28 SECTIONAL ADDRESSES. 



parallels of latitude as circuits of the earth, the charts used being those 

 of Neumayer's. A second investigation was made in 1904, and a third in 

 1908, the data used on the latter occasion being obtained from the magnetic 

 charts for the United States, and the line integrals calculated for areas 

 within the United States. In all cases vertical currents were indicated. 

 In 1920 a further investigation was made, and from the results it appears 

 that in certain areas there are upward electric currents, and in others 

 downward currents, the current density varying from +58x10"'' ampere 

 per square kilometre to — 53x10"^ ampere per square kilometre. It 

 appears very desirable that a definite answer should be obtained to this 

 question of vertical currents, but calculations show that to obtain a 

 decision it is necessary for measurements of the most precise kind to be 

 made. I have studied the calculations of Bauer, and conclude that, while 

 he is justified in his deductions, the data are not sufficiently reliable. The 

 values of the horizontal force are not all observed values, and those from 

 which they have been deduced have been obtained by many observers, 

 by many instruments, and under very varying conditions, so that the 

 probable errors are not small. My review leaves me exceedingly doubtful 

 of the existence of such currents. 



To illustrate the kind of precision required in the measurements, 

 imagine a very large circular area of radius r and assume that a vertical 

 current i is distributed uniformly over the area. The horizontal intensity 

 produced by such a current at a distance d from the centre of the circle is 

 2idjr'. If the current density is 30 X 10"'^ ampere per square kilometre, 

 the value of i is 27071 C.G.S. units of current, and the horizontal force 

 produced at 100 kilometres from the centre of the area is a little less than 

 2y, i.e. 0.00002 C.G.S. unit. The line integral round such an area would 

 be considerable, but it is easy to see that if there were no vertical currents 

 and measurements for one half of the circle were made with one instrument 

 giving correct results while those of the other half were made with a 

 second instrument giving values in error on an average of +4y, the 

 conclusion drawn would be that vertical currents of the values already 

 assumed really existed. 



Notwithstanding the difficulties, there is no doubt that sufficiently 

 precise measurements could be made over a carefully chosen area, which 

 would enable a definite decision to be reached with respect to such vertical 

 currents. The experiments would necessitate not only the use of similar 

 types of instruments but the establishment of small temporary stations 

 with magnetographs so that simultaneous values of the horizontal force 

 could be obtained. 



Variations of the Magnetic Field. 



Usually, when the cause of a large scale phenomenon like the earth's 

 magnetism is unknown, it is useless to look for the cause of the variations, ( 

 but in the case of terrestrial magnetism this is not so. Schuster showed 

 that the variations are not due to the same causes as those which produce 

 the main field, but to external causes. Some of the variations are 

 periodic in character and appropriate names such as daily, monthly and 

 annual variations have been applied, while others of a violent and non- 

 recurrent character are called magnetic storms. 



