564 



NATURE 



\SepL 26, 1878 



attempt at explanation — while everyone of them might 

 have been made part of his nature for ever afterwards, 

 by giving the simple reason in each case — if he does not 

 doubt the competence of his teacher, will have no more 

 to do with him. 



La Spectroscopic. By A. Cazin. (Paris : Gauthier- 



Villars.) 



The talented author of this work has passed away since 

 the MS. was completed. This is by no means a sys- 

 tematic treatise, but it contains a large amount of informa- 

 tion — some of it out-of-the-way information — and it will 

 repay perusal. As much of M. Cazin' s information on 

 the celestial applications has been gathered from Secchi' s 

 works its complete accuracy is not to be relied on, but 

 the explanation given of the different methods employed 

 is very clear. 



The part of the book which perhaps will be read with 

 the greatest interest is that dealing with radiation and 

 absorption spectra. In this part the author includes a 

 notice of much of his own work, which is of great in- 

 terest and importance. The historique of the question 

 as to the existence of double or multiple spectra is in- 

 teresting, ^and the author's leaning is against the view 

 held by Angstrom and Thal^n. He gives special obser- 

 vations of his own concerning nitrogen, and indeed was 

 engaged on an extension of them at the time of his 

 lamented death. 



LETTERS TO THE EDITOR 



yrhe Editor does not hold himself responsible for opinions expressed 

 by his correspondents. Neither can he undertake to return, 

 or to correspond with the writers of, rejected manuscripts. 

 No notice is taken of anonymous communications. 



[The Editor urgently requests correspondents to keep their letters as 

 short as possible. The pressure on his space is so great that it 

 is impossible otherwise to ensure the appearance even of com' 

 munications containing interesting and novel /acts.} 



Sun-spots and Rainfall 



If the sun-spot cycles were all of the same length the simplest 

 way of comparing the yearly sun-spot areas with the yearly 

 amounts of rainfall in order to see whether the two phenomena 

 were more or less numerically related, would be to find the 

 annual means of the greatest possible number of cycles, care 

 being taken to place the years of maximum and minimum in the 

 same two groups respectively, and the intervening years in due 

 succession. But as the sun-spot cycles are not of the same 

 length, we must in employing the method of arithmetical means, 

 make some modifications calculated to suit the circumstances of 

 the case. 



1. We may, for example, take any cycle whatever may be 

 its duration, and commencing with its first and ending with its 

 last year, compare with the sun-spots the rainfall at any station, 

 or the means of the rainfall at a number of stations for the 

 same years. 



2. If some of the cycles be of the same length, we may take 

 these alone and compare them with the rainfall, still taking care 

 that the years of maximum and minimum sun-spots shall be 

 respectively in the same groups. 



3. The average length of the sun-spot cycle being, as far as 

 is yet known, about eleven years, we may take any number of 

 cycles of different lengths and make two separate comparisons, 

 in one of which the maximum years are to be placed in the 

 same group, and in the other the minimum years in the same 

 group, the number of the other groups preceding and following 

 the epochal groups being determined by the fact that the mean 

 interval from minimum to maximum is about 37 years, and 

 from maximum to minimum about 7*4 means. 



I have tried these and several other methods, with, I think, 

 considerable success. The method just mentioned (3), which is 

 the old one of arithmetical means somewhat modified to meet 

 the conditions of the case, possesses several advantages, one of 

 the most important of which is that it enables us to compare 

 directly the rainfall with the sun-spots in the epochal years, and 

 in at least two years before and two years after them, thus 



affording a fair comparison for nearly the whole cycle of eleven 

 years — a most essential point, which, it would appear, has been 

 overlooked by some writers on the subject. 



Having in my last communication (Nature, vol, xvii. p. 448) 

 given an example of the fii'st half of the above method, together 

 with the results obtained by it for several localities, I intended 

 on this occasion to submit only one or two examples of the 

 second half. But finding that the method has been criticised by 

 Mr. Buchan, and having now Prof. Wolf's latest edition of his 

 relative sun-spot numbers, as well as the rainfall of Madras for 

 1877, it may be proper to give instances of the application of 

 the whole process. 



Let us begin with Wolf's relative numbers, which are so 

 arranged in the following table that those for the years of 

 maximum sun-spot, from 181 1 to 1877, are all in the sixth line. 

 This table has already been given by Mr. Buchan (Nature, 

 vol. xvii. p. 506), but as references will be made to it now, and 

 also in future discussions of the rainfalls of various parts of the 

 globe, it is desirable to reproduce it here. 



Table I. — Wolf's Sun-spot Numbers {Maximum Years in 

 Sixth Line). 



It will be seen that each of six of the columns in the above 

 table gives the sun-spot numbers for thirteen vears, and that the 

 first term of what is called the "mean cycle ' is obtained from 

 the expression, '^ + 2 + c ^ ■where a, b, c, are the means of the 



4 

 sun-spots for the first, second, and third years of the thirteen 

 years, the remaining terms being obtained in a similar manner. 

 The " variation" is simply the deviations from the mean value 

 of the sun spots for the " mean cycle." 



The next table gives the sun-spot numbers, from 1816 to 

 1872, arranged so that the minimum years are in the eighth 

 line : — 



Table II. — Uolf^s San-spot Numbers {Minimum Years in 

 Eighth Line). 



We can now compare the rainfall of Madras with the values 

 of the sun-spots in Tables I. and II., except for the years 181 1 

 and i8i2, for which there are no observations :— 



