168 Mr. Eliot's Remarks on Rainfall. [Nov., 



Mr. J. Eliot fully agreed with the remarks of Mr. Blanford. It would 

 not be possible, he believed, to obtain exact relations such as appeared 

 to be given in the paper from the examination of the variations of less 

 than five or six sun-spot cycles. The relations were derived from a 

 period of less than two cycles, and the results could only be regarded 

 at present as little more than interesting and suggestive coincidences. 

 The weight of evidence at the present time seemed to be in favour of the 

 assumption that the larger variations of rainfall in India were not direct- 

 ly and chiefly related to variations of sun-spot frequency, but to 

 changeable local meteorological conditions which might in part be due 

 to variations of the intensity of solar radiation. For example during 

 the past monsoon, half of Northern India including Bengal and Behar 

 has received about 10 per cent, less rain than the normal, whilst in the 

 other half comprising the North- Western Provinces, Central Provinces, 

 and the Punjab, he believed, the rainfall for the same period was in 

 excess to at least the same extent. By no valid process of reasoning 

 could such variations be directly correlated to the sun-spot frequency. 

 Hence unless much larger periods are dealt with, or the variations due 

 to changing local meteorological conditions are eliminated in such in- 

 vestigation, the conclusions based on such evidence would always be of 

 doubtful validity and hence valueless for practical meteorological work, 

 such as the prevision of the character of either the winter or summer 

 rains in India. And the more exact the relations so obtained, the great- 

 er would be the probability of their being mere coincidences, which would 

 disappear when additional evidence or data were utilized. 



Mr. R. D. Oldham made some enquiries as to the periodicity of the 

 sun-spot cycles, and their effect on meteorological and other phenomena. 



In reply to Mr. Oldham, the Pkesident said: — " The supposed varia- 

 tion of certain terrestrial phenomena, uniformly with the variation of the 

 sun-spots, rests at present on a purely empirical basis, and the degree of 

 confidence with which the reality of such variations may be accepted is 

 very different in different cases. That least open to doubt is the cyclical 

 variation of magnetic disturbances, which is not I believe questioned by 

 any physicist who has ever examined the evidence, and that of the aurora 

 borealis, the connection of which with magnetic disturbances is very 

 intimate. 



"No meteorological phenomenon (using the term in its most re- 

 stricted sense) exhibits the cyclical variation so undeniably and in so 

 marked a degree as does terrestrial magnetism : but in the case of tem- 

 perature, Koppen has shown with much probability that, in the tropics, 

 the mean annual temperature has varied by a small amount pari passu 

 with the varying frequency of sun-spots during two sun-spot cycles, 



