demonstrates his opinion in this way:

[Illustration]

Let there be a round vessel filled with water: in the middle of the surface
of the water place a slender iron wire on a perfectly round cork, so that
it may just float in aequilibrium on the water; let the wire be previously
touched by a magnet, so that it may more readily show the point of
variation, the point D as it were: and let it remain on the surface for
some time. It is demonstrable that the wire together with the cork is not
moved to the side D of the vessel: which it would do if an attraction came
to the iron wire by D: and the cork would be moved out of its place. This
assertion of the Englishman, Robert Norman, is plausible and appears to do
away with attraction because the iron remains on the water not moving
about, as well in a direction toward the pole itself (if the direction be
true) as in a variation or altered direction; and it is moved about its own
centre without any transference to the edge of the vessel. But direction
does not arise from attraction, but from the disposing and turning power
which exists in the whole earth, not in the pole or in some other
attracting part of the stone, or in any mass rising above the periphery of
the true circle so that a [Illustration] variation should occur because of
the attraction of that mass. Moreover, it is the directing power of the
loadstone and iron and its natural power of turning around the centre which
cause the motion of direction, and of conformation, in which is included
also the motion of the dip. And the terrestrial pole does not attract as if
the terrene force were implanted only in the pole, for the magnetick force
exists in the whole, although it predominates and excels at the pole.
Wherefore that the cork should rest quiescent in the middle and that the
iron excited by a loadstone should not be moved toward the side of the
vessel are agreeable to and in conformity {163} with the magnetick nature,
as is demonstrated by a terrella: for an iron spike placed on the stone at
C clings on at C, and is not pulled * further away by the pole A, or by the
parts near the pole: hence it persists at D, and takes a direction toward
the pole A; nevertheless it clings on at D and dips also at D in virtue of
that turning power by which it conforms itself to the terrella: of which we
will say more in the part _On Declination_.

* * * * *

CHAP. VII.

Why the variation from