222 SCIENCE PROGRESS. 



any delay in the work of testing. No other result obtained in 

 a tensile test shows more clearly, for instance, the utter un- 

 trustworthiness of wrought iron plate when stressed across 

 the grain if it is liable to shocks or sudden loads, or shows 

 more distinctly the superiority of steel plate in this respect. 



Limit of Elasticity. No term has given rise to more 

 confusion in dealing with the strength of materials than this. 

 One instance will suffice to show the kind of error produced 

 by false notions as to its meaning ; the well-known fact that 

 by stressing a bar in tension beyond its Yield Point we 

 raised its Limit of Elasticity in Tension was usually con- 

 sidered to represent the whole result of the action. We know, 

 however, from the researches of Bauschinger that this is a 

 very imperfect and misleading conclusion, a conclusion, too, 

 which may have caused much of the difficulty in understand- 

 ing some of the results obtained in the endurance tests of 

 Wohler. 



Bauschinger's experiments, carried out with perhaps 

 more exactitude than any previous experiments of this 

 nature, lead us to conclude that there is for any given 

 material a true natural Limit of Elasticity, understanding 

 by that term the Limit of the Elastic Condition according 

 to Hooke's law, nt tensio sic vis. But this natural Limit 

 may be varied in all sorts of ways, by strains set up in the 

 material during manufacture, by after working, or in a 

 testing machine. Unless, therefore, we know the whole 

 previous history of the bar we are testing we are quite 

 unable to say whether the Limit obtained in our test is 

 a natural one, or whether it is some artificial one pro- 

 duced by the treatment the bar has undergone during manu- 

 facture or afterwards. In one form this fact was appreci- 

 ated by all those interested in the question of testing ; it was 

 well known, for example, that cold rolling greatly raised the 

 Limit, two bars rolled from the same ingot would give very 

 different values for this factor, the one with the smaller cross- 

 sectional area and therefore during the last stages of manu- 

 facture rolled while partially cold would show a much higher 

 apparent Limit. But it was not, at any rate generally, under- 

 stood that such effects were only particular cases of a much 



