120 
a few well-selected stations. My results, as far as they go, 
point definitely to the region outside the surface of the 
earth as the locality of the periodic cause of the variation. 
It is easy to see that if electric currents parallel to the 
earth’s surface produce any disturbance, we can readily find 
out whether these currents are outside or inside the earth. 
As we pass through any current sheet the normal magnetic 
force remains continuous, but that tangential component, 
which is at right angles to the current, suffers a discontinuity 
depending on the intensity of the current. For a spherical 
current sheet these components will always be of opposite 
sign on the two sides. If we then find the distribution of 
magnetic potential on the surface of the earth from the 
horizontal components only, we should get by calculation a 
vertical component of different sign, according as the cause 
is inside or outside. A comparison with the observed values 
will at once decide the question. A more careful analysis 
is necessary, if the causes are partly outside and partly 
inside and we wish to determine their relative importance. 
I believe that few practical magneticians at the present 
day read Gauss’ memoir “ On the general theory of terres- 
trial magnetism,” and the loss which cosmical pli3^sics has 
suffered in consequence, is, as far as our generation is con- 
cerned, quite irretrievable. The memoir is a model of scien- 
tific reasoning, and full of suggestions, which are as valuable 
now as they were fifty years ago. The investigations of 
Gauss are founded on the assumption of a magnetic potential 
on the surface of the earth, and he deduces from this assump- 
tion the theorem, now well-known, that if on every place 
of the earth we know the component of magnetic force 
tending towards the west, we can determine from that 
component the direction and magnitude of total horizontal 
force, leaving only a quantity undetermined which may 
depend on the latitude, but which cannot depend on the 
longitude, and which, therefore, is easily found. It appears, 
