Wrought-iron is chiefly used in construction for parts subjected to a tensile strain, or that calculated to tear the fibres asunder, as in the case of the tie rod of a roof. The following gives the result of a few bars out of several experimented upon by Mr. Hodgkinson; the length of the bars being 10 feet, the cross section one square inch. With a weight of 3785 lbs., extension of the bar was .01690 in., the set of the bar being .0005 in.; with a weight of 6309 lbs., the extension was .02772 in., the set being .0007 in.; with a weight of 13,880 lbs., the extension was .05950, the set being .007; with a weight of 26,499 lbs., the extension was .1240, the set being .00680.

The following are some examples showing the resistance of wrought-iron to a cross or transverse strain, as in the case of a bar supported at both ends, stretching across a void space, the length of the bar being 14 feet 7½ inches, the depth in the direction of the pressure 1.585 in., breadth of the bar, or its thickness, 5.523 in., the distance between the supports or the span being 13 feet 6 inches. "With a weight of 28 lbs. applied and acting horizontally, the deflection after five minutes was .051 in.; with a weight of 56 lbs., the deflection was .112 in.; with a weight of 112 lbs., the deflection was .232 in.; with a weight of 224 lbs., the deflection was .458 in.; with a weight of 448 lbs., the deflection was .916 in.; with a weight of 1008 lbs., the deflection was 2.044 in. Wrought-iron, as already stated, is seldom used in constructions for parts subjected to direct compression, as a column or pillar. The student will find further on, while giving a few calculations on columns, some remarks of the resistance of wrought-iron to compression when used in this form.

With reference to the strength of wrought-iron plates, Mr. Hodgkinson says, "that the resistance of plates of the same length and breadth, but varying in thickness, was nearly as a cube of the thickness, or more nearly as the 2.878 power of it. Thus a plate of double the thickness of another would resist flexure or buckling with seven or eight times the force applied in the direction of its length." The mean breaking weight of wrought-iron plates in the direction of the fibre, in tons per square inch, may be taken from a series of experiments made by the above authority as 19.563 for the lowest, and 25.770 for the highest quality. The mean breaking weight across the fibre, in tons per square inch, being respectively 21.010 and 27.490. The tensile strain of wrought-iron plates, according to Mr. Fairbairn, was as follows : - With a thickness of plate of 1½ inches, the mean breaking strain per square inch in tons was 24.453, where the plate was 3 inches thick the strain was 25.031. The compressive strain of plates of the above dimensions show a mean ultimate pressure per square inch in tons of 90.967, and the mean ultimate compression per unit of length was .513 and .511.

As already stated, wrought-iron is not often used in the form of columns to resist compression, Mr. Hodgkinson states that the strength of square pillars of wrought-iron, long enough, became bent without the material being much crushed, is nearly as the 3.59 power of the lateral dimensions, or as d 3.59, where d is the side of the square, the length being constant. The law of resistance of wrought-iron is not widely different from that arrived at in my experiments with cast-iron pillars, the mean from pillars of this material being 3'6 nearly. From numerous experiments in the comparative strength of wrought and cast iron to bear pressure in the direction of the length, to determine the pressure which wrought-iron would bear as a column, "I find," says Mr. Hodgkinson, "that beyond 12 tons per square inch, it was of little or no use in practice." The experiments further show that cast-iron was decreased in length about double what wrought-iron was, of the same weight; but the wrought-iron sank to any degree with little more than 12 tons per square inch, whilst cast-iron required twice, or perhaps three times, the weight to produce the same effect.

We have in a preceding chapter given illustrations of the different forms of iron beams now generally used; we now give a formula adapted for calculating the breaking weight, of the form in which the lower flange is of greater breadth than the upper is that given by Mr. Hodgkinson, in which the "constant" was taken at 25, a representing the sectional area of the bottom flange in square inches, d the depth of the beam in inches, and l the length of the span or the bearing of the beam on the opening, also in inches, and W the breaking weight in the centre; W =a x a x c/ l in the approved form the sectional area of the top flange is equal to \ of that of the bottom. A good proportion between the span or length of the beam, between the bearing and its depth, is 1/12 the following is given by a high authority as a sound practical rule for proportioning the parts of iron beams : - "Take the breaking weight at from two to three times the weight estimated to be carried by the beam, then assume the depth of the beam at about 1/16th part (we have above stated it at 1/12th) of the distance between the supports for ordinary cases, and the sectional area of the bottom flange may then be found by the following proportion: as the depth, in feet, of the beam in the middle is to the distance between the supports, in feet, so is the - 1/26 part of the breaking weight, in tons, to the sectional area of the bottom flange, in square inches. Make the thickness of the bottom flange 1/12th the depth of the beam, and find the breadth by dividing the area by the thickness. Make the area of the top flange 1/5th part (we have above stated it to be 1/6th) of the area of the bottom flange, and half its breadth. Make the thickness of the web of the beam rather more than half the thickness of the bottom flange for the beam when cast, that the pattern may be made somewhat thinner.