Now suppose that the two plates were curved to an equal degree as in [Fig. 2443], and the bearing marks would extend all over both surfaces; but we may discover this error by turning one plate at a right angle, as in [Fig. 2444], in which case the bearing marks would show along the edges of No. 1 and along the middle of No. 2, and we may correct each with the file until both plates mark all across and from end to end when tried together lengthways as in [Fig. 2443], and one across the other as in [Fig. 2444]. But the plates may be curved to a different degree, as in [Fig. 2445], and it then becomes necessary to know which to file the most in correcting them and fitting them together, which we may discover as follows:—
Fig. 2445.
Fig. 2446. Fig. 2447.
We give one plate a light coat of red marking and rub it upon the other both sideways and lengthways. Suppose that on being separated and examined the bearing marks, shown as at a a and b b, [Fig. 2446], on one plate, and at c c and d d, [Fig. 2447], on the other, and as those at a a and b b are the narrowest, or in other words extend the least distance across the plate, it is proof that this plate is more concave than the other plate is convex, and therefore needs the most correction. This is plain because whatever part of a plate touches another, will, if the two are merely pressed together, only leave a bearing mark equal in area to itself, while this area will obviously be increased in proportion as one plate is moved about upon the other.