22 
land. Changes take place in the amount of aqueous vapour 
held by the air, and also probably in the molecular state of 
this aqueous vapour. But although we may not be able to 
point to the specific actions which produce electrical sepa- 
ration, we know that such separation implies a one-sidedness 
or heterogeneity; and since gravity will presumably act 
differently on the two things, we may probably suppose 
that one of the constituents which have caused this electrical 
separation may have a tendency to mount upwards in the 
atmosphere, while the other may have a tendency to move 
downwards. For instance, ir evaporation from the surface 
of the earth or sea be one cause of this electrical separation, 
we might imagine the land or sea to become electrified in 
one way, while the vapour electrified in the other direction 
might mount in the air, owing to its being specifically lighter. 
In fine, whatever be the cause of the electrical separation, 
we may presumably suppose that the one constituent will 
either remain below or find its way downwards, while the 
other, carrying with it its peculiar electricity, will mount 
upwards. 
Now, may not the earth be regarded as a Leyden Jar, the 
sea and earth forming one coating, and the upper, rarer, and 
hence electrically conducting strata of air forming the other 
coating ; and will not the tendency of the action above named 
be to charge the upper coating with one kind of electricity 
and the under with another ? 
Such a process would of course be continually going on, 
while on the other hand the earth, regarded as a Leyden Jar, 
would by means of thunderstorms, and possibly by other 
means, be continually discharging itself. 
Next, let us suppose that by some extensive local circum- 
stances a greater than usual electrical separation and charging 
of the Earth Jar has been going on. The effect of this local 
cause would, however, not be local, but would contribute to 
increase the charge of the earth as a whole — as one jar in 
act, so that the earth as a whole might, for a short period. 
