LXIII. 



Owing to the uncertain nature of this report, the Swiss 

 Government determined to test to destruction a bridge of the 

 same age as the one which collapsed, and the same type, viz., 

 Warren or trellis in V, but of a somewhat stronger construction. 

 This w r as the single line railway bridge over the Emme, at 

 Wolhusen, of 157 feet span. During the test it was found that 

 under a loading of one and a third times the load used in 

 calculating the bridge stresses the limit of elasticity had been 

 exceeded. The working stress used in the calculation was 4J 

 tons per square inch. 



Ultimately, this bridge collapsed suddenly, through the 

 rupture of the connection between a compression bar and the 

 upper boom, when the test load was only twice that used in the 

 ■calculation. 



To sum up, if in the case of a riveted lattice truss, and one on 

 the Vierendeel principle, the stresses be arrived at by the 

 ordinary method of supposing the joints articulated, in the 

 former the stresses will be at least 60 per cent, less than the 

 actual stress ; whilst in the latter, the calculated stress will 

 e-qual or exceed by 5 per cent, the actual stress. This fact lends 

 great weight to Professor Yierendeel's design, as, whether the 

 latter be calculated on the false hypothesis of articulation or the 

 •exact method, certainty of result is attained, and thereby what 

 an engineer should aim at, security for the structure. 



Mr. J. Bowman gave results of an approximate comparison 

 between an ordinary plate girder bridge with parallel flanges, and 

 the test bridge mentioned in the paper, being altogether in 

 favour of the plate girder. The weight of bridge under discus- 

 sion (which failed under an external distributed load of 404 tons), 

 is given as 62,703 kilos = 61*14 tons. 



The weight of plate girder bridge for same span, and which 

 will not be stressed beyond the elastic limit of 10 tons per square 

 inch under the ultimate load of 404 tons above, is only 40 tons. 



It is obvious that by developing the parabolic upper flange 

 in place of the parallel flange assumed, the weight could be 

 reduced even below 40 tons. The depth of plate girder is great, 



