A most remarkable phenomenon is produced in glass tubes, under certain
circumstances. When these are laid before a fire in a horizontal
position, having their extremities properly supported, they acquire
a rotatory motion round their axis, and also a progressive motion
towards the fire, even when their supports are declining from the fire,
so that the tubes will move a little way upwards to the fire. When the
progressive motion of the tubes towards the fire is stopped by any
obstacle, their rotation still continues. When the tubes are placed in
a nearly upright posture, leaning to the right hand, the motion will be
from east to west; but if they lean to the left hand, the motion will
be from west to east; and the nearer they are placed to the upright
posture, the less will the motion be either way. If the tube be placed
horizontally on a glass plane, the fragment, for instance, of coach
window glass, instead of moving towards the fire, it will move from it,
and about its axis in a contrary direction to what it had done before;
nay, it will recede from the fire, and move a little upwards, when the
plane inclines towards the fire. These experiments succeed best with
tubes about twenty or twenty-two inches long, which have in each end a
pretty strong pin fixed in cork for their axis.