77 
of Edinburgh, Session 1875 - 76 . 
keep this rotation, and thus continue the curve at the lower end 
of the chain much farther round than if the chain was not in 
motion. And, for very evident reasons, the quickest part of the curve 
is not at the bottom, but a short distance up the ascending side; and 
farther, the rotation of the link at the bottom is quicker than that 
corresponding to the curvature. These points may all be illustrated 
by a chain in which the links are short and the chain as thick as 
possible, so that the moment of inertia of the links round an axis 
perpendicular to the plane of the motion of the links is as great as 
possible. Such a chain when properly made gives a series of large 
and well-marked waves all the way up the one side of the loop and 
down the other. The length of the links also tends to change the 
form of the curve. If we have two chains of the same material 
and same size every way, except in the length of the link, then the 
larger the link the more the chain tends to open out the curves 
and take the circular form, and the smaller the links the nearer 
it approaches the form it has while at rest, and the more marked 
the reverse curve becomes. 
An elastic band in rapid motion will also tend to take up a 
circular form, because the strain at the quick part of the curves 
will tend to open them out, in the same manner as when the band 
was at rest. An elastic chain while in motion does not show the 
reverse curve like a chain, probably because the strain on the 
material prevents it doing so. 
In the previous experiments gravitation acted on the chains, so 
that whatever form we might impress on them, gravitation con- 
stantly tended to change that form and bring it back nearly to the 
form it would have if gravitation alone acted on it. An attempt 
was therefore made to get quit of the disturbing effect of gravita- 
tion. Different ways were tried of effecting this, but none of 
them were thoroughly successful. The next experiment shows the 
most successful method tried, namely, suspension. The chain is 
hung by means of a number of fine cords to a circular disc, capable 
of rotating about a vertical axis placed as far above the chain as 
possible. The chain is driven by means of a rapidly revolving 
horizontal pulley running on a vertical axis, and to give sufficient 
friction the chain is pressed to the pulley by means of an elastic 
wheel. The centre of suspension is so arranged that it can be 
