during their whole passage. Lastly, if one of the plates be
turned through 90 deg., no bands will be seen, and the spectrum will be
alternately bright and dark, as if no plates were used, except only
that the polarization is itself turned through 90 deg..

'If a wedge-shaped crystal be used, the bands, instead of being
straight, will cross the spectrum diagonally, the direction of the
diagonal (dexter or sinister) being determined by the position of
the thicker end of the wedge. If two similar wedges be used with
their thickest ends together, they will act as a wedge whose angle
and whose thickness is double of the first. If they be placed in the
reverse position they will act as a flat plate, and the bands will
again cross the spectrum in straight lines at right angles to its
length.

'If a concave plate be used the bands will dispose themselves in a
fanlike arrangement, their divergence depending upon the distance of
the slit from the centre of concavity.

'If two quartz wedges, one of which has the optic axis parallel to
the edge of the refractory angle, and the other perpendicular to it,
but in one of the planes containing the angle (Babinet's
Compensator), the appearances of the bands are very various.

'The diagonal bands, besides sometimes doubling themselves as with
ordinary wedges, sometimes combine so as to form longitudinal
(instead of transverse) bands; and sometimes cross one another so as
to form a diaper pattern with bright compartments in a dark
framework, and _vice versa_, according to the position of the
plates.

'The effects of different dispositions of the interposed crystals
might be varied indefinitely; but enough has perhaps been said to
show the delicacy of the method of spectrum analysis as applied to
the examination of polarized light.'

* * * * *

The singular and beautiful effect obtained with a circular plate of
selenite, thin at the centre, and gradually thickening towards the
circumference, is easily connected with a similar effect obtained with
Newton's rings. Let a thin slice of light fall upon the glasses which
show the rings, so as to cover a narrow central vertical zone passing
through them all. The image of this zone upon the screen is crossed by
portions of the iris-rings. Subjecting the reflected beam to prismatic
analysis, the r