and the Coming Season. 247 



time, out of his slender resources. In this diagram is shown the 

 monthly incidence of the 1902 rainfall. This diagram shows us 

 exactly during what months and over what areas the irregular 1902 

 rains fell. In the Cape Peninsula they culminated in June and 

 September. The heavy June rain of Cape Town was absent in 

 the south-west generally, but the September rains ran through the 

 whole of the southern and western coast districts. In the summer 

 rainfall areas of South Africa the rains came during spring and 

 autumn, except in Kaffraria, where there was rain in June. But 

 the most instructive part of the diagram is the " abnormality " one. 

 This shows that the abnormal rains throughout South Africa 

 occurred almost entirely in June and September. In July there 

 was a drop to normal rainfall. There was abnormal drought during 

 spring and autumn, and slight drought during July. 



The same diagram published with the Meteorological Commission 

 Eeport for 1903 shows that the excessive rain occurred in June in 

 the south-west and in November on the southern coast. Though 

 the rainfall amounted to 94 inches on Table Mountain (St. Michael's), 

 the general character of the year in Cape Colony was " an exception- 

 ally severe drought lasting throughout the year," the year's mean 

 for all Cape stations being 19 per cent, below the normal (Report of 

 Meteorological Commission for 1903). 



The real significance of these 1902 and 1903 irregular rains will 

 be seen when it is considered that the effect of them and of the un- 

 usually heavy last Meldrum cycle rains (1892) has been to raise the 

 mean rainfall of the Royal Observatory by nearly \ inch. Up to 

 1888 the mean rainfall at the Royal Observatory was 25*43 inches. 

 It is now 25*9 inches. 



CONCLUSIONS. 



In view of the more ample material that has accumulated since 

 1888 the following conclusions seem justified : — 



1. The three main weather cycles are of general application 

 throughout South Africa. I had considered in 1888 that the Storm 

 cycle brought practically no rain to the eastern stations, and Mel- 

 drum's cycle little or no rain to western stations. The experience 

 of the last 17 years shows that both may extend east and west 

 beyond their area of greatest influence. 



2. Observations from the northern stations are as yet too short to 

 draw safe conclusions, but they seem to indicate that the pulse of 

 heavier rainfall occurs a season earlier at northern stations (Trans- 

 vaal and Rhodesia). 



