746 REPORT— 1898. 



there -would also be an annual period. If the sun were transversely magnetised 

 we should, in addition, have two periodicities, neither of which, as I have lately 

 shown, is, as commonly supposed, equal to the synodical revolution of the sun, 

 but one of which is equal to the period of sidereal revolution, and the other as 

 much larger than the synodical revolution as that is larger than the sidereal period. 



In a Presidential Address to the Eoyal SocietyC Proc. Roy. Soc.,' vol. lii. p. 305), 

 Lord Kelvin made an appeal for the investigation of those periodicities which might 

 be due to solar magnetism. But, on entering into the question, I find that it would 

 be necessary to compare carefully the daily variations in the southern and northern 

 hemispheres. At present we know too little of the movements of the magnetic 

 needle in the southern hemisphere to attack the problem with success. I should 

 consider it, therefore, a matter of the greatest importance to obtain comparisons in 

 the daily curves on stations of approximately the same latitude north and south of 

 the equator. A few temporary observations at well-chosen places would be suffi- 

 cient for this purpose. 



My investigations on the diurnal variation have led me to the conclusion that 

 part of the eifect is due to currents induced inside the earth, and the results seemed 

 to satisfy the conclusion that the interior of our globe was a better conductor than 

 the outer layers. But daily variations have also been observed in the earth 

 currents which pass close under the surface of the earth, and it is doubtful at 

 present how much of the observed daily variation is due to these earth currents. 

 As the latter must be affected considerably by the nature of the ground, as regards 

 its electric conductivity, it seems to me to be of interest to make observations at 

 places not too far apart from each other on the same circle of latitude, and to 

 compare especially the daily variations on small islands surrounded by sea with 

 those on land stations. There, again, a short series of observations will be suffi- 

 cient. The possibility of obtaining by means of the diurnal period information as 

 to the electric conductivity of the interior of the earth seems to me to justify a 

 closer investigation than the present data allow us to enter into. 



That component of a possible solar magnetisation which is parallel to the 

 earth's axis will produce a permanent effect, adding itself to the normal elements 

 of terrestrial magnetism. Similarly, any magnetic forces which revolve with the 

 earth, either because they have their origin in electric currents in the atmosphere, 

 or because they are induced in space by the revolution of the earth, will produce 

 permanent effects which can only be separated from the intraterrestrial causes by 

 the analysis of spherical harmonics. Hence attention should be given to the 

 gradual perfection of the calculations which, originated by Gauss, have recently 

 been developed by Ad. Schmidt. 



But the first step towards the improvement in the analysis by spherical har- 

 monics must be the definite answer to the question as to whether the whole, or 

 nearly the whole, of the observed forces have a potential. As far as is shown by 

 direct observations in limited regions on land, the line integral of magnetic force 

 taken round a closed line vanishes. On the other hand, if the available observa- 

 tions over the whole of the globe are collected. Ad. Schmidt has shown that a 

 quite appreciable portion does not fit in with a potential. It might be thought 

 that electric currents traverse the earth's surface chiefly over the ocean districts 

 or in the polar regions ; but it seems almost impossible to reconcile the values 

 obtained by Schmidt for these currents with the known facts of atmospheric elec- 

 tricity. On the whole, it seems most probable that our measurements of magnetic 

 forces over some of the ocean regions are affected by a systematic error. 



It seems to me that the whole science of terrestrial magnetism is almost at a 

 standstill until this point has been cleared up. 



I believe that the greatest service that could be rendered to the science of 

 terrestrial magnetism would be a magnetic survey over some line, passing as nearly 

 as may be found possible along a circle of latitude round the earth. Two lines, 

 one in the northern and one in the southern hemisphere, would be better than one ; 

 but even leaving the southern hemisphere out of account, the question raised by 

 Ad. Schmidt should be settled as soon as possible by a direct determination of the 

 integral line of magnetic force round the earth. 



