UNIFICATION OF METHODS IN TESTING MATERIALS. XXXV. 
establish a fixed relation between the transverse section and the 
useful, or test, length of the piece. This relation, deduced by the 
law of similarity, shewed that the test length should be proportional 
to the square root of the cross section ; this law, as the author 
states, was agreed to by the convention, but it was not pointed 
out that the convention and the French Commission differed as to 
value of the coefficient, but the American Society ignored these 
formule, in as far, that they preserved a constant of 8” for the 
length of test piece. The French formula was Z = 8-16 vs, whilst 
the convention formula was J =11-3 vs; the former gave a diam. 
of 27:64 millimetres for a test length of 200 millimetres,’ whilst 
the latter gave 20 millimetres for same test length, or a propor- 
tion of 10 to 1 as stated by the author : the latter formula recom- 
mended itself on account of its simplicity due to the above men- 
tioned proportion, but the former or French formula had this 
advantage, that the area, the first item deduced in using the 
formula, was a simple number (in the present case 600 square 
millimeters). Now in order to simplify calculations, the adoption 
of a simple number to express the area of cross section which is 
the function used in calculations of strength, was more advisable 
than simplicity in expressing the diameter, the difficulty of 
measuring which was always the same. The American Society 
on the other hand used the dimensions credited by the author to 
the Conventions or rather European standards. 
The French dimensions were as follows—test lengths of 70, 100, 
141, and 200 millimetres ; cross sections of 75, 150, 300 and 600 
square millimetres; and diameters deduced from above of 9°77, 
13-82, 19-55, and 27-64 millimetres: the convention standards 
were those where the proportion of length to diameter is as 10 to 1, 
the length ranging from 100 to 250 millimetres. The American 
Society, as previously mentioned, preserved a constant test length 
of 8 inches, thus ignoring the difference which must occur in the 
measurements of total elongation, by testing with a uniform 
length and variable diameter. This length of 8 inches was chosen 
1 200 = $16 Vs . vs = 24°51 s = 6000 mm. d = 27— = 27°64 mm. 
