[ 163 ] 
IV. The Diurnal Variation of Terrestrial Magnetism. 
By Arthur Schuster, F.R.S. 
Received October 31,—Piead November 7, 1907. 
1. In a previous communication^ I proved that the Diurnal Variation of Terrestrial 
Magnetism has its origin outside the earth’s surface and drew the natural conclusion 
that it was caused by electric currents circulating in the upper regions of the 
atmosphere. If we endeavour to carry the investigation a step further and enquire 
into the probable origin of these currents, we have at present no alternative to the 
theory first proposed by Balfour Stewart that the necessary electromotive forces 
are supplied by the permanent forces of terrestrial magnetism acting on the bodily 
motion of masses of conducting air which cut through its lines of force. In the 
language of modern electrodynamics the periodic magnetic disturbance is due to 
Foucault currents induced in an oscillating atmosphere by the vertical magnetic 
force. The problem to be solved in the first instance is the specification of the 
internal motion of a conducting shell of air, which shall, under the action of given 
magnetic forces, determine the electric currents producing known electromagnetic 
effects. Treating the diurnal and semidiurnal variations separately, the calculation 
leads to the interesting results that each of them is caused by an oscillation of the 
atmosphere which is of the same nature as that which causes the diurnal changes of 
barometric pressure. The phases of the barometric and magnetic oscillations agree 
to about If- hours, and it is doubtful whether this difference may not be due to 
uncertainties in the experimental data. In the previous communication referred to 
I already tentatively suggested a connexion between the barometric and magnetic 
changes, but it is only recently that I have examined the matter more closely. In 
the investigation which follows I begin by considering the possibility that both 
variations are due to one and the same general oscillation of the atmosphere. The 
problem is then absolutely determined if the barometric change is known, and we 
may calculate within certain limits the conducting power of the air which is sufficient 
and necessary to produce the observed magnetic effects ; this conducting power is 
found to be considerable. It is to be observed, however, that the electric currents 
producing the magnetic variations circulate only in the upper layers of the 
atmosphere, where the pressure is too small to affect the barometer; the two 
* ‘Phil. Trans.,’ A, vol. 180, p. 467 (1889). 
VOL. CCVIII.—A 430. Y 2 22.4.08 
