424 THE SCIENTIFIC PAPERS OF 



confess at first appeared hardly credible to him, had since been 

 fully confirmed by experiments of his own. It was a remarkable 

 fact for engineers to consider that in constructing a bridge it came 

 to the same thing whether they took mild steel that would break 

 with 28 or 30 tons strain to the square inch and would come to its 

 elastic limit at about 1 5 tons, or whether they took a material that 

 would find its elastic limit only at 25 or 30 tons, and would break 

 at 50 or GO tons ; therefore the advantage that would be derived 

 from the hard and strong material would only be met with after 

 exceeding the elastic limit of the milder steel. He was not aware 

 of any experiments giving the influence of heat upon the coeffi- 

 cient of elastic extension of those materials. If it should so happen 

 that a bar of iron or steel at the temperature of 5 would follow a 

 less rate of elastic extension than a warmer bar, a steel bridge that 

 was ordinarily loaded to a strain of 8 tons per square inch, when 

 it cooled down, would contract rather more than in the ratio of 

 simple thermal contraction ; and afterwards, if the bridge changed 

 from the one temperature to the other, it would alter its length in 

 all its members due to the two ratios of absolute contraction and 

 a contraction due to a different rate of extensibility, It would be 

 interesting to know how the modulus of elasticity was influenced 

 by temperature. It was often thought that metals were fatigued 

 by strain ; and in the case of inferior classes of iron, it might be 

 true that, if a bar were strained a little beyond its elastic limit, its 

 fibres would be ruptured to some extent, and the material would 

 receive an injury ; but in the case of high-class material, such as 

 mild steel, there was an absolute advantage in straining it beyond 

 the limit of elasticity. It was possible, by careful manipulation, 

 to raise the breaking strain of a bar of a given sectional area to a 

 remarkable extent by gradually accustoming it to the strain. By 

 taking a bar of mild steel of 1 square inch sectional area and load- 

 ing with a weight of, say, 15 tons, and leaving the weight on 

 twenty-four hours, it would be found that both the elastic limit 

 and the breaking strength of the bar were materially increased. 

 That was a quality of high-class material which was, perhaps, not 

 yet thoroughly understood, and which certainly gave a great factor 

 of safety in favour of its employment. The chemical analysis, 

 which had not been dealt with in the paper, was of great import- 

 ance in reference to the object the author had in view. The 



