NOVESIBEK 11, 1904.] 



SCIENCE. 



641 



east- west component of the earth's mag- 

 netic force (Y), 0.001 C.G.S. unit, or about 

 one -fiftieth of the average value of Y. The 

 average effect on the horizontal intensity 

 is about one one-thousandth part, i. e., on 

 the order of the error of a field determina- 

 tion. However, the average effect on the 

 declination is about 0.2° — about six times 

 the error of a reduced field determination 

 of the declination on land, and about one 

 to two times the error of a determination 

 at sea by the most approved methods. 



Having given the results to be deduced 

 from a mathematical analysis of the earth 's 

 permanent magnetic field in accordance 

 with the principles laid down by Gauss, 

 let us now briefly turn our attention to 

 another mode of attack with the purpose 

 of deriving physical interpretations of the 

 various harmonic terms entering into the 

 Gaussian expression. The general title of 

 the series of the papers devoted to this 

 subject, of which the fourth number ap- 

 peared in the September issue of the jour- 

 nal Terrestrial Magnetism and Atmospheric 

 Electricity, is ' The Physical Decomposition 

 of the Earth's Permanent Magnetic Field.' 



The first harmonic finds a ready phys- 

 ical interpretation: it represents that en- 

 tire portion of the earth's total magnetiza- 

 tion which can be referred to a uniform 

 homogeneous magnetization of the earth 

 about a diameter inclined to the axis of 

 rotation. This term ■ represents about 65- 

 70 per cent, of the total field. Let us term 

 it the primary or 'normal' field. 



The diameter or axis of magnetization of 

 this field for 1885 made an angle of 11° 

 25.7' with the i-otation axis and pierced 

 the northern hemisphere in longitude 68° 

 30.6' W. of Greenwich. Its magnetic mo- 

 ment was 0.32298 B\ C.G.S. units, B being 

 the earth's mean radius. These figures 

 were dependent on Schmidt's analysis of 

 the earth's permanent magnetism, and a 



slight revision would be required in accord- 

 ance with his latest published Gaussian co- 

 efficients. However, as it was found that 

 these slight revisions are on the order of 

 error of determination, it will, therefore, 

 not be worth while at present to make any 

 change. 



In No. II. of the series of papers alluded 

 to, it was shown how the determinations of 

 the magnetic axis and of the magnetic mo- 

 ment were dependent upon the portion of 

 the earth considered in the calculations, so 

 that strictly the quantities adopted apply 

 only to the area embraced. Fortunately, 

 however, the effect of the neglected por- 

 tions of the earth— the polar regions — 

 diminishes rapidly with advancing latitude, 

 so that the values as adopted for the pri- 

 mary field, depending as they did upon 

 data from 60° N. to 60° S., will not differ 

 sufficiently from those obtained, had there 

 been data over the entire globe, to vitiate 

 the general deductions regarding the char- 

 acteristics of the 'residual' or 'secondary 

 field, ' i. e., that portion of the earth 's total 

 magnetization remaining after deducting 

 the homogeneous magnetization (the first 

 term). 



The map of this residual field has now 

 been constructed for three epochs; first, 

 for 1885 and recently also for 1842 and 

 1880, the first depending on Neumayer's 

 magnetic charts for 1885, the second on 

 Sabine's charts and the third upon Creak's 

 charts. The maps of the residual field for 

 the first two epochs agree well in all the 

 principal features with the one for 1885. 



The residual magnetization can thus be 

 broadly characterized: it consists chiefly 

 of two main magnetizations transverse to 

 the axis of rotation, one system lying in 

 the northern hemisphere, the north end 

 attracting pole (iVJ being east of the south 

 end attracting poles (jS/, S-^"), and the 

 other in the southern hemisphere, the di- 

 rection of magnetization being the reverse 



