We have seen how any type of polarized light can be turned into plane polarized, and it remains to consider how this can be investigated. The standard instrument is the nicol (see fig. 16). A rhomb of calcite, which is about 3 times as long as it is broad, is sawn across the line AC, and, after the cut has been polished, its two faces are cemented together with Canada balsam. The two ends AD and BC are also cut to a slightly different angle (68° in stead of 72°). A wave, entering parallel to the prism, is doubly refracted, and, because of the slope of the face AD, the waves go in somewhat different directions, the ordinary wave being most Dent. Canada balsam has refractive index intermediate between those of the ordinary and extraordinary waves, and, as the ordi nary wave meets it very obliquely, it becomes totally reflected and so falls on the side DC, which is blackened so as to absorb it. The extraordinary wave passes freely through, and so the emergent light is plane polarized. A nicol will polarize light over a certain range of angles of incidence. The limits are set by the condition that on one side both rays are reflected at AC and lost, and on the other that both are transmitted ; under favourable conditions the range of angles may be 25°. A typical experiment with polarized light involves the use of two nicols. The simplest example is the direct use of two "crossed" nicols. The first produces polarized light by absorbing one component ; the second is placed with its axis in the same line as the first but is twisted round through go°, so that the wave which was extraordinary wave in the first nicol is now the ordinary and so is absorbed. This device has been used for secret signalling. The signaller rotates his nicol through go°, and though to the naked eye his lamp appears to burn steadily, if it is viewed through a fixed nicol it will go out.
When we want to study the polarization of elliptically polarized light, it must first be made into plane polarized light by some means, and then extinguished by a nicol. There will always be two adjustments to be made, because the light has two parameters, the ratio of the axes of the ellipse and their position. There are two chief methods used. A quarter-wave plate may be rotated until it lies along the principal axis, so that it renders the light plane po larized; the light can then be extinguished by a nicol. The position
of the quarter-wave plate fixes the axes of the ellipse, and the angle it makes with the nicol gives the ratio of the axes. This method is specially convenient if we already know the axes of the ellipse, as is the case in some important experiments; in other cases the simultaneous making of two adjustments is troublesome. The alternative instrument is called the compensator. In this the analysis is in two fixed perpendicular directions, instead of along the axes of the ellipse, and we measure the relative lengths of the components of the light vector in those two directions and their phase-difference. The essential part of the compensator consists of two thin wedges of quartz, which is a rather weakly doubly re fracting crystal. One wedge is cut so that the crystal axis lies in a line parallel to the edge, and the other so that it is at right angles to it and in the face; (see fig. 17, where the lines and dots indi cate the directions of the axes). Incident light will be broken into two components in the first wedge, of which one will have a higher phase-velocity ; but on entering the second this will change about, so that, if the thicknesses are exactly equal as at the centre A, there will be no changes of phase. Thus, incident plane polar ized light will emerge as plane polarized at A. At any other point B this will not be so, because the light will have gone further through one crystal than through the other. If elliptically po larized light is incident, there will be places where it is plane polarized, and so, if we use an analyzing nicol set at the proper angle, we may observe the field crossed by dark bands. If one wedge is screwed over the other by means of the micrometer screw Al, the bands will shift be cause of the changing differences in the thickness of the wedges.
The reading consists in seeing how far the screw must be turned to bring one of the bands to the centre. This measures the phase difference in the components of the light vector along and across the wedge, while the ratio of their amplitudes is given by the setting of the nicol.