August ii, 1910] 



NATURE 



193 



that by far the greater part of a disturbance is due to 

 upper electric currents, tliough quite likely, in a subsequent 

 volume, he will consider the subterranean currents also. 

 Since the observed quantities actually to be operated with 

 appear to be resultant effects of both external and internal 

 forces, it is very desirable that we should know just what 

 proportion must be referred to one cause or the other. For 

 this separation we require, however, a knowledge of the 

 disturbances in the vertical intensity, and these are either 

 difficult to determine with sufficient accuracy or are not to 

 be had always at a sufficient number of stations, so that 

 Birkeland was perforce cotiipelled in his first treatment to 

 assume chiefly e.xternal currents. 



It was for these reasons deemed desirable to make 

 known promptly the deductions derived from the mathe- 

 matical analysis of certain typical cases of the class of " S " 

 storms. We have now the means of applying the first 

 decisive tests as to how far the Birkeland-Stdrmer theory 

 will account for the facts. 



There is a distinct advantage in treating, for the present, 

 simply these sudden beginnings of magnetic disturbances 

 for the reason that not only the time of beginning can be 

 sharply determined, but, what is equally important, the 

 actual magnitude and direction of the disturbance effect 

 on any particular element can be most accurately deter- 

 mined. As the effects we are here especially considering 

 do not extend, in general, over five minutes, we mav 

 readily scale off on the magnetogram the disturbance effect, 

 being, without essential error, simply the difference 

 between the ordinate to the curve at the point of beginning 

 of the disturbance and the ordinate to the particular point 

 of the disturbance considered. When, however, a mag- 

 netic storm extends over many hours, and even days, and 

 one wishes to know the magnitude and direction of the 

 disturbance at stated times, for example, every hour or 

 haif-hour, then what is called a " normal curve " must 

 be drawn from which the disturbance ordinate is to be 

 measured. This " normal curve " is supposed to represent 

 the curve of magnetic variations which would have resulted 

 had there been no disturbance ; but to determine such a 

 curve is far more difficult than is generally realised, and 

 usually an arbitrary assumption of some' kind must be 

 made to derive it. Of such assumptions we are practically 

 free in the disturbances considered. 



Application of Tests. 



The first fact of importance found from the analysis of 

 the disturbance of May S, 1902, was that the direction of 

 flow of the negative electric currents, which could account 

 for the external and internal magnetic disturbance systems, 

 was the same for both systems. While the strength of 

 the external system was about three times that of the 

 internal, nevertheless, the internal currents were not the 

 direct consequence of the outside moving negative charces, 

 i.e. they were not induced currents. Instead, for both 

 systems — outside and inside the earth— the flow of elec- 

 tricity wa.s eastwardly around the earth for negative charges 

 and westwardiy for positive charges. 



Having fixed the direction of flow of negative electricity, 

 let us inquire now whether kathode rays coming from the 

 sun will give the required direction. Birkeland, in his 

 experiments on a magnetised terrella when placed in a 

 Crookes's tube and subjected to a bombardment of kathode 

 rays, observed, among other interesting phenomena, the 

 formation, under certain conditions, of a ring of kathode 

 particles which encircled the terrella in the magnetic 

 equatorial regions. For an unmagnetised terrella there was 

 no such ring. Stormer, from his mathematical investi- 

 gations, found under what conditions a similar ring would 

 be formed when kathode rays from the sun came within 

 the deflecting influence of the earth's magnetic field. The 

 ring results when electric charges enter a magnetic field 

 perpendicularly to the lines of magnetic force, e.g. in the 

 magnetic equatorial regions. It was on account of the 

 possible formation of such an equatorial ring that Birkeland 

 was apparently led to the .-idoption of his term " equatorial 

 magnetic perturbation,", and to refer its cause to such a 

 ring. 



If we apply, however, the well-known law according to 

 which a negative charge would be deflected if entering the 

 earth's magnetic field from without at right angles to the 



NO. 2128, VOL. 84] 



lines of magnetic force, it is unfortunately found that the 

 deflection is to the west, and the moving negative electric 

 charges would accordingly encircle the earth from east to 

 west, hence opposite to what our mathematical analysis of 

 the disturbance of May 8, 1902, and January 26, 1903, 

 have shown must be the case to account for the observed 

 disturbances. 



Were we to assume, on the other hand, that the cor- 

 puscles are shot up into the earth's field instead of down- 

 wards, then those which struck the lines of magnetic force 

 perpendicularly would, after successive deflections, circulate 

 around the earth from west to east or eastwardly, and hence 

 harmonise with the observed facts. Thus far, then, we 

 should have to conclude that if the disturbances considered 

 are to be referred to kathode rays deflected by the earth's 

 field into more or less circular paths, the source of the 

 kathode rays would have to be within the earth itself, and 

 not without. 



But if the radius of the ring of moving corpuscles is 

 computed to conform with the time of propagation of the 

 disturbance around the earth (about 3I minutes), it is 

 found that the orbit would have to be distant from the 

 earth's centre 580 times the earth's radius, or 3,700,000 

 kilometres, or 2,300,000 miles, and thus the possibility of 

 a terrestrial origin of the kathode rays is likewise 

 eliminated. Furthermore, if we calculate the intensity of 

 the current which at that distance could produce the 

 observed effects of the disturbances of May 8, 1902, and 

 January 26, 1903, it is found to be 5,900,000 amperes. 

 Now Birkeland says on p. 311 of his book : — " In the case 

 of the greater storms, we found current-strengths that 

 varied between 500,000 and 1,000,000 amperes, or even con- 

 siderably more." Hence, to produce the comparatively in- 

 significant magnetic disturbance effects here considered, by 

 supposing a band of kathode particles circulating around 

 the earth, would require a current at least six times stronger 

 than that which Birkeland finds sufficient to account for 

 the much larger storm effects ! 



The hypothesis was next briefly examined on which the 

 disturbance effects considered might be referred to altera- 

 tions in the electrical conductivity of the atmosphere and 

 of the earth, either brought about by the secondary effects 

 from bombarding kathode particles, viz. the formation of 

 Rontgen rays, or, say, by the entrance into the earth's field 

 of the penetrating radiation (7 rays of radium). The 

 ionising effect and resultant alteration of electrical con- 

 ductivity of the regions involved might either be due to 

 the penetrating radiation from the sun or from the earth, if 

 only qualitative results are considered. It is therefore at 

 present not possible to state definitely whether the initial 

 cause of the disturbance of May 8, 1902, was due to a 

 terrestrial eruption or a solar one. First, further examina- 

 tions will have to be made of the disturbances of May 20 and 

 July 9, 1902, which were again closely coincident with the 

 Mont Pelte eruptions. The electric-conduction hypothesis 

 appears to satisfy, in general, the observed phenomena, and 

 accordingly it is to be subjected to a further rigid examina- 

 tion. It seems also to explain why some of the disturb- 

 ances take a westward path, although the majority of them 

 go eastward. 



Were we to suppose that the generated currents lie on 

 a sphere of radius approximately equal to that of the earth, 

 a velocity of the moving negative charges of 180 kilometres, 

 or no miles per second, results, hence a quantity of the 

 order of that for metallic conduction, or as found for the 

 kathode rays from glowing electrolytes. 



In conclusion, it should be stated that while it has been 

 shown that the class of simple disturbances discussed in 

 this paper cannot be referred to kathode rays in the way 

 Birkeland and Stormer have supposed, it should be dis- 

 tinctly understood that this in no wise vitiates other por- 

 tions of their theory, especially with reference to the larger 

 and more complicated magnetic disturbances and to the 

 origin and formation of polar lights. Before anything 

 definite can be said as to the validity of these portions of 

 their theory, it will be necessary to await the completion 

 of a similar analytical treatment to that made for the " S 

 disturbances. 



Such an analytical examination the writer has had under 

 way for more than a year, and a preliminary statement of 

 results was made at the meetings of the American Philo- 



