5 oS 



NATURE 



[July 



!9i3 



series, and AAs8i2 and 5828 in the fourth series, 

 as measured by Mr. Van Maanen : — 



A graphical representation of the results of this 

 series of measures is given in Fig'. 3, in which 

 the horizontal scale is that of solar latitude, and 

 the vertical scale that of displacements. The 

 agreement of the data of the foregoing table with 

 the theoretical curve derived from the average 

 of the ordinates at 45 is shown in Fig. 3c. 



It is considered that within the limits of preci- 

 sion the observations agree satisfactorily with the 

 displacements calculated for lines originating in 

 a source on the surface of a magnetised sphere, 

 and observed from a point in or near the plane of 

 the equator. On the assumption that the field is 

 due to the rotation of a charged body, or of a 

 body composed of neutral molecules which act 

 as though they carried a charge, it is concluded 

 that the sign of the dominant solar charge is 

 negative, and that the north magnetic pole of the 

 sun lies at or near the north pole of rotation. 



The determination of the strength of the sun's 

 field has presented some difficulty because the lines 

 measured are so weak in the arc or spark that 

 their separations in a magnetic field can only be 

 determined experimentally with the greatest diffi- 

 culty, or not at all. In the case of 5920/898, the 

 procedure has been to observe the separation of 

 the line in a sun-spot the strength of field of which 

 was determined from other lines; as indicated by 

 this line, the field strength of the sun at its pole 

 is 28 gausses. The line 5812 and a nickel line 

 583 1 '82 1 (not yet fully discussed), by comparison 

 with laboratory observations, lead to the values 

 48 and 29 gausses, but for comparison with the 

 foregoing determination these should be increased 

 by about 60 per cent, on account of systematic 

 differences between the measurers. The general 

 result is to indicate that the field strength at the 

 sun's pole is of the order of 50 gausses. Various 

 higher level lines, which show large Zeeman 

 effects in the laboratory, have hitherto failed to 

 show the effect of the sun's field, and it is there- 

 fore concluded that the intensity of the general 

 field falls off very rapidly in passing upward through 

 the reversing layer, more rapidly than in the case 

 of spots. 



NO. 228l, VOL. 91] 



Prof. Hale gives a careful discussion of the 

 question as to whether the magnetic fields indi- 

 cated by the observations are due to local pheno- 

 mena or to the magnetic effect of a rotating sphere. 

 In the case of spots, the Zeeman effect frequently 

 extends beyond the penumbra, and the structure 

 of the Ha flocculi sometimes suggests that local 

 magnetic fields may also be caused by invisible 

 spots, or by whirls in which no umbras or pen- 

 umbras have appeared. There is also some evi- 

 dence that the pores which occur in all parts of 

 the sun may be small vortices which develop into 

 j spots under favourable conditions. Reasons are 

 given for believing that the line displacements 

 in question are not due to any of these causes. 

 Right and left-handed whirls about spots are 

 equally common in the northern and southern 

 hemispheres, and there is no reason to suppose 

 that they could produce the systematic displace- 

 ments, of opposite sign in the two hemispheres, 

 which have been observed. In fact, in the 

 majority of the observations no spots whatever 

 and very few calcium flocculi were visible on the 

 sun. If the pores are electric vortices, like the 

 spots, there is no reason to suppose that pores of 

 one polarity preponderate in the northern hemi- 

 sphere and those of opposite polarity in the 

 southern, unless local differences in rotational 

 velocity are sufficient to account for such small 

 vortices as the pores may represent ; this possi- 

 bility, says Prof. Hale, deserves more careful con- 

 sideration than he has yet been able to give it, 

 but even if there were a clear preponderance of 

 pores of opposite sign north and south, it would 

 be difficult to account for the curve of observed 

 displacements. 



Serious objections having been urged against nil 

 theories of terrestrial magnetism, it can scarcely 

 be hoped that any one of them can be applied with- 

 out modification to the sun, especially in view 

 of its high temperature, low density, and gaseous 

 condition. In the case of spots, neutral molecules 

 cannot produce the observed fields unless an im- 

 probable degree of separation of the positive and 

 negative electrons is assumed. Harker's experi- 

 ments, 3 however, have led Prof. Hale to suppose 

 that in a spot there must be a flow of negative 

 electrons from surrounding regions into the cooler 

 umbra, and that the whirling of these in the 

 vortex may account for the strong magnetic fields 

 observed. 



An extension of Dr. Harker's work made at 

 Pasadena by Mr. King has further shown that 

 although the ionisation current decreases as the 

 pressure increases, it is still appreciable at pres- 

 sures up to twenty atmospheres. On account of the 

 greater mobility of the negative electrons, their 

 tendency to flow towards regions of lower tem- 

 perature, and the evidence afforded by Mr. King's 

 experiments that solar ionisation is not limited 

 to the region above the photosphere, it is evident 

 that the electrical and magnetic phenomena of the 

 interior of the sun must differ radically from those 

 of the earth. But since the negative electrons 

 will tend, on the average, to lie farther from the 



s Nature, July i8, ioi 2 , p. 5 i 7 . 



