118 
Or if we prefer to render the analogy between the ineteo- 
rological and magnetical systems more verbally complete we 
should say ascending negative currents at the equator, and 
descending negative currents at the pole. 
These vertical currents, being supposed to be confined to 
the upper regions of the atmosphere, we might imagine that 
they ought to render themselves visible at the magnetic pole, 
where they are most concentrated. If so they would appear 
as a luminous vertical curtain or fringe suspended in mid-air. 
This at once suggests to us that the well-known form and 
nearly continuous appearance of the aurora in these regions 
may be due to this cause, and may represent to us the vertical 
component of these currents, which we have here supposed 
to be the causes of the solar diurnal magnetic variations. 
It must not, however, be supposed that in making this 
suggestion we imply that phenomena of an auroral nature 
are not likewise connected with magnetic disturbances. 
It is to be remarked, in conclusion, that a system of 
atmospheric currents wiU act inductively on the terrestrial 
magnetic system, so that the final effect on the needle will 
be the conjoint effect of the currents above and of the mag- 
netic change below. In the case of the declination it is our 
inability to express the force that acts near the equator, or 
near the magnetic pole, in terms of any conceivable general 
change in the magnetic system that induces us to look to 
atmospheric currents as affording us a simpler mode of ex- 
pressing observed facts. This, however, does not hold for 
the horizontal force near the equator. A set of currents 
moving east in both hemispheres will produce, by induction, 
a definite and well-understood effect upon the terrestrial 
magnetic system. We do not, therefore, know how far the 
change produced by the sun upon this element is due to a 
cause above the needle, or how much to magnetic change 
below, and in this respect the conclusions we have deduced 
may require modification. 
