892 



ever small, for ever, but that the material is constantly 

 losing its sustaining power, or is, in fact, breaking by 

 imperceptible degrees. This idea is borne out by expe- 

 rience in cases where the weight sustained is not constant, 

 or does not press always in one direction ; thus, a spring is 

 somewhat broken every time it is bent. But such decrease 

 of strength either does not take place in regard to pressures 

 not approaching the breaking v/eight of the material, and 

 the direction of which is not changed, or so slowly that 

 it need not be considered. In fact, if such element of 

 calculation were admitted, we must make all permanent 

 structures infinitely strong. 



Experiments tend to prove, that in taking one-third of 

 the breaking weight as a safe quantity to be borne by 

 beams, is to err on the safe side ; and I am inclined to 

 think it would be abundantly sufficient that the breaking 

 weight of a beam, carefully ascertained, be equal to twice 

 the weight which would ever be put upon it ; but as vibra- 

 tion produces a very decided efi^ect, such limit ought not 

 to be passed. 



The general rule is, to make a beam at least three times 

 as strong as the weight it has to carry. This appears to 

 have arisen from an idea that one-third of the breaking 

 weight was the least weight which would produce a per- 

 manent defect of elasticity of a beam that is "a set." Mr. 

 Hodgkinson has, however, shown, that much less than one- 

 third the breaking weight produces a permanent set; that, 

 in fact, any weight laid on a beam produces a set, though 

 proportionately small. Any deduction from the production 

 of a set is fallacious. 



It appears by experiments made by Colonel Beaufoy, 

 at the Dock Yard, at Deptford, that the deflections of 

 beams of oak, fir, and pine, were very nearly in the ratio 

 of the weights, until about half the weights had been laid 



