It may happen in the course of the hammering that from b to b, becomes loose, whilst the extreme edge a a, is also loose, and that the intermediate part towards b, requires to be stretched. These minor differences cannot be told alone by bending the plate with the fingers, as errors frequently exist which arc too minute to yield to their pressure, and then, the eye and straight-edge are conjointly employed in the examination.

In a saw, the general aim is to leave the edge rather tight or small, as then, the small amount of expansion it acquires when at work, from heat and friction, will enlarge the edge just sufficiently to bring the saw into a state of uniform tension. Otherwise, if before the saw is set to work the edge is fully large enough, when expanded by the heat it is almost sure to become loose on the edge, and to vibrate from side to side, without proper stability; so as to produce a wide irregular cut, and make a flanking whip-like noise, arising from the violent vibration of the buckled parts of the plate, in passing through the saw kerf; the sides of the wood will then exhibit ridges like the ripple marks on the sandy shore.

In hammering all plates, preference should in the like manner be given to keeping the edge rather small or stiff, to serve as a margin or frame to the more loose parts within; it gives a degree of stability somewhat as if the object had a thickened rim, and when a rim really exists, the process of flattening is comparatively easy.

If by undue stretching, the edge is made too loose, the whole piece becomes flaccid and very mobile, and we seem to lose the governing power, or those retaining points by which the changes of the plate are both influenced and rendered apparent; the edge thould be therefore always kept somewhat tight, from being proportionally less hammered, especially as the edge more easily admits of expansion than the inner part.

As a general rule, it may be said that every part of the plate which is straight and tense, whilst others are curved and flaccid, denotes that every straight part is under restraint; and that its straightness is due to its being, as it were, stretched either lengthways or around its edges, by the other parts which are too loose, and therefore arched, and also strong. In such cases, the straight lines require to be extended in length, to allow sufficient room for the curves to expand to their proportional sizes. This refers not only to small local errors towards the inner part of the plate, as explained by diagram, fig. 288, p. 418; but should the one edge of a plate be tolerably straight, whilst the opposite is loose and flaccid, the rule also applies with equal truth, and the straighter side must be hammered; in this case the curved side is as it were a great bulge cut in two parts.

Should a circular saw have a sudden dent, such as at g, fig. 290 on the last page, standing the reverse way, and which may with the hammer. 421 result from its having rested upon a small lump of coke whilst in the fire, the first blows will be given on the hollow side, between the lines ii, to lessen the abruptness of the margin by stretching it to the dotted curve, and then it will he driven downwards by violent blows, to form a part of the general sweep or concavity; a little time is gained by these driving blows, over the mode of stretching by the hammer.

The foregoing descriptions have all referred to solid blows, upon the face of the hard anvil, but to expedite the process, recurrence is often had to blocking, which is only one application amongst many others of a wooden anvil or block with a narrow flat-faced hammer, such as fig. 263, page 399. In this case the blows are to a certain extent hollow, as the wood immediately beneath the hammer-face yields to the blow, whereas the margin around the same does not. Such blows are therefore unquestionably hollow, and bend with very little stretching.

The blocking is considerably employed in saw-making, after the loose parts have been entirely removed, as the hollow blows correct any flight errors of figure, by bending alone, and with little risk of stretching the plates, if the work be delicately performed.

Towards the conclusion, however, all the different modes of work are required to be used in combination, as the true condition of the plate is only the exact balancing of all the forces, or of the tension of the several parts; and it constantly happens that attention to one error causes a partial change and fluctuation throughout the whole. It therefore requires great tact to know when when to leave the anvil for the block, and when to return to the anvil, and so on alternately; and also which side of the plate should be upwards for the time, which particular points should be struck, and the required force of the blows.

Of course, within certain limits, a thick plate is easier to hammer than a thin one, as the latter is difficult from its excessive mobility; also a soft plate of iron is more difficult than a hard plate of steel, although the latter requires more blows to produce the same effect; but when the works are very thick they become laborious, and the difficulty always increases rapidly with the MM of the plate.

Those who may desire to practise this art should therefore commence with a plate some 4, 6, or 8 inches square, and moderately stout, and subsequently proceed to pieces larger and thinner. They will also find some advantage in raising the anvil to within about a foot of the eye, as the alterations can he then more easily seen whilst the work lies on the anvil, and the effect of any predetermined blows can be the better watched. One other observance is essential, namely, patience; as although the process is thoroughly reducible to system, and no blow should be struck in vain, the beginner will frequently find it necessary to pause, examine, and consider, especially as the errors decrease; whereas the accustomed eye will follow the fluctuations of the plate almost without intermission of the blows, and will also accomplish the task with the fewest possible number of blows, which is the great desideratum.

Indeed it may happen from hammering some parts of a plate excessively and improperly, that it is rendered so hard and rigid, as to make its correction very tedious, or indeed nearly impossible without previous annealing, as the plate might burst or crack from the extension being carried beyond the safe limit of malleability. As in raised works, the annealing is mostly done by a gentle red heat; but in hardened steel plates, a slight increase of temperature barely sufficient to discolour the plate, will make a perceptible difference; and this latter process is always the last step in making a saw, in order to restore, by a gentle heat, the proper elasticity which has been mysteriously lost in the grinding, polishing, and hammering required in its manufacture.