Results are given of direct-stress random-loading and block-loading fatigue tests on sharply notched mild steel specimens subjected to a tensile mean stress. It is shown that the random-loading tests are more damaging than the multi-level block tests simulating the same conditions. Methods of endurance prediction in variable-amplitude conditions are discussed. In random-loading tests the hypothetical S-N curve method gives the better prediction provided that a suitable value of the slope parameter, d, is chosen; this value is not predictable from two-level block tests. The Palmgren-Miner method gave optimistic predictions. There was little difference between the two prediction methods when applied to the results of the multi-level block tests and the value of d suggested by two-level tests was satisfactory.

No abstract

No abstract

No abstract

No abstract