On thoroughly annealing hardened steel it recovers the original, dimensions possessed before hardening; evidently therefore when a large steel tool is hardened, since the inner portion cannot possibly become cooled at the same rate as the outer film, a kind of strain must be developed amongst the particles; for, whilst the rapid chilling of the outer portion causes that part to occupy a larger volume than it otherwise would, this is not effected to so large a relative extent with the inner portions. The opposite result is necessarily produced during annealing. Accordingly, in the process of tempering steel goods, cracking and splitting or “buckling” and otherwise getting out of shape are inconvenient results that are apt to happen, especially when all parts of the surface are not equally and simultaneously heated or chilled, a result difficult to bring about with articles of certain shapes; the hardening process may in such cases be more safely effected by heating the article up gradually in a fluid bath such as melted lead (or in some cases oil) and then chilling in a cool fluid bath, preferably of oil; whilst the annealing may be performed by allowing the fluid in which the article is slowly heated up to cool down again spontaneously when the requisite temperature is attained. For special articles, c.y., saws, a particular manipulation is requisite to avoid buckling; in some cases the annealing of a flat plate, e.g., a circular saw, is effected between two solid flat masses of iron, which keep the plates from getting out of shape whilst cooling. In some instances when the point only of an article is required to be tempered, e. g., certain kinds of chisels, the whole mass is heated red hot and the point dipped into water or oil to harden it; on taking out, the heat from the body of the hot article is rapidly conducted to the point; as soon as its temperature is sufficiently raised the body is somewhat cooled by a short immersion so as to reduce the temperature of the whole approximately to uniformity, and then the article is allowed to cool altogether slowly; so that the hardening and annealing are effected by one heating only. For articles the temper of which requires to be somewhat exactly determined, the temperature attained is judged of by the tint of the iridescent film of oxide which forms on the surface (previously brightened) during the heating; the higher the temperature the thicker the film, the colours and temperatures being approximately as follows :— [table]

For instance, if a saw is to be tempered at a full blue it is heated over a clear fire or a mass of red hot iron or in a sand bath until the tint appears on the brightened surface, when it is allowed to oool. In many cases the tempering of tools is effected with sufficient accuracy without brightening them, by coating them with tallow, oil, or some similar composition, and then heating over a lamp or amass of hot metal until the tallow, &c., begins to decompose; with sufficient experience as to the size of the article, the mode of heating, and the character of the decomposition set up, much the same results are arrived at as when the colour is watched on the brightened surface. In some instances the requisite temper is given by one operation only, the point of the article to be hardened (a drill, say) being coated with tallow and heated in a flame until the tallow decomposes rapidly, and then plunged into a mass of tallow to chill it; the point is thus rendered harder than the body of the drill. By heating the point of a small tool nearly white hot and plunging it into a stick of sealing wax for a second, and then into another part of the wax, and so on till cold, steel may be made so hard as to be readily capable of boring into similar steel hardened in the ordinary way.

Instead of determining the temperature of the object to be annealed by the tint developed on it, various fusible alloys may be used, a series of such being placed in the annealing bath, so that by observing which are fluid the temperature is known approximately. Parkes gives the following compositions of lead-tin alloys for this purpose :— [table]

The first five of these alloys consequently correspond to the various yellow shades; the next three to brown, brown and purple spots, and purple; and the last four to the blue shades.

43. Strength and Tenacity of Iron and Steel.— Independently of the additional strength communicated to iron by the presence of small quantities of carbon, producing steels or so-called steels of various kinds, a considerable increase in this property is brought about by the removal of portions of cindery matter by fusion. Wrought iron that has simply been welded, reheated, and rolled a given number of times is, caeteris paribus, less tenacious than iron from which complete fusion has removed the films of silicate, &c., enveloping the component layers and fibres of the welded material, and preventing their complete union together. Thus the best qualities of weld iron, containing usually about 0 - 2 to 0 - 3 per cent, of carbon, possess a tensile strain of from 20 to 30 tons per square inch, whilst mild “steels” and ingot irons are usually considerably higher in this respect, viz., from 30 to 40 tons per square inch; harder steels of higher carbon percentage are proportionately stronger. The following numbers may be quoted, representing tensile strains in tons per square inch. [table]

Steel wire of as high a breaking strain as 206,170 lb, or 92 tons, per square inch section, has been drawn by Johnson and Nephew (Manchester), whilst wire of tensile strength of 70 to 80 tons per square inch is readily procurable. According to Collingwood, galvanizing increases the strength of freshly drawn wire by about 15 per cent., whilst ungalvanized wire has been found to gain nearly 5 per cent, by merely being allowed to remain at rest for a week or so; the increase of strength in each case appears to be simply due to internal adjustment of the strains produced during drawing.

It does not necessarily follow that the resistance to percussive force exhibited by a given sample of metal will be in the ratio of its tensile strength. Thus phosphorus when present together with only minute quantities of carbon (forming the so-called “phosphoric steels ”) does not very materially decrease the tensile strength, such steels containing 0 ’ 3 per cent, of phosphorus being often but little inferior to soft non-phosphorized steels in this respect; but when tested by a “falling weight” (a mass of known weight falling once or oftener from a known height upon the centre of the bar or rail firmly gripped in supports a known distance apart) the phosphorized metals generally show themselves considerably inferior to the non-phosphorized ones. The same remark applies to silicon. According to

Dudley the effect of phosphorus, silicon, and carbon in hardening iron and making it less capable of resisting percussion are nearly in the proportions of 3, 2, and 1 relatively to one another. The presence of manganese diminishes this deleterious effect of non-metals; whilst, if more than minute quantities of carbon be present, the tensile strength as well as the resistance to percussion is greatly diminished by the additional presence of phosphorus or silicon in proportions beyond certain small limiting amounts. Accordingly it is the usual practice to test rails, bars, &c., not only by the determination of the breaking strain for tensile force (measured by pulling asunder, preferably by hydraulic power, a bar turned to known definite dimensions, and made into the shape of fig. 65), but also by a falling weight,—a “monkey” (somewhat like a pile driver) being raised to a known height and let drop upon the rail.

The particular tests applied in different instances vary much; for instance, some little while ago the official falling weight test for Bessemer rails at Gratz (Austrian South Railway Company’s Works) was to permit a weight of 1000 kilos (about a ton) to fall from a height of 15 feet upon the centre of the rail supported by two rests 3 feet apart, any amount of bending being allowed, but not fracture, whilst a test for elasticity or resistance to permanent deflexion was applied by placing a weight of 17,500 kilos on the middle of the rail similarly supported. The North-Eastern Railway (England) similarly at one time tested rails by allowing a weight of 1800 lb to fall from 4 feet height, the number of blows requisite to produce rupture and the permanent bending produced by each being noted. In other instances the test applied has been a ton weight falling a greater height, such as 20 feet or even 30 feet, the rail being required to stand one such blow only, or a succession, the particular details of the test to be applied being usually specified in each particular case; thus the Midland Railway Company has tested steel rails by allowing a weight of 1 ton to fall three times from a height of 12 feet, the supports being 4 feet asunder.

In just the same way as regards the determination of tensile strength, the dimensions of the piece to be tested (6 inches, 8 inches, 10 inches long, &c.) are usually specified, and the strain which the metal will just stand without becoming permanently elongated {limit of elasticity) determined, as well as the total strain requisite to produce rupture, together with the “ductility” or amount of permanent extension of the test piece and the diminution in section of the bar at the point of rupture. Thus for instance the following numerical data were obtained by Kirkaldy with a particular specimen of West Cumberland Bessemer steel plate, three pieces being tested, each 10 inches long, and respectively ∣, ⅜, and ∣ inch in thickness. [table]

These numbers illustrate, amongst other things, the effect upon the final values produced by variations in the diameter of the test pieces; the shorter and thicker the piece the greater in amount is the permanent extension. In calculating the rupturing strain per square inch, the dimensions of the metal as originally employed before permanent alteration was brought about are employed; by taking the diminished area at the point of fracture as the section, a much higher value is obtained as the tensile strength per unit area of the extended metal. Tempering steel greatly increases its breaking strain and limit of elasticity, but decreases the permanent extension; thus the following values represent certain results obtained with Creuzot steels of the A class, in tons per square inch (see Engineering, 1875, p. 119). [table]

In making a contract for the supply of steel of a particular quality, the details of the tests of strength to be applied should consequently be duly set forth; for instance, a short time ago the French Government required certain steel navy tubes, of which the limit of elasticity and breaking strain tested in a particular way were respectively to be 21 and 38 tons per square inch. Similarly in the construction of the Mississippi great bridge the cast steels used were contracted to be of the following qualities :—“To be of crucible cast steel; the staves of the tubes to stand a compressive strain of 60,000 and a tensile strain of 40,000 lb per square inch section without permanent set, and to stand a tensile strain of 100,000 lb per square inch without fracture. Modulus of elasticity to be between 26 and 30 million lb, preferably nearer the lower limit, and as constant as possible; bars of the same modulus to be selected for the tubes, so that each side shall have same power of resistance; each bar to be tested and modulus stamped on it. Steel pins, rods, bolts, eyewashers, rivets, &c., and the ⅛ inch steel plates for enveloping the staves to stand a tensile strain of 40,000 fb per square inch without permanent set, and an ultimate tensile strain of 100,000 lb without fracture.”

Practical tests of the capability of metal to stand bending double or through some given angle, or twisting round and round in the cold without fracture are often applied, as are also tests of the capability of being bent hot, forged, welded, &c. A test as to the power of resisting repeated bending strains backward and forward through a given angle is sometimes applied; a particular mechanical arrangement for effecting this has been described by Olrick.

James Price has constructed a machine for testing rails as to durability under rolling wear and tear, consisting of a pair of metal rollers 5 feet in diameter and 16 inches wide, weighing 45 cwts. each, supporting a frame weighing 6⅜ tons, connected with a centre boss and vertical axle, so that the rollers are driven round in a circle, one bearing with 5 the other with 6 tons pressure; the rails to be tested are bent into a circle or preferably a polygon, to equalize the wear of the rollers, which are driven over them at a speed of 13 or 14 miles per hour until the rails are broken or wear out.

It by no means follows that the rail which possesses the greatest tensile strength will resist wear and tear and rolling friction best, although this might be anticipated if, other things being equal, increased tensile strength corresponds to greater hardness; on the other hand, experience does not always indicate that the most carbonized rails last the longest, although the superiority of ingot metal (Bessemer steel, &c.) over weld iron (not fused) rails is well demonstrated; it is probable that the interposed film of cinder between the metallic fibres in the latter case greatly facilitates the destruction and wearing away of the upper surface, just as the accumulation of dust and sand on the rail between the passage of consecutive trains, especially conjoined with moisture, conduces to mechanical abrasion. Price Williams has calculated from the results of various British lines that the average tonnage lives of wrought iron and Bessemer rails (i.e., the traffic in tons requisite to wear away f inch of the head of the rail) are respectively close to 17⅜and 161 million tons, the latter being thus more than nine times as lasting as the former.

Numerous observations have been made on the effect on the strength of iron and steel of punching and dialling holes, of notching and otherwise removing part of the surface, and of shearing, with the general result of indicating that the disturbance produced in the relative positions of the constituent particles by forcibly punching and shearing in the cold materially decreases the strength of a bar or plate (apart from the actual removal of substance); but that drilling does not effect the strength in the same way (see various papers in the Journal I. and S. Inst., Iron, and Engineering during the last few years). Annealing restores the strength to a considerable extent; if the plates be punched whilst red hot the annealing takes place spontaneously.

The strengthening effect upon soft steels and ingot irons of hardening by heating and plunging into oil is often very marked, the tendency to crystallinity observable in large masses of cast metal being thus largely removed. A valuable paper 0 on the causes and effects of hardening iron and steel, by Professor Akermann, is to be found in the Journal I. and 8. Inst., 1879, 504; whilst the Research Committee of the Institution of Mechanical Engineers has recently issued reports containing much information on thiä subject.

Effect of Temperature on the Strength of Iron and Steel. —Many observations on this point have been made by different experimenters, with the general result of indicating that at 0° C. and below the tensile strain and resistance to percussion of iron and steel bars, rods, &c., are substantially the same as at the average ordinary temperatures of 15° to 20°, but that what difference there is is usually in the direction of diminution in strength; the numerical values obtainable are considerably variable with the composition of the metal, Ac.; thus Webster finds that a severe cold of - 15° C. does not affect the tensile strain of wrought iron and steel, although it slightly increases the ductility by about 1 per cent, with iron and 3 per cent, with steel; the power of resisting transverse strain is, however, some 3 per cent, lower, whilst the flexibility and the resistance to rupture by impact are reduced by the following amounts :— [table]

A committee of engineers appointed by the Russian Government for the purpose of examining carefully into this question has recently found that, when the amount of phosphorus, silicon, and carbon in Bessemer and Siemens-Martin steel rails exceeds jointly about 0 ’ 54 (varying from 0 ’ 44 to 0 ’ 67) per cent., the rails are decidedly more brittle at temperatures near to - 20° C. than at the ordinary temperature ( 4- 10° to 20°); whilst the effect of low temperature in producing brittleness is not marked when the phosphorus, silicon, and carbon jointly amount to only about 0 ’ 41 (0 , 37 to 0 ’ 55) per cent. Expressing the amounts of non-metals on the scale proposed by Dudley (3 parts of carbon, 2 of silicon, and 1 of phosphorus being considered as equivalent to one another, so that a rail containing carbon = 0'30, silicon — 0 - 20, phosphorus = 005, would be equivalent to one containing phosphorus = 0 ’ 25), these results may be expressed as follows: when the percentages of carbon, silicon, and phosphorus are jointly equivalent to an average of 0 ’ 19 per cent, of phosphorus, the effect of low temperatures is not marked; but when they are equivalent to an average of 0’31 per cent, of phosphorus, the rails are more brittle at temperatures near - 20° than at ordinary average temperatures near to + 15°.

Breakages of axles, crankshafts, pumprods, &c., exposed to strains and vibration appear to be more common in frosty weather than at other seasons of the year; it is considered by many that exposure to vibration, &c., and low temperature simultaneously tend to diminish tenacity and set up a brittle structure in a way not observed when only one of the two causes alone acts; direct evidence on this point is, however, wanting. It is noticeable that large masses of cast metal (cast iron, true steel, or ingot metal) if cooled too quickly are apt to have the internal portions in a high state of tension or strain; for the outer portion, when solidified, prevents the contraction taking place that would otherwise ensue during the solidification and cooling of the inner portion; hence the amount of extra strain requisite to produce rupture is much diminished, so that the want of elasticity of bearings, foundations, &c., connected with the almost rigid ground during hard frosts in certain cases causes the strain applied during use to exceed the reduced amount which the metal can then bear without fracture. Chilled castings, case hardened iron, and tempered steel, moreover, are affected by low temperatures in another way; the outer harder portions do not expand at exactly the same rate as the inner softer parts; and consequently alteration of temperature produces such variations in the internal strain as in some cases to lead to fracture either spontaneously or by the superaddition of the strain due to ordinary use. On the whole it appears that no clear evidence is as yet extant proving that vibration either alone or concurrently with low temperature does actually cause a brittle crystalline structure to be...