270 WILLIAM FIELD HOW. 
heat, would, by the heating and quenching process, become 
annealed, and after having been so treated, both tensile and 
bending tests would be satisfactory. 
If the plate be hard, a high tensile result and low elongation 
‘would be obtained and consequently show the material to be 
unsuitable for constructional works. 
Elastic Limit. 
When testing a specimen,: under a certain stress the metal will 
sensibly yield, and this “yield point” may, for all practical pur- 
poses, be considered to be the “Limit of Elasticity,” and is 
recorded as such by many experienced experimentalists ; but it is 
well known that the term “Limit of Elasticity” should be applied 
to the lowest stress to which a specimen can be subjected without 
creating a permanent set, and this is reached before the yield point. 
Material used for constructional purposes should have an elastic 
limit of at least twice the stress the member is designed to carry, 
but it has been found from experiments that if material is loaded 
frequently, its elastic limit and ultimate tensile strength are 
increased, and it is not, therefore, actually dangerous should the 
‘stresses occasionally exceed the original elastic limit of the material, 
but it is not advisable to so strain iron or steel, as by so doing, its 
‘ductility is decreased. 
Ultimate tensile strength and elongation appear to be sufficient 
recommendation of iron and steel for constructional purposes ; a 
high elastic limit can be obtained from hard and brittle material. 
In fact, the elastic limit varies with the treatment to which the 
material or specimens have been subjected before testing. 
Considerable inconvenience has been caused by some engineers 
adopting different standards of length in which to ascertain the 
elongation of specimens. Mr. David Kirkaldy, the pioneer of 
scientific mechanical testing, introduced the system of taking the 
elongation in a length of ten inches, and this was generally 
adopted until mild steel began to successfully compete with iron, 
when for some reason unknown to the author, eight. inehes — 
