The rough grinding is accomplished by means of a lead tool and coarse emery; the size of grain may be such as will pass a sieve of 60 threads to the inch. The process of grinding is quite similar to that previously described, but it goes on comparatively quickly. The rough grinding is checked by the spherometer, and is interrupted when that instrument gives accordant and correct measurements all over the surface. The fine grinding may be proceeded with by means of a glass-faced tool as before described, or the labour may be reduced in the following manner. A slate tool, which must be free from green spots (a source of uneven hardness), is prepared, and this is brought nearly to the curvature of the roughly ground speculum, by turning or otherwise. It is finished on the speculum itself with a little flour of emery. The fine grinding is then carried on by means of slate dust and water, the slate tool being the grinder. The tool is, of course, scored into squares on the surface. If the casting process has been carried out successfully, the rough grinding may take, say six hours, and the fine grinding say thirty hours for a disc a foot in diameter. The greatest source of trouble is want of homogeneity in the casting, as evidenced by blowholes, etc. In general, the shortest way is to discard the disc and start afresh if there is any serious want of perfection in the continuity or homogeneity of the metal. Fig. 62. The finely ground surface must, of course, be apparently correct in so far as a spherometer (with 3 inches between the legs for a disc 1 foot in diameter) will show. Polishing and figuring are carried out simultaneously. Half an hour's polishing with a slate-backed pitch tool and rouge and water will enable an optical test to be made. The most convenient test is that of Foucault, a simple appliance for the purpose being shown in the figure (62). It essentially consists of a small lamp surrounded by an opaque chimney (A) through which a minute aperture (pin-hole) is made. A small lens may be used, of very great curvature, or even a transparent marble to throw an image of the flame on the pin-hole. A screen (B) is placed close to the source, and is provided with a rocking or tilting motion (C) in its own plane. The source and screen are partly independent, and each is provided with a fine adjustment which serves to place it in position near the centre of curvature. The screen is so close to the pin-hole in fact that both the source and a point on the edge of the screen may be said to be at the centre of curvature of the mirror. The mirror is temporarily mounted so as to have its axis horizontal, in a cellar or other place of uniform temperature.