JrxE n, 1895.] 



SCIENCE. 



G75 



tilde ■? Does this again imply that the ro- 

 tation of the magnetic earth is an important 

 factor ? 



If we connect by lines all the places on 

 the earth's surface having the same de- 

 parture (with due regard to sign) fi'om the 

 values as computed from above formula' 

 we get a series of curves that converge 

 around two foci of maximum and minimum 

 departures. I have carried out this idea 

 with the aid of my collected data in the 

 case of the inclination for three epochs, 1780, 

 ISSO and 1SS5. I call the curves thus ob- 

 tained lines of equal departing inclination, 

 or. briefly, ' isapoclinios.' It is especially 

 remarkable that these lines close around 

 two points not on opposite sides of the 

 equator, but on the same side.* Their pre- 

 liminary positions are : 



Latitiicie. I.ongilude. 

 For 1885. 



North end atti-acting focus, 20°S. 40°"\V of Or. 



South end attracting fix'us, 5°S. 40°E of Or. 



For 1780. 



X. F. 0° 50°W. 



S. F. 0° 60°E. 



These positions are subject to a slight re- 

 vision. The main part, however, is brought 

 out very clearlj' in both cases, viz. : that the 

 chief cause of distortion of the prima)-y sym- 

 metrical field can he represented as due to a 

 secondary polarization apjiroximately equatorial 

 in direction. 



I then showed that tlie isapoclinics obey 

 in a remarkable degree the laws governing 

 a magnetic sj-stt'm. They do not run at 

 random. Thus, for example, the foci or 

 poles of this secondary system fall nearly 

 on the agonic lines of the actually observed 

 field, and the secondary magnetic equator 

 running roughly nortli and south marks 

 out approximately tlie places where occurs 



* Similar results have been obtained hy von Bezold 

 in the paper cited, and by A. von TUlo as seen in his 

 preliminary -paper in Comptes Rendus, Oct. 8, '91, 

 pp. 597-599. It is very much to be hoped that von 

 Tillo's charts will soon V>c pul)li.«licd. 



the maximum declination. In a word, the 

 magnetic field ivhich ive actttally observe can be 

 marly obtained by super-imposing a secondary 

 (ijuatorial field U])on a primary polar one. 



By comparing the maximum horizontal 

 intensities of the the two systems, as found 

 in the respective magnetic equators, I lind 

 that the polar field is about fire to six times 

 stronger than the secondary, and that the axis of 

 the restdtant system would make an angle of 

 about 10° with the rotation axis. 



Furthermore, the secular variation phe- 

 nomena can be qualitatively explained bj- 

 the shifting of just two such poles as be- 

 long to the secondaty system. It cannot 

 be explained by the disturbance of poles on 

 opposite sides of the equator. 



Weshould thus have to refer both thedistribution 

 and the secular variation to apparently the same 

 kind of a polarization. 



This harmonizes with the empirical con- 

 clusions at the beginning of this paper. 



Since the intersection of tlie agonic lines 

 with the equator fall so nearly together 

 with the positions of the isapoclinic foci, a 

 fair idea, perhaps, can be obtained of the 

 shifting of these foci from the motion of the 

 agonic lines along the equator. I find that 

 both agonic lines have been moving west- 

 wardly along the equator for the last 300 

 years at the average rate of about 0. °2 per an- 

 num. If the motion continues around the 

 equator at this rate the resulting period 

 would be about 2000 years, but I do not 

 wish to be understood as asserting that this 

 is the secular variation period. 



A possible third field, which has been 

 made probable by Dr. A. Schmidt's beau- 

 tiful researches, was also pointed out. 

 Schmidt found, namelj-, that not the en- 

 tire observed magnetic effect on the earth 

 can be refen-ed to a potential; currents 

 tli&t pierce the earth's surface seem to make 

 themselves felt. Perhaps his currents can 

 be explained tlius : If an arl)itrarily mag- 

 netized sphere rotates in a conducting fluid, 



