322 REPORTS ON THE STATE OF SCIENCE, ETC. 
of universal use, it would not only be of scientific value to do so, but also be of com- 
mercial use. Possibly one of the simplest ways to effect this without interfering too 
much with existing testing machinery would be to remodel each briquette by the 
addition of a short parallel part to the waist of sufficient length to enable the 
central part to develop a purely tensional stress in it. 
Photo-elastic experiments show that the minimum length required in the case of 
the British Standard is 0°88 in., and cement briquettes of this shape, when tested, 
indicate by the uniformity of lateral contraction across the central part that pure 
tension stress exists at this cross-section, so that the mean average stress and the true 
stress are one and the same. 
Further exploration might possibly suggest other and better ways of solving this 
difficulty of the want of uniformity in standard methods of testing cement and other 
materials, which can only come about by the research and co-operation of the repre- 
sentative bodies now existing in most countries. 
In this brief survey of one branch of an extensive field of research, it has not been 
possible to give detailed accounts of the experimental and theoretical data utilised, 
and for this and other information the reader may be referred to the following papers, 
some of which have been quoted above. 
APPENDIX. 
(1) ‘The Determination by Photo-Elastic Methods of the Distribution of Stress 
in Plates of Variable Section, with Some Applications to Ships’ Plating.’ E. G. Coker, 
Trans. Inst. Naval Architects, 1911. 
(2) ‘ Photo-Elastic Measurements of the Stress Distribution in Tension Members 
used in the Testing of Materials.’ E.G. Coker, Min. Inst. Civil Engineers, 1918-19. 
(3) ‘ Tension Tests of Materials.’ E. G. Coker, Engineering, January 1921. 
(4) ‘Contact Pressures and Stresses.’ Prof. EK. G. Coker, K. C. Chakko, and 
M. 8. Ahmed, Proc. Inst. Mech. Engineers, 1921. 
(5) ‘On the Approximate Solution for the Bending of a Beam of Rectangular 
Cross-Section under any System of Load.’ Prof. L. N. G. Filon, Trans. R.S., Series A, 
vol. 201. 
(6) ‘ The Effects of Scratches in Materials.’ EH. G. Coker, Proc. of the Engineering 
Conference of the Institution of Civil Engineers, 1921. 
(7) ‘Stress Concentration due to Notches and like Discontinuities.’ Prof. E. G. 
Coker and Dr. Paul Heymans, B.A. Report, 1921. 
(8) ‘The Distribution of Stress at the Minimum Section of a Cement Briquette.’ 
E. G. Coker, International Association for Testing Materials, New York Congress, 1913. 
(9) ‘Des Recherches Récentes sur la Photo-Elasticimétrie ayant rapport 4 son 
application dans les Problémes Posés en Construction.’ E. G. Coker, La Sociéié des 
Ingénieurs Civils de France, 1922. See also Hngineering, June 16, 1922. 
II. A Standard Form of Test-Piece. 
By Professor W. E. Dausy, F.R.S. 
1. The Flanged Test-Piece. 
During my researches on the elastic properties of metals, I found that the best 
form of test-piece to use was one in which the gauge length is defined by flanges, as 
shown in fig. 8. When the gauge length is defined by centre dots on the test-piece, 
clips have to be used, each carrying a pair of pointed screws, which are driven into the 
centre dots, and the points of the screws deepen the dots, or, better, make their own 
centre dots. The pointed screws must be spring-loaded in order that they may 
maintain their grip on the test-piece as its diameter contracts under test. 
I have tried many forms of these clips, and have found some fairly satisfactory, 
but no form of spring-loaded pointed clip is so satisfactory as flanges turned on the 
test-pieces themselves. 
The practical disadvantage of clips for elastic tests is that the points cannot be 
satisfactorily driven into hard steel, and in soft metal, like copper, the pits made by 
the points elongate as the test-piece stretches. The single advantage of using clips 
is that the extension is measured on the central portion of a bar of uniform diameter. 
But when the extension is small, as in elastic testing, the flanges have a negligible 
jnfluence on the result, and therefore the elastic modulus determined from a gauge 
