Avoid Blue-heat.

With steel the tensional, transverse and compressive stresses vary greatly with the composition of the material. As already stated the more carbon it contains, the less elastic, but the harder is the steel, while the less carbon it contains, the more elastic, but less hard will it be. The ultimate limits for tension per square inch are given in Table IV from 42000 pounds to 108000 pounds and for compression from 90000 pounds to 150000 pounds, all per square inch.

Steel like wrought-iron is stronger in bars than in plates (plate-steel being about 80 per cent of bar-steel), and is stronger with the grain than across it. The proportions for tension and shearing across or along the grain being about the same in steel as in the wrought-iron. In regard to re-working, according to Mr. Clay, the strength of steel is increased to the fourth piling, and then declines, until at the seventh re-working it is weaker than after the first working. Or, if the tensional strength of the first working is = 1,0 the fourth is= 1,253 and the seventh = 0,94. This, however, is quite as dubious as the advantage of re-working wrought-iron very often. Welded joints in steel are very much weaker than the original metal, having only some 40 per cent to 55 per cent of the strength of the unwelded parts.

If steel plates are worked cold they lose strength greatly. (Unwin, however, denies this for mild steel if working is uniform.) Punching them cold decreases their tensional strength some 33 per cent, if very hard steel; the milder the steel the less will the loss be. But this loss can be restored again by annealing after punching. Annealing instead of damaging steel plates greatly increases their tensional strength, particularly if the steel is very hard, adding some 50 per cent to same; but is of Blight, if any, advantage in mild steel. But after all said, the only sure method for the architect to pursue is to make careful tests to see whether the material possesses the necessary qualities and strengths he desires. Chemical tests should therefore give way to practical tests. The tensional or compressive ultimate strength can be readily ascertained by testing small pieces in testing-machines, and making proper allowances for the differences in sizes of test-pieces and full size members. These machines are of various sorts, but generally by hydraulic power or weights, greatly increased by leverage or screw-action, they tear small specimens apart, or else crush them, the effects on the pieces being carefully noted and recorded. Some of the machines make automatic records. The kind of machine used is not so essential to the architect so long as he has ascertained its reliability. From the tensional and compressive ultimate strength the cross-breaking strength of the specimen can be calculated; or the modulus of rupture can be obtained by breaking specimens across with centre loads and calculating the modulus of rupture by the following Formula :

Steel varies with carbon in strength.

Do not punch steel cold.

Tests for strength.

Modulus of rupture from tests.

k = 3/2.w.l/b.d2

(102)

Where k = the ultimate modulus of rupture of a material, in pounds, per square inch.

Where w= the load or amount of pressure, in pounds, applied at the centre, required to break a specimen of rectangular cross-section, lying on two supports.

Where l = the clear length, in inches, between supports.

Where b = the width of specimen, in inches, measured (across the specimen) at right angles to the line of pressure.

Where d = the depth of specimen, in inches, measured along the line of pressure.

Where the test specimen is exactly square in cross-section and of one square inch area, that is one inch by one inch, and where the supports are exactly twelve inches apart, and the load or pressure applied in the centre, the modulus of rupture will always be eighteen times the load or pressure, or: k = 18.w (103)

Where k = the ultimate modulus of rupture of a material, in pounds, per square inch.

Where w = the load or pressure, in pounds, applied at the centre, required to break a test specimen, of one inch by one inch (square) cross-section, lying on supports exactly twelve inches apart. The shearing strength is generally found by direct tests.

Test specimens of steel, wrought or cast iron, should never be broken off suddenly, or by blows; nor should they be jarred, as otherwise the fibres crystallize more or less and this greatly affects the result. They should be carefully planed off in the machine-shop.

Another important point is to be sure to have all test specimens, when not cut off the actual pieces being used, of the same material and general thickness as the pieces.

In cast-iron, frequently, test-pieces are cast on to each piece, these are broken off and tested. They should not be broken off, however, except in the presence of the architect or his inspector. If they are of the same thickness as the piece, they offer a fair test ; if, however, they are much smaller and thinner than the piece no reliance can be placed on the result.

In wrought-iron and steel it is best to roll the pieces a little too long and plane off the superfluous ends for testing.

Besides the tests as to strength there are many tests resorted to to ascertain approximately the quality of the material. A few will be given here. If cast-iron be struck on its edge a sharp blow with a chisel or hammer, it is of soft and good quality if it can be indented. If it breaks it is very hard and brittle. If it rings out clearly, it is a good casting ; if the sound is dull it is full of sand-holes, air-bubbles or flaws.

If the surface is smooth, even and hard, and the edges sharp and perfect it is a good casting. If the edges or surfaces are uneven and wavy it is an indication of unequal shrinkage and cooling or more likely unskilful ramming of the sand in the. flask around the pattern or mould. If hollow pieces, when tapped, show uneven thicknesses on opposite sides the core has sagged or floated in the mould.1