A.— JIATHEMATICAL AND PHYSICAL SCIENCES. 21 



constitution of the earth's interior, the difficulty has to be faced that at 

 high temperatures iron loses its magnetic properties, and that increase of 

 pressure depresses rather than raises the critical point. As the interior 

 of the earth is at a high temperature the theory therefore loses support, 

 unless it is assumed that at exceptionally high pressures and temperatures 

 there is some restoration of the magnetic properties. 



If the earth is taken to be a uniformly magnetised sphere its magnetic 



moment, as already stated, is given by - Trrl where I is the intensity of 



magnetisation. The effect is that of a magnetic doublet and could be 

 produced equally well by a spherical shell of thickness (r — a) in which 

 case if a, the thickness of the shell is 10 miles, the intensity of magnetisa- 

 tion would be about 50 to produce the magnetic forces observed. In a 

 shell of such thickness the temperatures and pressures would not be so 

 great as to destroy the magnetic properties of iron. 



While the theory of an iron core or an outer shell permanently 

 magnetised has little to support it there is little doubt that the 

 irregularities in the earth's field are largely related to the structure of the 

 «arth's crust, and large masses of iron and iron ore must play an important 

 part in the irregularities which are observed. Geophysical surveys by 

 magnetic methods support this view. 



Electric Currents circulating round the Earth. 



The next simplest theory is that the magnetic field is due to electric 

 currents circulating round the earth, and this naturally gives rise to the 

 question of the seat of origin of the electromotive forces necessary to 

 maintain such currents. If the currents are uniform in density throughout 

 the volume of the earth the magnitude of this density would be about 

 10""* ampere to produce the necessary intensity of magnetisation. If we 

 suppose that there was once a source of electromotive force but it has 

 long ceased to operate, the currents produced would take a very lono- time 

 to die down owing to self-induction. 



In the case of a copper sphere of the size of the earth, Lamb has shown 

 that three million years after such currents were generated and the cause 

 removed, they would still have one-third of their original intensity. The 

 electrical conductivity of the earth is, however, much less than that of 

 copper, and if remnant currents are responsible for the earth's present 

 magnetism, their value in the past must have been very great, and the 

 cause of their origin would still be a mystery. But it is much more 

 profitable to look for a possible electromotive force not only to produce 

 but permanently to maintain a current system. Such a possible source 

 was indicated by Larmor at a meeting of this Section of the British 

 Association in 1919. Larmor pointed out that in the case of the sun, 

 surface phenomena indicated the existence of a residual internal circula- 

 tion mainly in meridian planes. If this circulating conducting material 

 cuts a magnetic field which in direction is the same as that of the earth 

 circulating currents would be set up in such a direction as to augment the 

 magnetic field and eventually a condition of equilibrium would be set up 

 between the producing electromotive force and the attenuation effects. 

 The system is, in fact, that of a self-exciting dynamo, and the energy of 



