COMPLEX STRESS DISTRIBUTIONS IN ENGINEERING MATERIALS. 213 
Comments by F. Rogers. 
It is worth noting at the outset that, rather differently from us, the French often 
mean by ‘fatigue’ what we would call simple overstrain ; e.g., such as occurs in a 
tensile test piece strained beyond the elastic limit. In this particular series of 
articles, however, this use of the term has not occurred considerably. Further, 
there seems to be no reference to any question of an ‘ ageing’ effect in metals, either 
under steady stress or in the absence of stress. 
Most writers, as is to be expected, agree explicitly or implicitly in answering (1) in 
the affirmative. 
There is no agreement as to unconditional scrapping after a specified life (measured 
in time, or else number of stressings). This is as one would expect, since in most cases 
the aim is to design for permanent use, except where wnavoidable wear is concerned. 
Railway cranks and axles are, however, used for a definite mileage only. 
I consider that the suggestions made by Grenet are the crux of the problem, 
although they are not individually novel. There is room for more research on resilience 
from this point of view, particularly the values of elastic modulus in relation to en- 
durance of repeated stress and repeated shock, and the safe limit of repeated shock. 
I have shown (No. 62) the increase of elastic modulus with brittleness due to 
overheating of steel. 
Periodical annealing during the useful life is only applicable to certain materials 
and to certain forms. No large pieces, such as shafts, can be so dealt with, on account 
of distortion and scaling. Small iron articles, such as chains, are annealed. I have 
shown that at a certain stage annealing is in any event incapable of restoring a mate- 
rial. The original heat treatment of steels is often such that their properties would 
be hopelessly ruined by any process which could fall under the vague term ‘ annealing.’ 
Springs are, however, sometimes annealed and re-tempered ; it is questionable whether 
any advantage is gained. 
The microscope is a valuable aid to research, but it is only in exceptional cases 
that it is of assistance in finding hair-cracks in existing structures. In my experience 
the hair-crack period is of short duration, and tests to destruction are the most reliable 
index of the state of the material; but in certain cases it would be worth while to keep 
vital points polished ready for periodical examination (and varnished), and approach- 
ing fatigue could then be satisfactorily detected at a considerably earlier stage than 
the appearance of hair-cracks. 
In regard to Guillet’s dictum that the original treatment of the metal is usually to 
blame when metals fail under alternating stress, I do not think it is possible to arrive 
at a general conclusion ; that is to say, each of the following three main classes contain 
the usual sources of trouble, and many cases in practice have been traced to eaca of 
these causes :— 
(a) Flaws, including pipe, fissures, blow-holes, impurity, and non-metallic en- 
closures. 
(b) Faulty original heat-treatment of pure metal. This includes, as a special case, 
strains set up in manufacture, and overwork in the working processes. 
(c) Under-estimation of stresses to be expected on the part of the designer. This 
includes, as a special case, insufficient allowance for the effect of repetition 
of a stress which would be harmless if applied once or steadily maintained. 
SPECIAL PROBLEMS. 
The Resistance of Tubes to Collapse. By GitBERtT Coox, M.Sc. 
(The small figures in the text refer to the bibliography.) 
When a thin cylindrical tube is subjected to a gradually increasing 
external pressure, a point is reached at which the equilibrium becomes 
unstable, any further increase in the pressure resulting in the collapse of the 
tube. This pressure is known as the collapsing pressure. The subject of its 
determination is one which has, from time to time, received considerable 
attention both from mathematicians and engineers. Yet, in spite of the 
