Memoirs presented to the Cambridge philosophical society on the occasion of the jubilee of Sir George Gabriel Stokes, bart., Hon. LL. D., Hon. SC. D., Lucasian professor.

AND THE INFLUENCE OF A MAGNETIC FIELD ON OPTICAL PHENOMENA. 393
be anticipated that the equation of the free periods would involve the orientation
of the molecule with regard to the magnetic field. But if that were so, these
periods would not be definite, and instead of a sharp magnetic resolution of each
optical line there would be only broadening with the same general features of polarization. To that extent the phenomenon was in fact anticipated frorn theory, except as
regards its magnitude. The definite resolution of the lines is however an addition to
what would have been predicted on an adequate theory, and thus furnishes a clue
towards molecular structure.
A POSSIBLE ORIGIN OF SERIES OF DOUBLE LINES.
The definiteness and constancy in the mode of decomposition of a molecule into
atoms shows that these atoms remain separate structures when combined under their
mutual influence in the molecule, instead of being fused together. Each of them will
therefore preserve its free periods of vibrations, slightly modified however by the
proximity of the other one. For the case of a molecule containing two identical atoms
revolving at a distance large compared with their own dimensions, each of these
identical periods would be doubled*: thus the series of lines belonging to the atom
would become double lines in the spectrum of the molecule. It has been remarked
that the series in the spectra of inactive elements like argon and helium consist of
single lines, those of univalent elements such as the sodium group where the molecule
consists of two atoms, of double lines, while those of elements of higher valency appear
usually as triple lines.
In other words, a diad molecule consists of the two atoms rotating round each
other with but slight disturbance of the internal constitution of each of them. Their
vibrations relative to a system of axes of reference rotating along with them will thus
be but slightly modified: relative to axes fixed in space there must be compounded
with each vibration the effect of the rotation, which may be either right-handed or
left-handed with respect to the atom: thus on the same principles as above each line
will be doubled. If the lines of a spectral series are assumed to belong to a definite
atom in the molecule, those of a molecule consisting of two such atoms would thus
be a system of double lines with intervals equidistant all along the series, but in this
case without definite polarizations.
But if the constituents of the double lines of a series were thus two modifications of the same modes of the simpler atomic system, it would follow that they
should be similarly affected by a magnetic field. This is not always the case, so that
* In illustration of the way this can come about, revolution is different, and each single undisturbed period
consider two parallel cylindrical vortex columns of finite becomes two adjacent disturbed periods. Analogous consection in steady rotation round each other. Each by siderations apply to the interaction of the two atoms of the
itself has a system of free periods for crispations running molecule, rotating round each other.
round its section: when one of them is rotating round the According however to Smithells, Dawson, and Wilson,
other, the velocity of the crispations which travel in the Phil. Trans. 1899 A, it is the molecule of sodium that gives
direction of rotation is different from the velocity of those out the yellow light, that of sodium chloride not being
that travel in the opposite direction: thus the period of effective.
VOL. XVIII. 50