April \, 1878J 



NATURE 



449 



years after it, or thirteen years in all. Then, with 

 the view of reducing the effects of the so-called non- 

 periodic variations, I took a mean of the rainfall 

 in the first and third of the thirteen series, and a 

 mean of that mean and of the rainfall in the second series, 

 and so on. This gave me eleven new means, which 

 I called the " mean cycle." Again, starting with a mini- 

 nimum year, I took eight years before it and four years 

 after it, and found eleven other new means in the same 

 way. To each set of results, or to a combination of both 



of them, I then applied interpolation formulae, and found 

 a well-marked coincidence between the sun-spot and rain- 

 fall variations. 



As the sun-spot cycles are not all of the same length, 

 it is evident that by starting from any one year and going 

 backwards over a long period, always using the same fixed 

 number, a maximum and a minimum year might fall into 

 the same group. 



Let me, by an example, explain my method more fully. 

 I take the Madras rainfall : — 



It will be seen that the individual years of maxi- 

 mum sunspots, 1 8 16, 1829, 1837, 1848, and i860, are 

 all in the same vertical column, and that all the years of 

 minimum sun-spots, except iSro, contribute to the forma- 

 tion of the beginning and end of the cycle. No doubt it 

 would have been better to have placed 1836 farther in 

 advance, but this would have altered the position of the 

 maximum year 1829. 



The results given by the above method show a double 

 oscillation of the Madras rainfall during the sun-spot 

 period, and I see (Nature, vol. xvi. p. 333) that Mr. J. 

 Allan Broun has found such an oscillation for Trevandrum 

 as well as for Madras, and this may be a characteristic of 

 the rainfall of the whole of Southern India. We have, 



however, evidence of a single rainfall oscillation for other 

 parts of India. Even Mr. Blanford now admits that there 

 is a periodic variation underlying all irregularities and 

 more or less coincident with the sun-spot variation. 



In consequence of the method adopted by Dr. Hunter 

 the years of maximum sun-spots, in place of being all in 

 the same group, are spread over three or four of the 

 groups from which he derived his mean cycle, and it is 

 probably owing to this that he missed what, I think, is the 

 real character of the rainfall cycle at Madras, as shown in 

 Table I. The annual average rainfall for each year of his 

 cycle, together with the deviations from the mean, is as 

 follows, and it will be seen that his cycle is very different 

 from the one given by my method : — 



According to Dr. Hunter's cycle, the rainfall of Madras 

 was in excess in the fifth year to the extent of ten inches, 

 whereas, according to mine, it was nearly three inches in 

 defect. He takes, it is true, the years 1868-76, which I 

 do not take, preferring to wait till I get the rainfall of 

 1877 ; but although 1870, which he puts down in his fifth 

 group, was a very wet one, the double oscillation still 

 exists, one of the minima occurring soon after the sun- 

 spot maximum. 



I must now come to Dr. Hunter's letter in Nature 

 (vol. xvii. p. 59), But, first of all, I may be excused for 

 saying that I do not think, some remarks he made about 

 a periodicity of cyclones in a former letter (vol, xvi. p. 455) 



were altogether calculated to put the matter in its proper 

 light. He says (p. 456) : — " M. Poey called the attention 

 of the French Acaddmie des Sciences to the subject five 

 years ago, and published, as far back as 1873, a list of 

 hurricanes in the West Indies from 1750 to 1873, in sup- 

 port of his views. Dr. Meldrum has worked the same 

 question as regards the [East] Indian Ocean," Now 

 if these words are meant to convey the impression 

 that M. Poey preceded me, and that I followed with 

 the cyclones of the Indian Ocean, all I can say is 

 that M. Poey himself gave a different version of the 

 matter. 



In his second letter (vol. xvii. p. 59) Dr. Hunter states 



AA 2 



