230 report — 1878. 



Report on Sunspots and Rainfall. By Charles Meldrum, F.R.S. 



[A communication ordered by the Council to be printed in extenso among the 



Reports.] 



1. In 1873 and 1874 (see British Association Reports for those 

 years) I submitted tables of the rainfalls of various parts of the world, 

 and expressed the opinion that there was strong evidence of a connection 

 between rainfall and sunspots. 



2. Having received additional observations, I now beg to submit the 

 principal results obtained by comparing the rainfalls of different countries, 

 and the levels of some of the rivers of Central Europe with Wolf's rela- 

 tive sunspot numbers. 



3. Probably the best method of comparing the sunspots with the 

 rainfall is that of the harmonical analysis. In a paper which was com- 

 municated to the Royal Society in January, 1876, I applied that method 

 to the annual mean rainfalls of the greatest possible number of stations 

 scattered over the globe, and to the mean annual depths of some of the 

 rivers of Central Europe, and found not only that there was a rainfall 

 cycle of nearly the same length as the sunspot cycle, but also that the 

 two cycles had the same characteristics with respect to the intervals 

 between the epochs of minimum and maximum and maximum and mini- 

 mum, a circumstance which strongly pointed to a causal connection. 

 But, as the method is laborious, I have not yet had time to apply it to 

 the rainfalls of single stations, or even to the mean rainfalls of different 

 countries. I hope to be able to do so soon, and to communicate the 

 results on another occasion. 



4. In the meantime the probability or otherwise of a connection 

 between sunspots and rainfall may be shown by the old method of arith- 

 metical means. 



5. Although the mean length of 'the sunspot cycle is about eleven 

 years, yet, in employing the method of arithmetical means, it would be 

 objectionable to commence with any year whatever in a long series of 

 observations, and taking the greatest possible number of periods of eleven 

 years each, compare the annual mean rainfalls with the annual mean sun- 

 spots ; for by doing so the maximum and minimum years might be so 

 much dispersed over the common eleven-year period thus formed as to 

 conceal any periodic variation that might exist. It is essential to refer 

 the comparisons to the epochs of maximum and minimum, and this can- 

 not well be done by commencing with any year whatever. 



6. With a view of avoiding that objection as far as possible, and at the 

 same time of obtaining a simpler and more expeditious method than that 

 of the harmonical analysis, I make two comparisons, in one of which the 

 maximum years of sunspots are taken for the point of reference, and in 

 the other the minimum years. 



7. As the epoch of maximum sunspots occurs on an average 3" 7 years 

 after the epoch of minimum, and the epoch of minimum 7"4 years after 

 the epoch of maximum, the maximum years in the first comparison are 

 all placed in the sixth of thirteen terms or series of years, while in the 

 second comparison all the minimum years are placed in the eighth or 

 ninth of other thirteen terms or series. Then, with the object of dimi- 

 nishing the effects of so-called accidental irregularities in the rainfall, the 



