, G.^-ENQIN BERING. 



133 



made, for example, to replace the very hard steel used at present 

 for connecting-rods by duralumin or some other forged aluminium 

 alloy. It seems worth trying ; but who, in our present state of ignorance 

 of tne real properties of metals, will say if the experiment will be a 

 success '! 



How difficult it is to prophesy may be illustrated by the results of 

 two empirical tests on durahuniu and steel sheets of the same thick- 

 nesses. The ultimate strengths and elongations of the steel and the 

 duralumin were roughly equal. The lantern slides show that under 

 reverse-bend tests they both follow the same law, the steel being the 

 better. But under the cupping test they follow opposite laws. 



The suitability of different materials presumably depends on their 

 fundamental physical properties. These may be many, but some 

 physicists think that they are probably really very few, and that, 

 knowing these few, it may be possible to deduce all the complex 

 properties required by the engineer and to state with certainty how 

 materials will behave under any conditions of service. This is the most 

 fundamental problem which needs solution to enable the knowledge 

 of the strength of materials to be put on a sound foundation. It will 

 need the co-operation of able physicists, metallurgists, and engineers 

 to solve it. 



While urging the importance of research in the fundamental theories 

 of stress and fundamental properties of materials, I wish to lay special 

 stress on the nature of the researches required. Engineers are intensely 

 practical men, and their practice has generally been ahead of their 

 theory. The difficulties they have met have been dealt with, often with 

 the greatest ingenuity and skill, as special problems. They have seldom 

 had time or opportunity to solve the general problems, and as a result 

 they are used to making their experiments and trials as close a copy — 

 usually on a smaller scale — of the real thing as possible. The results 

 obtained in this way, while they are applicable to the particular 

 problem, are of little general use. They depend on many factors. The 

 researches I am now advocating must be of a diametrically opposite 

 description. They must be absolutely general, and the results must 

 depend on one factor only at a time, so that general laws may be 

 established which will be applicable to all special problems. 



There are many other similar gaps in our knowledge to which I 

 have not time to refer to to-day. I have tried to show that we need 

 most of all a real knowledge of the fundamental properties of materials, 

 from which we shall be able to deduce their behaviour in any condition 

 of service, so that we may be able to compare the relative merits of 

 diverse materials for any particular purpose. 



Secondly, that we need a practical method of calculating the stresses 

 in parts of any form, so that concentrations of stress may be avoided 

 or that their magnitudes may be known and allowed for. 



Thirdly, that we need a rational connecting link between the tests 

 made on materials and the stresses they will bear in service, to replace 

 the factor of safety. I have suggested two tests, the Proof Load and 

 the Fatigue Limit, which might be used directly in estimating the allow- 

 able working stress. 



