UNIFICATION OF METHODS IN TESTING MATERIALS. XXXVII. 
reason being that that condition was all that was necessary to 
to make two tests comparable on bars or rivetted pieces. The 
American Society in this class of tests required that the process 
should be the same as in tensile tests dividing the useful or test 
length into small sections in such a manner that the loss in value 
Sustained, might be determined from its elements, and that the 
_ calculation of the modulus and coefficient of elasticity should be 
made under like conditions. 
As regards mixed tests (shearing and punching) the Convention 
was silent, but as stated by the author, the French Commission 
recommended further study ; the Convention only recommended 
that certain plates, such as those for boilers, of wrought iron, 
should be submitted to the punching test, but claimed that such 
a test was useless for plates of mild steel or ingot iron, neither of 
which should ever be punched. 
The author stated, inter alia, “it is proposed to consider the 
nature of the mechanical tests rather than the results obtained, 
and generally to deal with the subject in regard to the data which 
it is sought to obtain in the tests and the precautions to be taken 
in order to eliminate disturbing influences.” ‘The author then 
proceeded, under the clause headed “ Tension,” to enumerate the 
various points which it was essential should be accurately deter- 
mined in connection with the commercial testing of materials of 
construction, and enumerated under the sub-head (6) “the yield © 
Point,” adding in a further paragraph as apparently distinct from 
being essential that for accurate scientific testing the limit and 
coefficient of elasticity be also determined. 
On this point, Mr. Haycroft could not agree with the author, as 
he considered it was much more essential to know the limit of 
elasticity than it was to know the yield point. He was of course 
aware of the considerable difference of opinion which existed 
ee Mepaeaa ip ee to this preneh of the subject, and 
andv ful consideration. 
Tt was well known that great difficulty has always existed in deter- 
mining on a stress diagram the exact point where the limit of 
