XXXVI. DISCUSSION. 
by the American Society, as well as the decimal part of an inch 
diameters, for the purpose of approaching as nearly as possible 
the measures of the metric system. : 
As regards the accuracy of machines and errors permissible, 
the American Society made no recommendation, but the French 
adopt a rule not quite so exacting as that suggested by the Con- 
ventions. 
As regards the rate of testing, it was the French Commission, 
not the American Society, who considered that the duration of 
the test should be, in a certain measure, a function of the volume 
of the test piece; the Convention, whilst making no recommenda- 
tion on this subject, strongly urged the necessity of noting with 
what rapidity the diagram was traced. ‘The resolutions of the 
Convention and the American Society provided for the use of 
diagrams similar to that exhibited by the author, but the French 
Commission did not consider them necessary, at least in current 
practical tests with a view to ascertaining the quality of the 
metal tested from a determination of the useful surface they pre- 
sent, and this is the opinion of some English scientists, amongst 
others Prof. Unwin. This remark of course only applies to 
diagrams taken by one or other of the autographic recorders. He 
uses diagrams however, which are plotted from the known stresses 
and the various observed elongations. 
Under the heading of compression, the author for the first time 
referred directly to the results of the French Commission; the 
first recommendation was correct, as far as the elastic limit is 
concerned, but the mention of cubes or prisms of 25 millimeters 
a side should have been given as a recommendation of the Con- 
vention and not of the French Commission, whose recommendations 
on this head were that the diameters be reduced to 19:56 milli- 
meters equal to a cross section of 300 square millimeters, and the 
useful length to 20 millimeters. As regards tests on long pieces, 
which fail by buckling, the Convention had issued no instructions, 
but the French Commission recommended the requirements 48 
stated by the author, but the reason for which he omitted, said 
