ock in the evening, when it becomes
nearly constant for the rest of the night. In the southern hemisphere
the direction of the movement is reversed.

When the pointing of the needle is compared with the direction of the
moon, it is found that there is a similar change. But, instead of
following the moon in its course, it goes through two periods in a day,
like the tides. When the moon is on the meridian, whether above or
below us, the effect is in one direction, while when it is rising or
setting it is in the opposite direction. In other words, there is a
complete swinging backward and forward twice in a lunar day. It might
be supposed that such an effect would be due to the moon, like the
earth, being a magnet. But were this the case there would be only one
swing back and forth during the passage of the moon from the meridian
until it came back to the meridian again. The effect would be opposite
at the rising and setting of the moon, which we have seen is not the
case. To make the explanation yet more difficult, it is found that, as
in the case of the sun, the change is opposite in the northern and
southern hemispheres and very small at the equator, where, by virtue of
any action that we can conceive of, it ought to be greatest. The
pointing is also found to change with the age of the moon and with the
season of the year. But these motions are too small to be set forth in
the present article.

There is yet another class of changes much wider than these. The
observations recorded since the time of Columbus show that, in the
course of centuries, the variation of the compass, at any one point,
changes very widely. It is well known that in 1490 the needle pointed
east of north in the Mediterranean, as well as in those portions of the
Atlantic which were then navigated. Columbus was therefore much
astonished when, on his first voyage, in mid-ocean, he found that the
deviation was reversed, and was now towards the west. It follows that a
line of no variation then passed through the Atlantic Ocean. But this
line has since been moving towards the east. About 1662 it passed the
meridian of Paris. During the two hundred and forty years which have
since elapsed, it has passed over Central Europe, and now, as we have
already said, passes through European Russia.

The existence of natural magnets composed of iron ore, and their
property of attracting iron and making it magnetic, have been known
from the remotest antiquity. But t