Then begins the careful and tedious process of cooling. Fairly rapid until the mass is solid enough to prevent the formation of fresh striæ, the cooling is continued more slowly, in the furnace or after removal to the annealing oven, until the crucible is cool enough for handling, the whole process generally taking a week or more. Then the real trouble begins. The crucible is broken away and there is found a more or less cracked mass of glass, sometimes badly broken up, again furnishing a clear lump weighing some hundreds of pounds. This glass is then carefully picked over and examined for flaws, striæ and other imperfections. These can sometimes be chipped away with more or less breaking up of the mass. The inspection of the glass in the raw is facilitated by the scheme shown in elevation Fig. 37. Here _A_ is a tank with parallel sides of plate glass. In it is placed _B_ the rough block of glass, and the tank is then filled with a liquid which can be brought to the same refractive power as the glass, as in Newton’s disastrous experiment. When equality is reached for, say, yellow light, one can see directly through the block, the rays no longer being refracted at its surface, and any interior striæ are readily seen even in a mass a foot or more thick. Before adding the liquid a ray would be skewed, as _C_, _D_, _E_, _F_, afterwards it would go straight through; _C_, _D_, _G_, _H_. The fraction that passes inspection may be found to be from much less than a quarter to a half of the whole. This good glass is then ready for the next operation, forming and fine annealing. The final form to be reached is a disc or block, and the chunks of perfect glass are heated in a kiln until plastic, and then moulded into the required shapes, sometimes concave or convex discs suitable for small lenses. Then the blocks are transferred to a kiln and allowed to cool off very gradually, for several days or weeks according to the size of the blocks and the severity of the requirements they must meet. In the highest class of work the annealing oven has thermostatic control and close watch is kept by the pyrometer. It is clear that the chance of getting a large and perfect chunk from the crucible is far smaller than that of getting fragments of a few pounds, so that the production of a perfect disc for a large objective requires both skill and luck. Little wonder therefore that the price of discs for the manufacture of objectives increases substantially as the cube of the diameter.