298 



NA TURE 



[August 25, 1923 



quite comparable with that of well-seasoned magnets. 

 The loss occurs practically all in the uniform portion 

 of the earth's magnetic held, parallel to a diameter, 

 inclined at present to the axis of rotation about 

 1 1 '5" ; the loss is not made up by any material gain 

 in the non-uniform, heterogeneous portion of the 

 earth's magnetic field.* The annual percentage loss 

 is nearly the same for the polar and e(iuatorial 

 components of the uniform magnetic field ; and aa 

 the equatorial component is only about one-fifth that 

 of the polar component, the absolute annual loss in 

 the earth's magnetic moment results almost entirely 

 from the polar component, i.e. the uniform portion 

 of the earth's magnetic field, symmetrical about the 

 earth's axis of rotation, suffers nearly the entire loss. 

 The system of forces which must be superposed 

 upon the uniform internal magnetic field of^ 1885 in 



order to obtain the observed field for 1922 proves 

 to be a demagnetising system, the magnetic axis of 

 which is directed almost diametrically opposite to 

 that of the primary uniform field. A similar result was 

 found in 1904* for the period 1890-1900. In brief, the 

 secular-variation system shows the characteristics 

 of the self-induced field of a uniformly magnetised 

 body. 



Let us next inquire briefly into which of the 

 quantities in formula (i) so large an annual rate of 

 change for the earth's field as 1/1500 part is to be 

 attributed ? Certainly not to the angular velocity, w, 

 or to the radius, r, or to the density, D, since changes 

 on the order of 1/1500 part in one of these quantities, 

 or in their combined product, would not escape 

 detection by other means. We must conclude that 

 the physical factor, /, contains within itself the kernel 

 for the observed change, but what does this imply ? 



Let us suppose next that in the factor / we have 

 embodied some physical relation upon which both 

 the earth's magnetic field and its gravitational field 



• Ttrr. Mag. andAtm. EUct., voL 8 (1903), p. 107, and voL 28 (1923), p. at. 



• Terr. Mag. and Attn. Elect., vol. 9 (1904), pp. 181-186. 



NO. 2808, VOL. I I 2] 



depend. Then on the basis of the large average annual 

 loss during the past eighty years in the strength of 

 the earth's magnetic field, we can immediately say 

 that magnetism and gravity are not related to each 

 other as the first power of the factor, for otherwise 

 a correspondingly large annual change in gravity 

 would likewi.se have been observed. Again, while 

 gravity b greater over the oceans than over continents, 

 the equivalent intensity of magnetisation is, on the 

 average, .somewhat less for ocean areas than for con- 

 tinental areas, so here again there is no immediate 

 relation between gravity and magnetism. According 

 to Sutherland's theory" which was based on a slight 

 modification of the laws of electrodynamics, magnetism 

 would depend on the first power of a small fraction 

 ^ (about 2-6 X 10-"), and gravity upon the second 

 power : this quantity /i would enter into the factor/ of 

 formula (i). Accordingly the annual 

 decrease of 1/1500 part in magnetism 

 would imply, on Sutherland's theory, 

 only a decrease of the square of 1/1560 

 part, or about one-half of a millionth 

 part in gravity, and this is a quantity 

 which may readily escape detection 

 with our present gravity appliances, 

 unless the accumulative effect over 

 many years be carefully observed at 

 several standard stations. Hence, a 

 theory involving gravity and magnet- 

 ism in the manner prescribed by 

 Sutherland's hypothesis might be ad- 

 missible. But the observed decrease 

 in the earth's magnetic field-strength 

 would then have to be referred to a 

 corresponding change in /*. But what 

 makes /S change ? It was only meant 

 to represent a very slight variation in 

 the law of action between electric 

 charges ; if ii changes, so must the new 

 assumed law of electro-dynamics. We 

 have under investigation various hypo- 

 theses to account for the observed secu- 

 larchanges in the earth's magnetic field. 

 t»» Sufficient has been given to show 

 with what extreme care a theory of 

 the earth's magnetic field will have to 

 be formulated and how exhaustively it 

 will have to be examined in the light 

 of the data now known to us. No one 

 who will familiarise himself with the 

 facts will lightly announce the dis- 

 covery of a new theory of the origin of 

 the earth's magnetism. New and inter- 

 esting matters may confidently be expected from the 

 discovery of the true cause. 



In conclusion. Fig. 2 is presented to show the 

 positions of the following points : MA(I). north end 

 of magnetic axis of the earth's uniform internal 

 magnetic field in 1922, latitude 78° 32' N., and 

 longitude 69° 08' W. ; ]SIA(E), north end of magnetic 

 axis of the earth's uniform external magnetic field 

 in 1922, latitude 768° N. and 121-4° ^^ ■ '< and N.M.P., 

 the approximate position of the North Magnetic Pole 

 in 1904, latitude 70-5° N., and longitude 95"5° ^^■ 

 As will be seen, the line of maximum auroral frequency 

 passes to the south of the three jxjsitions. (The 

 other lines shown are the routes of the Carnegie.) 

 It will be noticed that the displacement of the E-axis 

 is about 52° west of that of the I-axis, and that the 

 N.M.P. is about midway in longitude between I and E. 

 From the amount and direction of displacement of 

 the E-axis with reference to the I-axis, we may deduce 

 further important facts bearing upon the theory' of 

 the earth's magnetic field and the possible conductivity 

 of interplanetary' space. 



'• Terr. Mag. and Atm. Ekct., voL 9 (1904), pp. 167-172. 



