424 . KEPOET— 1882. 



handle plans and drawings well knew. Similarly, a thin sheet of iron or 

 steel, bent into a tubular form, without further stifiening, offered as high 

 a resistance per square inch to compression as the most heavily braced 

 rectangular strut. The author recently tested a piece of ordinary stove- 

 pipe, 4 inches in diameter and 2 feet long, made of sheet iron only about 

 a fortieth of an inch in thickness, and found it stood, without buckling, a 

 compressive stress of ]5"9 tons per square inch, whereas one of the Britannia 

 Bridge rectangular cells, 18 inches square and 8 feet long, made of plates 

 and angles half an inch thick, crippled under a stress of 136 tons, or say 

 15 per cent, less stress than that sustained by the piece of stove-pipe. If 

 flat plates, as much as half an inch in thickness, behaved so badly in a 

 rectangular cell only 18 inches square, it is hardly necessary to speculate 

 as to what would be the case in tubes 12 feet square, which would be the 

 size required for the Forth girders. With rectangular struts formed of 

 four corner-pieces and lattice sides, the required strength of the lattice- 

 work has been found by experiment to be considerably greater than theory 

 would indicate, and the form, therefore, is a very disadvantageous one as 

 compared with a circular cross-section where eveiy particle of metal per- 

 forms useful work. In a long span bridge, it is essential to reduce the 

 secondary bracing to a minimum, because the weight of metal itself con- 

 stitutes the chief load. For that and for many other reasons, including 

 the comparatively small resistance offered by a curved surface to the wind, 

 the author, after carrying out not a few experiments himself, and passing 

 in review the numerous experiments made by others during the past thirty 

 years, was satisfied that a circular form of cross- section was the proper 

 one in the case of the Forth Bridge. All the main compression members, 

 therefore, will be tubes varying in diameter fi'om b feet to 12 feet, and only 

 the wind bracing, subject to alternate compressive and tensile stresses, 

 will consist of rectangular latticed members. 



Between the two main girders, as described above, the doiible line of 

 railway will be carried on an internal viaduct, supported by trestles and 

 cross- girders. The permanent way will consist of heavy bridge rail.«? 

 on longitudinal sleepers bedded in four steel troughs, the outer pair of 

 which serve also as the top members of the girders of the internal viaduct, 

 the inner pair being simply rail-bearers. The width of trough is such 

 that, in the event of derailment, the wheels will drop into the troughs, 

 and run along the timber sleepers clear of obstruction. A buckle plate 

 floor and parapet or wind screen will be provided of ample strength to 

 ensure the safety of the trains. No guard rails will be introduced, as the 

 engineers and the Board of Trade are in accord in considering them a 

 source of danger during high winds. 



It is hardly necessary to state that the whole of the superstructure will 

 be of steel. For the tension members the steel used is to have an ultimate 

 tensile strength of not less than 30 tons nor more than 33 tons per square 

 inch, with an elongation of 20 per cent, in a length of 8 inches. For the 

 compression members the strength is to be from 34 to 37 tons, and the 

 elongation 17 per cent. In making the tubes and other members, all plates 

 and bars which can be bent cold are to be so treated, and, where heating 

 is essential, no work is to be done upon the material after it has fallen to 

 a blue heat. The steady pressure of hydraulic presses is to be substituted 

 for hammering wherever practicable, and annealing will be required if the 

 steel has been distressed in any way. No punching or shearing will be 

 allowed, and all plates will be planed at the edges and butts, and all holes 



^ 



