EXPERIMENTS UPON WROUGHT-IRON GIRDERS. 28? 



Railways. Since that time some thousands of bridges have been built en- 

 tirely of wrought iron. The introduction of a new material, and the uncer- 

 tainty of its durability, led the Board of Trade to determine that the strain 

 should not exceed 5 tons per square inch on any part of the structure. 

 These requirements appeared to be founded on no fixed principle ; and the 

 bridge recently erected across the River Spey having been objected to as 

 not in accordance with this standard, it was resolved (with the consent and 

 at the expense of the Board of Trade) that the question whether the conti- 

 nuous changes of load, and the vibration by which they were accompanied, 

 did or did not lead to fracture. This was the question for solution ; and the 

 experiments now recorded have in a great measure determined to what extent 

 bridges of this kind can be loaded without incurring danger from fracture. 



It is well known that the power of resistance to strain of wrought-iron 

 plates in combination depends upon the principle on which they are united ; 

 and unless the parts are permanently established, the five-ton tensile strain per 

 isquare inch might lead to error. For the purpose of ascertaining the effects 

 of the changes of load and vibration causing rupture, a small iron-plate beam 

 of 20 feet clear span, and 16 inches deep, representing the proportions of one 

 of the girders of the Spey Bridge, was constructed, and exposed to strains and 

 conditions similar to those produced by the passage of heavy trains over a 

 girder bridge. 



The beam, as already described (page 46 of the Report of the Oxford 

 Meeting), was first loaded with one-fourth its breaking weight, and with this 

 load it sustained about one million changes without injury. The load was 

 then increased to nearly one-half the breaking weight, when it broke after 

 5175 changes. From this it appeared that bridges were not safe when loaded 

 to one-half the weight that would break them*. Having arrived at this 

 result, the beam was taken down and repaired, and the experiments renewed 

 with two-fifths the breaking weight, when 158 changes were made to bring 

 the parts repaired to their bearing. The load was then reduced from 4-6785 

 tons to 3-54 tons, when 25,900 changes were effected. After this the load 

 was again reduced to 3 tons, one-fourth the breaking weight, when 3,150,000 

 changes were recorded. Ultimately the load was increased to 4 tons, or 

 one-third the breaking weight, when it broke by tension across the botfcom 

 flange after sustaining 313,000 changes of that load. 



In calculating the strain on the area of the metal after deducting the rivet- 

 holes, which, it must be remembered, were larger in proportion in the small 

 beam than in bridges full size, it was ascertained that the beam suffered no 

 deterioration with strains of 7^ tons per square inch ; but with 10 tons it 

 broke with only 51 72 changes, as may be seen in the following Tables of ex- 

 periments. 



Table IV. — Beam repaired. 



The beam broken in the preceding experiment was repaired by replacing 

 the broken angle-iron on each side, and putting a patch over the broken 

 plate equal in area to the plate itself. Thus repaired, a weight of 3 tons was 

 placed on the beam — equivalent to one-fourth of tiie breaking weight ; that is, 



Lever 4470 lbs. 



Shackles 74 „ 



Half weight of beam 1815 „ 



Scale and 8 lbs 4'34 „ 



6793 

 With this weight the experiments were continued as before. 

 * See Report of Thirtieth Meeting, page 48. 



