78 



NATURE 



[March i6, 191 i 



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 o/ anonymous communications.] 



The Non-simultaneity and the generally Eastward 

 Progression of Sudden Magnetic Storms. 



Under the above hcadint*. Dr. L. A. Bauer contributes 

 to Naturk of .March 2 a letter of five columns. Referring 

 to his Table II., p. 10, he says:—'* It will be noticed that 

 Kew is not included, for the simple reason that although 

 Dr. Chree scaled the required data some months ago, he 

 has not vet published them nor forwarded them to me " ; 

 and later he adds : — " Dr. Chree could not have done 

 better than immediately to have published his own data in 

 the same open manner that Mr. Paris had done. Instead, 

 he labours to discredit the Coast and Geodetic Survey 

 observations, and withholds his own from public scrutiny. 

 In half the interval of time between the first and second 

 presentation of his paper, had Dr. Chree chosen, he could 

 have had at his command data from Europe and Asia 

 which, combined with his own, would have served admir- 

 ably to have tested the main contentions." The real facts 

 are as follows. In July, 1910, I consulted the Kew curves 

 for the dates of the fifteen disturbances treated by Mr. 

 Faris and Dr. Bauer, and, as stated in the paper read 

 before the Physical Society (Proc, vol. xxiii., part i.) on 

 November 11, "I was able to identify ten of the fifteen 

 disturbances with reasonable certainty." The measure- 

 ments I then took were confined to the times of commence- 

 ment of these ten disturbances. Dr. Bauer's request for 

 data, which reached me in January, included, not merely 

 the times, but the amplitudes of all the movements. I 

 supposed, mistakenly, as it proves, that before publishing 

 anything Dr. Bauer would await the data from the more 

 remote stations, which could not well reach him in less 

 than two or three months. I thus gave precedence to 

 official work of urgency. Also, to get the best results 

 possible, I had an independent set of measurements made 

 by my chief assistant, and took a fresh set myself, and 

 considered carefully all cases in which the times obtained 

 differed by m<5re than one minute, the limit of accuracy 

 I hoped to attain. In a good many cases I was doubtful 

 which of several small movements was the one intended, 

 so curves were drawn indicating the movements and times, 

 and amplitudes were given for the various alternatives. 

 The results of this very considerable labour were dis- 

 patched to Dr. Bauer on February 25, no hint having 

 meantime been received from him that the data were 

 urgently wanted. 



The charge that I laboured to discredit the Geodetic 

 Survey observations is equally unjustifiable. As anyone 

 who reads my paper will readily recognise, my criticisms 

 were directed, not against the observational data — which, I 

 may say, struck me as quite up to the usual standard of 

 observations — but against the use made of them. In 

 bringing forward these criticisms, I hoped to do better 

 service than by collecting a number of other miscellaneous 

 data, affected by similar uncertainties, the course which 

 Dr. Bauer thinks it was my duty to adopt. 



As Dr. Bauer takes no notice of my criticisms, I should 

 like to indicate briefly their nature, to assist the compre- 

 hension of your readers. 



Mr. Faris deduced times of commencement and corre- 

 sponding velocities of propagation from Horizontal Force 

 (H), Declination (D), and Vertical Force (V) curves 

 separately. On the average, the H times preceded the D 

 by 044 and the V by 23 minutes. I indicated the 

 improbability of the disturbing force commencing with a 

 component along one only of the three fundamental direc- 

 tions at a station. Normally, one must expect com- 

 ponents in all three directions ;" thus systematic differences 

 in the times of occurrence in the D. H, and V curves can 

 hardly be real. To be visible, a disturbance must attain 

 a certain amplitude. So-called " sudden commencements " 

 really take several minutes, as a rule, to reach their 

 maximum, and the apparent differences in time mean, pre- 



No. 2159. VOL. 86] 



sumably, differences in the amplitudes of the three c<.. 

 poncnts, or difference of sensitiveness in the three magn- 

 graphs. The neglect of this point of view may lead 

 serious error. To see this, take a simple case. Supp' 

 a number of similar compasses to be placed on a strai^in 

 line radiating out from an electromagnet. Suppose each 

 needle frictionless, but its displacement imperceptible uniil 

 it attains i'. Let, now, a sk>wly increasing curr 

 actuate the electromagnet. As the current rises, m<. 

 ments become visible in one compass after another, •■ 

 an unscientific observer might infer that a disturbs; 

 was being propagated outwards from the magnet with 

 velocity which might be very small if the current incre.i 

 very slowly. The time, however, when any compass 

 visibly affected is determined, not by the velocity of j> 

 pagation of electromagnetic waves, but by the sensiti\ 

 ness of the needle, its orientation, and the rate of increase 

 of the current actuating the electromagnet. Disturbances 

 of the type selected by Mr. Faris are usually largest in 

 H and least in \\ and it thus appeared to me highly 

 suggestive that the times derived from the H curves were 

 so markedly the earlier. The tendency in all cases must 

 be for the time shown to be late, the error being greater 

 the smaller the disturbance and the less sensitive the 

 magnet. If, then, on any given occasion, a disturbance 

 nowhere large were smaller in Europe than America, we 

 should expect the time of commencement shown by the 

 average European magnetograph to be a little behind that 

 shown by the average .American instrument. 



Another criticism I made was this. Originally — so far 

 as I could understand — Dr. Bauer and Mr. Faris assumed 

 differences in times at any two stations to depend only on 

 differences of longitude. This implied that the disturb- 

 ance travelled either due east or due west, appearing 

 simultaneously at all places in the same longitude. Five 

 American stations were arranged in two groups, the mean 

 longitudes (79° 9' W. and 146° 42' W.) of these being 

 regarded as belonging to two central stations 68° apart. 

 Call these two imaginary stations A and B, the former 

 being the nearer to Greenwich. It was recognised by Mr. 

 Faris that a disturbance travelling, say, westwards, might 

 pass A first, reaching B after travelling 68°, or it might 

 pass B first, reaching A after a journey of 292°. In the 

 former case it originated in the wide zone, 267°, separating 

 the extreme stations Honolulu (158° W.) and Porto Rico 

 (65° W.) ; in the latter case it originated in the narrow 

 zone, 40°, separating Baldwin (95° W.) from Sitka 

 (135° W.). The peculiarity I dwelt on was that no fewer 

 than nine out of the fifteen disturbances were treated as 

 if originating in the narrow zone, whereas one would 

 have expected only two or three. Dr. Bauer does not 

 refer to this point, but the method he now follows seems 

 different. He again takes two central stations, this time 

 one in America, the other in Europe, but he determines 

 the velocity, apparently, not from the difference of longi- 

 tude, but from the arc 75° on the connecting great circle. 

 Further, he tacitly assumes that all the disturbances com- 

 mence in the wide zone between the westmost .American 

 and eastmost European station, as all are assumed to 

 traverse the 75° arc and none the 285° arc. This may 

 or may not be a better plan than that first adopted, but 

 it is totally different. A hypothesis which makes all the 

 fifteen storms originate between Honolulu and Katharinen- 

 burg is obviously incompatible with one that makes nine 

 out of the fifteen originate between Baldwin and Sitka. 



Dr. Bauer's second method would avoid the difficulty 

 I pointed out of supposing that the velocity near the poles 

 is not merely moderate, but actually small. It has, how- 

 ever, this obvious drawback, that it cannot give a true 

 velocity at all unless we suppose the disturbances to 

 actually travel along one particular great circle connecting 

 two arbitrary points on the earth's surface. 



One of my criticisms related to inconsistencies between 

 the times shown by different stations and the conclusions 

 reached as to the direction and velocity of propagation. 

 This can be better illustrated by some data used more 

 recently by Dr. Bauer himself (Terrestrial Magnetism, 

 December, 1910, Table IV., p. 225), which are much 

 more open to criticism than those from the Coast and 

 Geodetic stations. He takes seventeen disturbances given 

 by Mr. Ellis as recorded at eight stations between 1882-9. 

 Dr. Bauer forms two groups of these stations, the .second 



