DEPARTMENT OF TERRESTRIAL MAGNETISM. 327 



which are shown by the Earth's magnetism, pecuHar to the Earth, or are they 

 effects from forces bound up with the direction and speed of rotation of a re- 

 volving body? If the latter be true, then the Sun's magnetic field may also be 

 found to be a complex one with its magnetic poles not simply non-coiacident with 

 its poles of rotation, but, as in the case of the Earth, not even diametrically 

 opposite. A preliminary analysis of Hale's published observations seemed to 

 indicate that the Sun's field may be as complex as that of the Earth. However, 

 the method of mapping a magnetic field by observing the Zeeman effect in the 

 glowing Sun's vapors is a much more difficult matter than mapping the Earth's 

 field by means of instruments the results of which can be definitely interpreted. 

 Accordingly, it will be well to defer drawing definite conclusions and confine 

 our studies at present mainly to the accumulated data for the Earth's field. 



Several years ago the author found that the constant of the Earth's magnetic 

 field proportional or equivalent to the quantity termed ''intensity of magneti- 

 zation, " varied in a systematic manner from parallel to parallel, increasing 

 with approach towards the equator. The average value of the constant, 

 expressed in c. g. s. units, for the parallels 60° north and south, was 0.070, 

 whereas for the equator it was 0.082, hence an increase of about 17 per cent on 

 the former value. The conclusion reached was as follows: That if a law be 

 formulated involving a constant term supplemented by one proportional to 

 the square of the sine of the co-latitude, the observed increase from parallel 

 to parallel towards the equator is closely represented. It will be noted that 

 this second term would thus vary in the same way as the radial component of 

 the Earth's centrifugal force. Furthermore, the law stated is similar, as will 

 be recalled, to that of Faye, which represents the variation in the angular 

 velocity, from parallel to parallel, of points on the Sun's surface. The law 

 was found to give a representation of the observed magnetic facts on the Earth 

 to within 1 per cent. 



A re-investigation of this interesting matter has been carried out, and it has 

 been found that the constant referred to above shows the systematic increase 

 towards the equator quite generally over the entire Earth. The effect appears 

 to be too large, as judged from laborator}^ facts, to be accounted for chiefly by 

 some centrifugal action resulting from the Earth's rotation. The matter is, 

 therefore, being examined into from various standpoints. There is some 

 indication that the distribution of land and water plays an important part. 



Intercomparisons of the standard instruments at magnetic observatories, 1905-1914. L. A. 

 Baner and J. A. Fleming. (Presented before the American Physical Society, Wash- 

 ington meeting, April 1915. The data and results on which this paper was based are 

 published in full in Publication No. 175, vol. ii, 211-278, Carnegie Institution of 

 Washington. Cf. pp. 321-322.) 



For a world-wide magnetic survey, such as that undertaken by the Depart- 

 ment of Terrestrial Magnetism of the Carnegie Institution of Washington, it is 

 manifestly of the highest importance to know the interrelationship of the 

 various magnetic instruments employed both by the observers of the Carnegie 

 Institution of Washington and of the cooperating organizations. It has been 

 found repeatedly that magnetic instruments, though constructed with the 

 utmost care, may, nevertheless, give values of the magnetic elements which 

 differ, for the various instruments employed, by amounts considerably larger 

 than the purely observational error with any one instrument. The instru- 

 mental differences shown, for example, by the very best dip circles, may reach 

 a magnitude 5 to 10 times the observational error. Indeed, cases of dip circles 

 have been found which required a correction, on some carefully determined 

 standard, by an amount equal to the secular change for 10 years and more. 

 Unfortunately, the correction is generally not a constant one, but a function 



