Now when a stria has been noted mark its location as to the surface, and, moving the eye a little, look for parallax of the fault with respect to the surface mark. If it appears to shift try a mark on the opposite surface in the same way. Comparison of the two inspections will show about where the fault lies with respect to the surfaces, and therefore what is the chance of working it out. Sometimes a look edgewise of the disc will help in the diagnosis. Numerous barely detectable striæ are usually worse than one or two conspicuous ones, for the latter frequently throw the light they transmit so wide of the focus that it does not affect the image, which could be greatly damaged by slight blurs of light that just miss focus. Given a disc that passes well the tests for striæ and the like the next step is to examine the perfection of the annealing, which in its larger aspect is revealed by an examination in polarized light. [Illustration: FIG. 43.—Testing a Disc in Polarized Light.] For this purpose the disc is set up against a frame placed on table or floor with a good exposure to skylight behind it, and inclined about 35° from the vertical. Behind it is laid a flat shiny surface to serve as polarizer. Black enamel cloth smoothly laid, a glass plate backed with black paint, or even a smooth board painted with asphalt paint will answer excellently. Then holding a Nicol prism before the eye and looking perpendicular to the face of the disc, rotate the prism on its axis. Fig. 43 shows the arrangement, _A_ being the eye, _B_ the Nicol, _C_ the disc, and _D_ the polarizer behind it. If annealing has left no strain the only effect of rotating the Nicol will be to change the field from bright to dark and back again as if the disc were not there. Generally a pattern in the form of a somewhat hazy Maltese cross will appear, with its arms crossing the disc, growing darker and lighter alternately as the Nicol is turned. If the cross is strongly marked but symmetrical and well centered the annealing is fair—better as the cross is fainter and hazier—altogether bad if colors show plainly or if the cross is decentered or distorted. The test is extremely sensitive, so that holding a finger on the surface of the disc may produce local strain that will show as a faint cloudy spot. A disc free of striæ and noticeable annealing strains is usually, but not invariably, good, for too frequent reheating in the moulding or annealing process occasionally leaves the glass slightly altered, the effect extending, at worst, to the crystallization or devitrification to which reference has been made.