28 PECULIAR PROPERTIES OF IRON AND STEEL. Art. 25 



being equivalent to 100% of the live load stress. In specifica- 

 tions of the last type, there is practically but one working unit 

 stress, while in others there may be many ; in one case the impact 

 is allowed for by varying the impact stress, and in the other (to 

 some extent) by using different unit stresses for different parts 1 . 



The usual factors of safety for iron and steel structures are 

 4 and 5, being 4 for buildings in general and for ordinary high- 

 way bridges. For cast iron the factor of safety is usually taken 

 as 8, in columns. The factor of safety for timber is 4 to 10, 

 depending on the kind of stress 2 



In any case, the factor of safety should be suck as to bring 

 the working stress well within the elastic limit. The important 

 point is to have an equal margin of strength for all parts of a 

 structure. 



25. Peculiar Properties of Iron and Steel in the Non- 

 elastic State. When stressed beyond the elastic limit, iron and 

 steel show some very interesting qualities. A stress which pro- 

 duces an appreciable permanent set in a body is said to overstrain 

 it. In order to understand the phenomena of overstrain, it 

 should be remembered that with increasing ultimate strength in 

 steel, there goes decreasing ductility, that is, less elongation. Steel 

 of high ultimate strength is called hard steel. The ultimate 

 strength of steel, in manufacture, is chiefly regulated by the 

 pecentage of carbon the more carbon, the harder the steel. Steel 

 is also made harder by heating and quenching in a cooling bath. 

 It has been found that overstrain also hardens steel in a remark- 

 able way, but impairs elasticity, and if repeated often produces 

 "fatigue." Rest and heat restore elasticity. 



26. Shifting of the Elastic Limit. If a bar of steel or iron 

 be stressed beyond the yield point, the stress removed, and then 

 immediately applied again, it will be found that the elastic limit 

 has been very much lowered but that the yield point has been 

 raised, and is now equal to the maximum previous stress. If, 

 however, a period of rest intervenes .between the loadings, the 

 elastic limit will also rise (time being an important factor) and 

 may even reach the new yield point ; the yield point also rises a 



! For a comparison of the main features of a number of railway 

 e siwoifioations see an article by the author in Engineering News, 

 Vol. 50, p. 444. 



*See "Cambria." 



