looked at from either end. It thus happens that if the rotation is determined by the direction of the twist, two molecules of the same variety will always conspire whether they are arranged as ab, cd, or, to take the extreme case, as ab, dc (with the second molecule reversed). The assumption of any particular arrangement of molecules is thus not necessary in explaining the rotation. The average effect produced by a large number of active elements interspersed in an inactive medium will thus be the same in all directions, and proportional to the number of molecules traversed by the ray. As there is no polarity in the molecule, a right-handed element will always produce the same kind of rotation, say, to the right of an observer travelling with the ray. The rotation produced when the ray is reversed by reflection will thus be in an opposite direction, and the two rotations will neutralise each other.
But if the molecules exhibit any polarity, that is to say, if the effects produced by the two ends of the same molecule are opposite, the resultant effect produced by a number of such molecules arranged in haphazard directions, will be zero. In order that the effects produced by the molecules may conspire, it is necessary that they should take up a special arrangement like the disposition of molecules in a magnetised rod. It is seen that in this case the rotations of the direct and the reflected rays are in the same direction, and the resultant rotation is therefore doubled. There is some analogy between the action of such polarised molecules and of substances which, when placed in a magnetic field, rotate the plane of polarisation.
