386 On Rainfall in connexion with Sun-spots. [Mar. 9, 



present investigation warrant the inference that, generally, the rainfall 

 of the next two or three years will be below the average rainfall of the 

 years 1870-3, and that, though some places may suffer from floods, more 

 will suffer from drought. Prom all the information I have been enabled 

 to collect, it is not improbable that the rainfall of this year (1875), re- 

 markable though it is for destructive floods in various parts of the world, 

 will be less than the average. Floods are, apparently, characteristic of 

 the setting in of the minimum period of sun-spots, especially on the 

 continent of Europe. But these floods are local, and the large vintage 

 in Central Europe is an indication that, generally, the rainfall there for 

 at least several months of the year has not been excessive. 



A completely satisfactory proof of a connexion between sun-spots 

 and rainfall requires that the variation in the rainfall must necessarily 

 follow from the variation in the sun-spot area. We may show that the 

 two periods, as far as observation extends, are equal, that the intervals 

 between their maxima and minima epochs are equal, and that the times 

 of the epochs themselves are exactly what might be expected. It might 

 also be argued that periodic constitutional changes in the sun, as in- 

 dicated by the solar spots, must produce atmospheric changes. But 

 granting all that, it may be said that so long as there is a bare possibility 

 of explaining the phenomena otherwise, the evidence is incomplete. 

 There may be two independent causes at work, one producing a sun-spot 

 and the other a rainfall cycle, and the two causes and effects may have 

 run nearly parallel to each other for many years ; but it does not follow 

 that they will always do so. 



Such a possibility may exist, but it seems to be a rather remote 

 one ; .and in order, if possible, to make it still more remote, it may be 

 desirable to compare the amount of sun-spots and rainfall for each day 

 and month. 



But supposing it were fully established that the rainfall and sun- 

 spot variations are produced by one and the same cause, it may be asked 

 whether the corresponding eleven-year period of magnetic variation, dis- 

 covered by Sir Edward Sabine, has also the same cause; and if so, by 

 what process are such apparently dissimilar effects as rainfall and mag- 

 netic variations produced by a common agent? I would hazard the 

 opinion that the proximate cause of the sun-spot, magnetic, and rainfall 

 cycles is a secular variation in the intensity of solar heat, the solar spots 

 or cyclones being, like terrestrial cyclones, most frequent when the 

 temperature is greatest, that the sun radiating more heat in some years 

 than in others evaporation and rainfall are greater in the former than in 

 the latter, and that the changes in solar radiation affect the earth's 

 magnetism. This opinion seems to receive some support from the cir- 

 cumstance that, according to some remarks recently made by Sir George 

 Airy at a meeting of the Institution of Civil Engineers, and to researches 



