114 SCIENCE IN AFRICA 



tions in view of the fact that in 1930 and 1931, when the sunspots 

 were nearing their minimum, the levels of all the lakes rose in an 

 unexpected manner when they should have been falling. 



It is important to note that some authorities do not believe that 

 a connection between the lake levels and sunspots has yet been 

 proved, particularly in view of the unexpected developments of 

 the last few years. Dr. H. E. Hurst, Director of the Physical 

 Department in Egypt, for instance, who has paid particular atten- 

 tion to the effect on the variation in level of Lakes Victoria and 

 Albert on the flow of the Nile, points out that Brooks' correlation 

 co-efficient of 0-83 between the level of Lake Victoria and sunspot 

 numbers for the period 1 896-1 922, has fallen in the period 1923- 

 1934 to o-o6, and for the whole period 1896-1934 to o-6o which is 

 scarcely significant. Furthermore, he shows mathematically that 

 there should be a difference of a quarter of a period, two and a 

 half to three years, between the maximum sunspot numbers and 

 the maximum lake level, whereas Brooks found these coincident. 

 For these reasons Dr. Hurst regards the correlation between lake 

 levels and sunspots as a chance relation which has now disappeared. 



If it can be concluded in general that the recent abnormal rise 

 of the great lakes is a temporary variation from a permanent cycle, 

 then we should expect that records of rainfall over long periods in 

 other parts of the continent would show similar cyclic variations. 

 The fifty-year record at Dakar, according to Welter (1930), shows 

 a striking correspondence with sunspot numbers (see page 105), 

 but appears to be the only case; Schumann and Thompson (1934) 

 in their study of South African rainfall conclude that 'cycles of 

 fixed periodicity are not in evidence in South African rainfall. If 

 cycles do exist, they are more or less completely masked by the 

 much more prominent irregular secular variations.' They are 

 dealing, however, with a part of Africa where conditions are those 

 of temperate rather than tropical latitudes. 



It is interesting to consider the effect of these periodic changes 

 on biological cycles and even cycles of human activity, though 

 here we are mainly in the realm of speculation. It has been sug- 

 gested that the eleven-year cycle influences plant growth and so 

 the incidence of locust invasions. Simmons (1929) has pointed out 

 the striking fact that the periods of food shortage in Uganda, 



