396 



NA TURE 



[September 30, 1909 



range, 740 7, actually recorded, represent merely the 

 full width of the photographic paper. How much 

 these rantjes were exceeded it is impossible to say, 

 but, judging by the look of the curves, the excess was 

 probably considerable. The vertical-force trace got off 

 the sheet only on one side, and this element would 

 appear to have been less disturbed than the other two. 

 Still, as the trace was off the sheet continuously for 

 nearly an hour after 3.35 p.m., the chances are 

 that the true range exceeded somewhat largely the 

 range 530 7 actually recorded. The duration of the 

 storm was comparatively short, but whilst it lasted it 

 exhibited an energy which has been very seldom 

 rivalled at Kew. The oscillatory movements were 

 quite as rapid as those of October 31, 1903, and the 

 range of the elements has probably not been exceeded 

 during the last twenty years, not even during the great 

 storm of Febiuary 13, 1892. 



Magnetic storms such as the present inevitablv 

 create an interest in the explanations that have been 

 advanced to account for the phenomenon. The 

 theories of Arrhenius and of Nordmann, the theories 

 and researches of Birlceland, and the deductions made 

 by .Maunder from the Greenwich disturbances all point 

 to the sun as the ultimate source, and to some form 

 of discharge — ions, electrons, or such like carriers of 

 electricity — as the immediate vehicle. The electrical 

 nature of aurora is difficult to dispute, and the 

 fact that storms like the present appear to be in- 

 variably associated with aurora visible far outside 

 the polar regions unquestionably supports in some 

 ways theories such as those of Birkeland or 

 Arrhenius. 



When we come, however, to details, difliculties 

 present themselves. If magnetic storms are directlv 

 due to the electrical currents which render the upper 

 atmosphere luminous, how comes it to pass that the 

 visual phenomena of aurora are so constantlv chang- 

 ing, whilst even in the most conspicuously variable of 

 magnetic storms the larger movements of the magnets 

 take usually 5, 10, or 20 minutes to accom- 

 plish, the force appearing to alter at a nearlv 

 uniform rate for minutes on end? The relatively 

 .gradual nature of the magnetic change is a true pheno- 

 menon — as clearly indicated bv the short-period 

 magnets of the Eschenhagen pattern, as in the larger 

 Kew magnets with periods of 10 seconds or more. 

 There is, again, the very remarkable fact that when 

 we go to high latitudes, where aurora and magnetic 

 disturbance are both almost daily occurrences, the 

 association of the two phenomena becomes much 

 more difficult, if not impossible, to recognise. The 

 absence of visible aurora during active magnetic dis- 

 turbances may be reasonably accounted for during 

 the Arctic summer, when the sun is above the horizon, 

 but it is a different matter when we find the magnets 

 rather quieter than usual during the occurrence of a 

 bright aurora. Unless we are to assume a funda- 

 mental difference of type between auroras presenting 

 the same spectroscopic lines, or a varietv of sources 

 for different magnetic storms, there is a difficultv 

 which is not easily surmounted. The onlv explanation 

 that has occurred to me is the possibilitv that the visual 

 phenomena may represent merelv intense local con- 

 centration of electrical current, and that the main 

 portion of the discharge frequentlv makes no appeal 

 to the eye, and is of a much more steady and persistent 

 character. Another difficulty in regarding the pheno- 

 mena of magnetic storms as entirelv and directly due 

 to the action of electrical currents associated with 

 aurora is that it is a frequent occurrence — as on the 

 present occasion — for the horizontal force to be con- 

 siderably depressed below the normal value when the 

 • Storm has apparently ceased and for some considerable 

 NO. 2083, VOL. Si] 



time thereafter. It is possible, of course, that the 

 external currents have partly demagnetised the earth, 

 or at least modified its distribution of magnetism, 

 and that there are recuperative tendencies tending to 

 cause reversion to what is for the time being- a more 

 stable distribution; but if this be the true explanation, 

 the demagnetising action and the recuperative ten- 

 dencies are presumably in action during the course 

 of the storm, and profoundly modify the magnetic 

 phenomena. To many minds subscription to some 

 theory may be a necessity for intellectual comfort, 

 but in the case of magnetic storms reservation of 

 judgment appears at present the more scientific 

 attitude. 



In addition to the foregoing we have re- 

 ceived the following communication from Prof. 

 X. Fowler, of the Imperial College of Science 

 and Technology, South Kensington : — The pos- 

 sible occurrence of a magnetic storm and auroral 

 display on September 24 or 25 was suggested 

 by observations of the large spot which was 

 then on the sun's disc. On September 24 the spot was 

 a little west of the central meridian — vvhich appears to 

 be the most favourable position in relation to magnetic 

 disturbances — and spectroscopic observations showed 

 that it was of the same disturbed type as the spot 

 associated with the great magnetic storm of October 

 31, 1903 (X.viURE, vol. Ixix., p. 6). 



On Friday evening (September 24) the sky was over- 

 cast, and it did not then occur to me to test the 

 possible presence of aurora by the spectroscope. On 

 Saturday evening, however, although the sky was at 

 first completely clouded over, the spectroscope gave 

 unmistakable evidence that an auroral display was in 

 progress. From 6.40 to about 7.30 (the sun set at 

 5.52), the whole sky was filled with a feeble light, with 

 brighter patches here and there, and the characteristic 

 green line of the auroral spectrum was seen in every 

 direction. The greatest intensity was at first near the 

 zenith, but the line was easily visible over the entire 

 sky, and was even seen in the light reflected by a 

 pocket handkerchief. This condition continued with 

 diminished brightness until near 8 p.m. Between 8 

 and 9 o'clock the display was very feeble, but shortly 

 after 9 the auroral line was again faiily distinct in a 

 faintly luminous belt about 10° above the northern 

 horizon. .After 9.30 no evidence of aurora was 

 obtained, although the sky was then p.-irtially clear. 



The general distribution of the green line over the 

 heavens in clearer skies has been occasionally noted 

 bv Angstrom and others, but I have not yet found any 

 previous record of such a wide diffusion of the auroral 

 light when the sky was completely clouded. If wholly 

 above the clouds, the aurora must have been of 

 extraordinary brightness in order to produce this 

 effect. 



Besides the green line, there were three fainter 

 nebulous lines or bands in the green and blue, which 

 have been frequently mentioned by previous observers. 

 A careful search was made for the red line which 

 appears in " crimson " aurorae, but its presence was 

 not even suspected. 



As to the sun-spot, there was a brilliant re- 

 versal of the C line of hydrogen over one of 

 the umbrae when I observed it at 12.20 p.m. 

 on .September 24, and on opening the slit it 

 was clear that this appearance was produced by 

 a very bright overlying prominence. Reversals of 

 the chromospheric lines ti,, and 1474 K were also sus- 

 pected, but the observations were stopped by clouds. 

 .According to Tacchini and Lockver, it is the promin- 

 ence, rather than the spot, which should be considered 

 as related to the magnetic disturbance. 



