The scarf may be produced by filing, sawing or chiseling the ends, although this is not good practice because it is then impossible to give the desired upset and additional metal for the weld. This added thickness is called for by the fact that the metal burns away to a certain extent or turns to scale, which is removed before welding.

When the two ends have been given this shape they should not fit as closely together as might be expected, but should touch only at the center of the area to be joined (Figure 51). That is to say, the surface of the beveled portion should bulge in the middle or should be convex in shape so that the edges are separated by a little distance when the pieces are laid together with the bevels toward each other. This is done so that the scale which is formed on the metal by the heat of the fire can have a chance to escape from the interior of the weld as the two parts are forced together.

If the scarf were to be formed with one or more of the edges touching each other at the same time or before the centers did so, the scale would be imprisoned within the body of the weld and would cause the finished work to be weak, while possibly giving a satisfactory appearance from the outside.

Fluxes.--In order to assist in removing the scale and other impurities and to make the welding surfaces as clean as possible while being joined, various fluxing materials are used as in other methods of welding.

For welding iron, a flux of white sand is usually used, this material being placed on the metal after it has been brought to a red heat in the fire. Steel is welded with dry borax powder, this flux being applied at the same time as the iron flux just mentioned. Borax may also be used for iron welding and a mixture of borax with steel borings may also be used for either class of work. Mixtures of sal ammoniac with borax have been successfully used, the proportions being about four parts of borax to one of sal ammoniac. Various prepared fluxing powders are on the market for this work, practically all of them producing satisfactory results.

After the metal has been in the fire long enough to reach a red heat, it is removed temporarily and, if small enough in size, the ends are dipped into a box of flux. If the pieces are large, they may simply be pulled to the edge of the fire and the flux then sprinkled on the portions to be joined. A greater quantity of flux is required in forge welding than in electric or oxy-acetylene processes because of the losses in the fire. After the powder has been applied to the surfaces, the work is returned to the fire and heated to the welding temperature.

Heating the Work.--After being scarfed, the two pieces to be welded are placed in the fire and brought to the correct temperature. This temperature can only be recognized by experiment and experience. The metal must be just below that point at which small sparks begin to be thrown out of the fire and naturally this is a hard point to distinguish. At the welding heat the metal is almost ready to flow and is about the consistency of putty. Against the background of the fire and coal the color appears to be a cream or very light yellow and the work feels soft as it is handled.

It is absolutely necessary that both parts be heated uniformly and so that they reach the welding temperature at the same time. For this reason they should be as close together in the fire as possible and side by side. When removed to be hammered together, time is saved if they are picked up in such a way that when laid together naturally the beveled surfaces come together. This makes it necessary that the workman remember whether the scarfed side is up or down, and to assist in this it is a good thing to mark the scarfed side with chalk or in some other noticeable manner, so that no mistake will be made in the hurry of placing the work on the anvil.

The common practice in heating allows the temperature to rise until the small white sparks are seen to come from the fire. Any heating above this point will surely result in burning that will ruin the iron or steel being handled. The best welding heat can be discerned by the appearance of the metal and its color after experience has been gained with this particular material. Test welds can be made and then broken, if possible, so that the strength gained through different degrees of heat can be known before attempting more important work.