_Experiment_ 3.--If we now take the plate of selenite thus prepared, and place it under the piece of brass on the stage, we shall see, instead of the alternate black and white images, two coloured images composed of the constituents of white light, which will alternately change by revolving the eye-piece at every quarter of the circle; then, by passing along the brass, the images will overlap; and at the point at which they do so, white light will be produced. If, by accident, the prism be placed at an angle of 45° from the square part of the stage, no particular colour will be perceived, and it will then illustrate the phenomena of the neutral axis of the selenite, because when placed in the relative position no depolarisation takes place. The phenomena of polarised light may be further illustrated by the addition of a second double image prism, and a film of selenite adapted between the two. The systems of coloured rings in crystals cut perpendicularly to the principal axis of the crystal are best seen by employing the lowest object-glass. _Biaxial Crystals._--To show perfectly the beautiful series of _rings and brushes_ which biaxial crystals exhibit, it becomes necessary to convert the microscope, for the time being, into (so to speak) a wide-angled telescope. [Illustration: Huyghenian Eye-piece. Inner draw-tube. Objective in draw-tube. Analysing Prism. Objective. Specimen under Examination. Sub-stage Condenser. Polarising Prism, fixed in sub-stage below. Fig. 184.--Diagrammatic arrangement of the Polarising Microscope. _In Sub-stage_: P, polarising prism; C, sub-stage condenser on stage; M, mineral or crystal. On nose-piece: O^1, objective, 4/10-inch; A, analysing prism. _In Draw-tube_: O^2, 2 or 3 inch Objective; H, Huyghenian eye-piece.] For the purpose, screw on a low-power objective to the end of the draw-tube (Fig. 184).[31] As the light requires to be passed through the crystals at a considerable angle, a wide-angled condenser should be employed, but it need not be achromatic. The objective most suitable is a 4/10-inch, of ·64 numerical aperture, but a 1/4-inch of ·71 numerical aperture, or a 1/3-inch of ·65 numerical aperture, will answer the purpose equally well. As the whole of the back lens of the objective should be visible through the analysing Nicol prism, the back lens of the objective must not be too large; thus a 1/2-inch of ·65 numerical aperture will not be so effective. The analysing prism may be placed either where it is in the drawing, below the stage, or above the eye-piece. It works equally well above the objective, the position it ordinarily occupies in the microscope.