B24 REPORTS ON THE STATE OF SCIENCE, ETC. 
them, so that E can be determined from a test-piece of this form from the load- 
extension diagram with an accuracy sufficient for all practical purposes. 
2. The influence of the flanges is to slightly reduce the ultimate extension in com- 
parison with the extension of a plain bar of equal gauge length. In the soft steel 
used extension may be taken at 3 per cent. less, allowing a plus or minus error of 
1 per cent. for variation of the quality of the material. 
I Dotted Flanges 
= Peal as made 
- Mo. thick 
ey 
en et 
Bod 
pat} 
y 
H 
ih 
(b) The next step was to investigate the validity of the law of similarity when 
applied to a flanged test-piece. 
Test-pieces were prepared in which the gauge length was eight times the diameter. 
The longest gauge length was 6 in. and the shortest was 2 in., the corresponding 
diameters being } in. and } in. Strict similarity was departed from only in the 
diameter and thickness of the flanges, which were | in. diameter and 74 in. thick in 
each specimen. ‘The test-pieces were cut from the same bar and all were annealed at 
900° C. before turning. 
Having regard to the slight difference in quality of the bar from outwards to its 
core, and remembering that the quality of the core is practically concentrated in the 
Extension on 6” 
a 
c 
o 
= 
= 
ao} 
0 
3 
4 
Length between Flanges = 8 diameters 
0 10 20 3 
' Extension —per Cent 
Extension between Flanges 
Fia. 10. Fie. 11. 
smaller test-pieces, the results of the experiments show that each bar gave the same 
percentage elongation within the limits of experimental error. 
Further details of this test will be found in my book on the ‘Strength and Structure 
of Materials,’ from which fig. 9, which shows a series of bars, fig. 10, a series of load- 
extension diagrams, and fig. 11, a set of corresponding stress-extension diagrams, 
are taken. 
