248 



NATURE 



[Al'RIL 28, 1910 



LETTERS TO THE EDITOR. 



[The 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 intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.] 



Precursors of Magnetic Storms. 



In kindly noticing (Nature, December 30, 1909, p. 259) 

 my short account of the magnetic storm of September 25, 

 attention was directed to the fact that what I called the 

 precursor was experienced all over the world. 



As I have little or no opportunities to compare our 

 curves with others, may I be allowed to ask for informa- 

 tion through the columns of Nature? 



I call a " sudden start " of the magnet a movement 

 which occurs after several hours of perfect calm, and 

 causes the trace to make a " sharp angle," that is, so 

 sharp that I can unhesitatingly tell the time of the occur- 

 rence to the nearest two minutes. Many disturbances, 

 large or moderate, have a " sudden start " — at least 

 here. 



My impression has long been that disturbances with a 

 " sudden start " very generally have a kind of " pre- 

 liminary tremor" some hours before the start; the curve, 

 which we suppose to be quite smooth, is interrupted by 

 a short movement, which lasts but a few minutes, after 

 which the curve resumes its smoothness for the remaining 

 hours. The tremor may be very small indeed, but the two 

 characteristics, to be found on a smooth curve, and to be 

 of very short duration, make it quite easy to point it out 

 and tell the time. 



The start of the " preliminary tremor " is in the same 

 direction as that of the disturbance itself, at least as a 

 rule. 



I do not venture to hold an opinion as to the connection, 

 fortuitous or otherwise, between the two phenomena, but 

 I should be very glad to know whether the " precursor " 

 is also observed in other countries. The following is the 

 list of all the " sudden starts " of H during the last 

 fourteen months of Zi-ka-wei (January, 1907, to February, 

 1908, inclusive), with the time of the start and that of the 

 preliminary. A comparison with the curves of some other 

 observatory is invited. Probably the traces will not be 

 found so smooth as here, and some of my " sudden starts " 

 will correspond to progressive starts, but on the whole I 

 hope that a comparison will be possible. 



I use Greenwich time. 



Remarks 



H not quite smooth. 

 D very smooth. 



Curves not quite smooth. 

 Curve not quite smooth. 



Curves not smooth. 



There is another pre- 

 liminary at 3^. 

 Curves not smooth. 



And also 29at i ih. 12m. 



The curve of H was 

 lost, until loh. But 

 D does not show any- 

 thing. 



If we put aside No. 21, we have twenty-three occur- 

 rences. In eighteen of these there is no doubt; in the 

 three cases in which no precursor was found, and in the 

 two more or less doubtful cases, the curves were not 

 smooth, that is, the " start " was not quite sudden. 



In Zi-ka-wei, during the fourteen months considered, 

 the eighteen disturbances which began with quite a sudden 

 start all had one or two preliminaries. The interval 

 between the forerunner and the disturbance ranges from 

 two to seventeen hours of perfectly smooth trace. 



Zi-ka-wei, China, March 11. 



J. DE MOIDREV. 



Centre of Gravity of Annual Rainfall. 



The mere a priori criticism which Mr. Watt, in his 

 letter to Nature of April 14, has bestowed on my letter 

 to Nature of March 31, dealing with a large class of 

 concrete physical facts, is, to my mind, far from satis- 

 factory or sufficient. In the last sentence of my letter 1 

 anticipated that it might meet with some such " simple " 

 algebraic criticism. 



I did not, as is implied, assert that the use of the C.G. 

 of the year's rainfall would dispense with the ordinary 

 graphic representation of the monthly distribution, but 

 that it appeared to be a convenient method for comparing 

 the rainfall distribution at a number of stations in the 

 same country for the same 3'ear, as well as at a single 

 station, or for the mean rainfall of the same country, for 

 a succession of years. Obviously, therefore, Mr. Watt's 

 simple types of rainfall for his imaginary stations A, B, 

 C, which belong to very different regions on the earth's 

 surface, have no application, even in theory, to my 

 " suggestions." 



As for the practical character of my proposals, the 

 following may be given. They are quite as useful as 

 the comparison of the variation of rainfall with that of 

 population for a decennium in India, which was the sub- 

 ject of a paper given some j'ears ago in the Journal of 

 the Royal Meteorological Society, London. A comparison 

 of the variations in the C.G. of the rhean monthly rain- 

 fall for the past 50 or 100 years with the agricultural 

 results of those years in the British Isles, &c., for which 

 the data may be available ; an examination of these 

 variations in connection with the much-discussed question 

 of weather cycles, so commonly based on rainfall statistics ; 

 the detection of serious clerical slips in the tabulation of 

 rainfall ; the interpolation of the probable rainfall figures 

 for a month in the event of a rain-gauge or measure-glass 

 being temporarily unserviceable ; the detection of the 

 ignorant or inadvertent use of a wrong measure-glass — a 

 matter of frequent occurrence in India, and possibly not 

 unknown in this country. 



Rain falls with such seeming irregularity of quantity 

 and date, even in India — and much more so in this " un- 

 speakable " Scotland — that it would appear prima facie 

 impossible that there could be any approach to constancy 

 of the date around which the whole year's rainfall 

 balances. If we look at the tables of monthly rainfall for 

 a large number of stations (in the same country) and see 

 that the figures are not even approximately the same for 

 the months or for the whole year, if we consider the 

 difficulties connected with the measurement of rainfall, 

 which are discussed in many of the volumes of " British 

 Rainfall," it is surely surprising that, in spite of all these 

 things, the reported year's rainfall should balance round 

 a date which does not differ by more than a few days 

 for a great many, if not all, the stations in any one 

 year, and that for another year the displacement of this 

 central date is so nearly the same for all of them. By 

 merely looking at the monthly figures or at the graphs 

 of those figures, we cannot accurately estimate either the 

 central date or the amount of its displacement. 



The causes which determine the times and amounts of 

 rainfall for any place or country are known only in a. 

 very general wav indeed. They are so elusive that investi- 

 gators in their despair have even had recourse to sun-spots 

 or comets' tails as a possible cause of special excess 

 or deficiency of rainfall. They are so elusive that even in 

 India, with its comparatively regular rainfall-seasons and 

 with its special equipment of experts, the problem of , 



