566 



NATURE 



\_Sept. 26, 1878 



this kind we can scarcely venture to go beyond actual ex- 

 perience. 



It would be easy to multiply similar examples and results of 

 the method which I ventured to submit in my former paper to 

 Nature — a method the main object of which was to refer, as 

 nearly as possible (without using the more laborious method of 

 the harmonic analysis), the rainfalls of remote localities to the 

 epochs of maximum and minimum sun-spots. I will, however, 

 for the present only give, further, the results that have been 

 obtained for the Paris rainfall from 182410 1867, and from l8l6 

 to 1870. 



Table VII. — Rainfall of Paris {^Maximum Years in Sixth 

 Line). 



The above table shows that, whether we take the "means" 

 or the "mean cycle," the rainfall was greatest in the years of 

 maximum sun-spot ; that it was least in the ninth year of the 

 "mean cycle;" and that, on the whole, the rainfall and sun- 

 spots, notwithstanding some discrepancies, increased and de- 

 creased together. 



The next table, in which the arrangement is inverted, gives 

 similar results for Paris. 



We see that the minimum rainfall occurred, on an average, in 

 the year immediately preceding the year of minimum sun-spots, 

 as at Edinburgh, but that the variation was not so regular. 



As formerly remarked, the rainfalls of Edinburgh and Paris 

 — especially that of Edinburgh — are more favourable to the 

 theory than the rainfall of Madras. 



Mr. Buchan considers the method which has now been 

 sketched a new one, and, "as such, deserving of a careful ex- 

 amination as to how far it is applicable to the data submitted for 

 discussion." This examination consists almost wholly in showing 

 that by placing the maximum years in the same line or group the 

 minimum years are spread over six out of thirteen groups, and 

 that by placing the minimum years in the same group the maxi- 

 mum are also spread over six groups. Hence he concludes 

 that this double arrangement is inferior to a single one in which 



the maximum and minimum years together are "compactly" 

 spread over six out of eleven groups. But it seems to me that 

 he has in Igreat measure lost sight of what should be a main 

 object of compariscns of sun-spots and rainfall, namely, the 

 Closest possible reference of the rainfall to the epochs of maxi- 

 mum and minimum sun-spots, and that however compact the 

 arrangement he recommends may be considered, it is funda- 

 mentally objectionable. By placing the maximum and minimum 

 years respectively in the same groups there is certainly a much 

 greater chance of finding any connection that may exist between 

 the two phenomena than by spreading them over six groups 

 out of eleven. 



How far the method defended by Mr. Buchan is applicable to 

 the data will appear from the following table of the Madras 

 rainfall, in which Dr. Hunter's arrangement is adopted. The 

 maxima and minima years are marked with an asterisk. 



Table IX. — Rainfall of Madras {Maximum and Minimum 

 Years in Six Groups). 



The mean "rainfall for the cycle is 48*3 inches. Now the 

 mean rainfall for the fifth group is 58*3 inches, and the mean 

 value of the sua-spots for the same group 45*0, which is the 

 maximum. It is thus made to appear that, on an average, the 

 maximum rainfall of Madras coincides with the maximum of 

 sun-spots. But this is contrary to fact. We know, as a matter 

 of observation (see Table III.), that the mean rainfall of Madras 

 in the maximum years was not above the average, and yet the 

 arrangement recommended by Mr. Buchan makes it ten inches 

 above the average. 



Applying the same arrangement to the rainfall of Edinburgh, 

 we get the following results. 



Table X. — Rainfall of Edinburgh {Maximum and Minimum 

 Years in Six Groups). 



In the above table the sun-spot maximum occurs in the 5th 

 group, and the rainfall is made to be i*2 inch below the mean ; 

 that is, according to this arrangement the people of Edinburgh 

 are supposed to get less than their average allowance of rain in 

 the maximum year. But according to the observations pub- 

 lished by the Scottish Meteorological Society, that again is 

 contrary to fact, for (see Table V.) the rainfall of Edinburgh i^ 

 not I '2 inch below the mean for the maximum years, but, taking 

 the thirteen "means" five inches above it, and, taking the 

 " mean cycle " 3 inches above it. 



