Appearance of Fatigue Failure The structural engineer has abandoned cast iron as a constructional material largely because of its lack of ductility in the tensile test and demands nowadays that all metals used for load carrying parts should possess appreciable ductility. Ductility is of course an essential requirement for fabrication but with the use of ductile metals has grown up the idea that a structural part will exhibit appreciable deformation and so give warning before it fails. This is true enough if failure is due to an overload once accidentally applied but large numbers of failures in service occur every year which give no such warning, are not preceded by any visible deformation even in materials with 20 per cent. or more percentage elongation, and are not caused by loads in excess of those for which the part has been designed. The fracture appears as a crack, that is a clean break, and its surface is generally of smooth, velvety appearance. An example of two such cracks in one welded beam is shown in Fig. I. All such failures are caused by a very large number of load applications, often all of them in the range of permissible stresses. The phenomenon is known as “Failure from Fatigue.” R. Weck
Commenting on Mr. S. P. Banerjee’s contribution to the discussion, Professor A. L. L. BAKER writes: On page 24, reference 21, it is stated that moments M b and M b + M c are approximately proportional to their corresponding deflections. The closeness of the approximation depends on the degree of similarity of shape of the bending moment diagrams. When the ratio of Y b / M b is greater than Y c / M c which is usual on account of the relative shapes of the M b and M c diagrams, the approximation is on the safe side. Even in special cases, when this is not true, the vagueness of K values for the soil and their distribution, and EI values for beams and their distribution necessitates the use of extreme safe limiting values, which justifies using the above convenient approximation.