A.— MATHEMATICAL AMJ PHYSICAL SClEiNCES. 25 



an extent comparable with that of the earth, but apart from this restriction 

 with regard to outer space, the sun's magnetic field is strikingly similar to 

 that of the earth, for its direction of magnetisation is the same and the 

 magnetic axis is inclined to the sun's axis of rotation at about 4°. More- 

 over, the magnetic axis is not fixed in position relative to the axis of rota- 

 tion, but revolves slowly round it, so that, whereas the period of rotation 

 of the solar axis is about 31 days, that of the magnetic axis is about 

 26 days. The intensity of the field at the poles is about 50 Gauss. 



These observed similarities between the magnetic fields of the earth 

 and the sun, especially as the physical conditions are so different, naturally 

 lend support to the theory that the magnetisation is brought about by 

 rotation, and the fact that the axes of rotation and magnetisation do not 

 coincide, while disturbing, may possibly be explained by reasonable 

 assumptions. 



If rotation of matter is necessary to produce the magnetic fields of the 

 earth and the sun, the angular velocity, the radius, and the density must 

 be important factors. If the magnetic effect is proportional to Dor" 

 where D is the density, the calculated intensity of the sun's field agrees 

 with that observed, taking the earth's field as the standard. Unfortunately, 

 owing to the square of the radius being involved in the expression for the 

 field, an effect proportional to Dcor"^ cannot be tested by experiments in 

 the laboratory, as a value of o) necessary to produce a measurable effect 

 could not be obtained. A magnetic effect proportional to Dor can be 

 and has been tested in the laboratory, but the effect is far too small to 

 account for the earth's magnetism. 



There are a number of ways in which small magnetic fields may arise 

 in a spherical rotating conductor. For example, centrifugal force may 

 result in the free electrons moving towards the surface until equilibrium 

 is brought about by the resulting electrostatic forces. As an alternative 

 to this, gravity may pull the electrons towards the centre of the sphere 

 until again the resulting electrostatic forces restore equilibrium. In the 

 first case Swann has shown that the horizontal intensity would change 

 sign as an observer travels from the equator to the pole. In the second 

 case the magnetic field at the poles would be of the reverse sign to those 

 of the earth's field. 



A theory which has been tested by laboratory experiments is one 

 depending on gjToscopic action. If the magnetic condition of iron arises 

 from the rotation of the electrons in the constituent atoms, the axes of 

 rotation should tend to become parallel to the earth's axis of rotation. 

 Only a slight change in orientation can be expected because of the forces 

 due to adjacent molecules, but the net result must be to cause each 

 molecule to contribute a minute magnetic moment parallel to the earth's 

 axis of rotation. When a steady state has resulted there will be an angle 6 

 between the two axes, and the axis of rotation will prccess, i.e. it will 

 trace out a cone. The net result so far as the magnetic effect is concerned 

 is to cause each molecule to contribute a minute magnetic moment parallel 

 to the earth's axis of rotation. The effect will be proportional to the 

 angular velocity and not the radius, so that the effect can easily be tested 

 in the laboratory. Barnett first succeeded by laboratory experiments in 

 showing that magnetisation was produced by rotation, and that the 



