AbstractData are presented from fatigue and creep–fatigue crack growth rate tests carried out at 650°C on an aged type 347 stainless steel weld. Fully reversed tests were carried out under displacement control, with constant displacement hold periods, to simulate closely the thermally induced straining relevant to power plant applications. Tensile hold periods of up to 192 h were used in the creep–fatigue tests; a 1 h hold period was sufficient to induce internal intergranular cracking in the test specimens. In the creep–fatigue tests, crack growth occurred during both the cyclic excursions and the hold periods. Cyclic rates were also dependent on hold time, the acceleration in fatigue crack propagation rates being associated with the extent of creep damage produced ahead of the main crack as a result of the hold periods. Such an interaction between fatigue and creep was supported by a sequential creep–fatigue crack growth experiment. Empirical equations were obtained which describe the total crack growth rates as the sum of a cyclic and a dwell contribution. The cyclic contributions were described in terms of a stress intensity factor range, while the dwell components were correlated with a creep crack growth correlating parameter C*. It was shown that the reference stress approximation to C*accurately predicts measured values as long as creep strain rate data were derived from constant displacement stress relaxation data rather than from forward creep data.MST/782