accuracy the exact proportionality of the rotation to the distance
traversed by the polarized beam. Thus in one series of experiments
where the rotation required by the direct beam was 12degrees, that
acquired by three passages through the glass was 36degrees, while that
acquired by five passages was 60degrees. But even when this method of
magnifying was applied, he failed with various solid substances to
obtain any effect; and in the case of air, though he employed to the
utmost the power which these repeated reflections placed in his
hands, he failed to produce the slightest sensible rotation.

These failures of Faraday to obtain the effect with gases seem to
indicate the true seat of the phenomenon. The luminiferous ether
surrounds and is influenced by the ultimate particles of matter.
The symmetry of the one involves that of the other. Thus, if the
molecules of a crystal be perfectly symmetrical round any line
through the crystal, we may safely conclude that a ray will pass
along this line as through ordinary glass. It will not be doubly
refracted. From the symmetry of the liquid figures, known to be
produced in the planes of freezing, when radiant heat is sent
through ice, we may safely infer symmetry of aggregation, and hence
conclude that the line perpendicular to the planes of freezing is a
line of no double refraction; that it is, in fact, the optic axis of
the crystal. The same remark applies to the line joining the
opposite blunt angles of a crystal of Iceland spar. The arrangement
of the molecules round this line being symmetrical, the condition of
the ether depending upon these molecules shares their symmetry; and
there is, therefore, no reason why the wavelength should alter with
the alteration of the azimuth round this line. Annealed glass has
its molecules symmetrically arranged round every line that can be
drawn through it; hence it is not doubly refractive. But let the