

to Transverse Strain in Beams. 427 



strain, and without the necessity for any other resistance to 

 flexure. 



And considering that if we regard an ideal vertical section 

 through the centre of the beam, the whole of the resistances 

 must be represented by the molecular forces between the 

 particles on the opposite sides of that section, it is obvious 

 that these are practically none but ordinary tensile and com- 

 pressive stresses "*. 



It is, however, only by considerable overstrain at the lower 

 side of the beam that these forces can be enabled to supply 

 the resistance which the facts prove to exist. And it is well 

 known, as has already been observed, that an overstrain is 

 always followed by more or less set after the pressures causing 

 it are removed. And it is also well known that frequent 

 alternations of strain and release cause such a modification of 

 the internal structure, that with the lesser strains the variations 

 between the proportions of stress and strain are reduced 

 and the elasticity becomes more nearly perfect. And with 

 the greater strains the probable effect will be of a similar 

 nature. 



The capacity of the material for overstrain will be reduced 

 without any increase of the tensile strength ; and if the extreme 

 overstrains were frequently repeated, it is more than probable 

 that the beam would be broken with a much less load than it 

 supported without breaking on its first trial. Such an 

 experiment may perhaps never have been made as to subject 

 a beam repeatedly to anything near the breaking load. But 

 the vibrations to which a beam is exposed when employed in 

 a railway-bridge probably cause intermolecular strains of as 

 great an amount with a similar change in the internal 

 structure, and the failure of such beams with loads no greater 

 than they have frequently borne before is unfortunately too 

 common a circumstance. 



The obvious conclusion is that the excess of strength for 

 which Mr. Barlow, and it may be added the present writer, 

 has been at such pains to account, and which, taking Mr. 

 Barlow's theory for granted, practical engineers have treated 

 as a real and important element of security, should not be 

 regarded by them at all, but the efficient transverse strength 

 of a beam should be computed from the tensile strength by 

 the old method of calculation, and any extra strength exhibited 

 in the trial of new castings disregarded as transitory and 

 delusive. 



* There is probably also an infinitesimal resistance to flexure resulting 

 from the lateral displacement of the particles. This must, however, be 

 so extremely small as to be inappreciable in comparison with the re- 

 sistances arising from tensile and compressive stress. 



