MECHANICAL TESTING OF IRON AND STEEL. 225 



produce very serious results, perhaps deciding into which of 

 two important groups of results a particular experiment 

 is to be placed. 



When it is remembered, that apart from inertia effects, 

 the instantaneous stress produced by a sudden load, if it 

 does not exceed the limit, may be twice as great as the 

 static stress produced by the same load when applied 

 gradually, it will be conceded how important the point is 

 in such calculations ; the real stress due to the load may 

 be well within the limit of natural elasticity, and yet the 

 instantaneous stress set up may exceed it, and thereby 

 when often enough repeated seriously change the mole- 

 cular condition of the bar. 



This point is dwelt upon, because it seems to be too often 

 forgotten when discussing the effects of repeated loads, and 

 the results they produce in parts of machines or structures, 

 and also in estimating the so-called factor of safety. 



The lines therefore, to sum up, upon which further 

 scientific testing should proceed are two : {a) The deter- 

 mination of the real natural Limits of Elasticity of different 

 qualities of iron and steel, and the influence upon it of the 

 various chemical constituents, and the way in which it is 

 altered by various methods of treatment, that is, the question 

 of artificially raised or lowered limits. In this connection it 

 may be interesting to remember that the Modulus of Direct 

 Elasticity is the one function which seems most constant, no 

 matter how the quality of the iron or steel varies, still it 

 does seem to change as the carbon percentage is increased, 

 being gradually diminished in value as that percentage rises. 

 May it not, therefore, turn out that there is a direct 

 connection between the values of the natural Limit of 

 Elasticity and the Modulus of Elasticity ? 



(b) The study of the effect of repeated loads, continued 

 on the lines of Wohler, Spangenberg, Bauschinger, and 

 others, and in particular the solution of the question as to 

 what is the molecular change induced by repeated loading 

 beyond the range of the Limit of Elasticity, and why it is 

 that the broken parts of the bar when tested at once in an 

 ordinary testing machine show apparently no molecular 



