New mechanistic interpretations to rationalize fatigue crack growth retardation due to load excursions are presented. It is reasoned that crack closure arising from residual tensile displacements is not the primary mechanism for growth attenuation following a peak tensile overload.A new mechanism for retardation is discussed in terms of a "micro-roughness" model. Quantitative analyses are provided to estimate the extent of reducti ons ineffective dri vi ng force in the reta rded growth regi on due to possible crack branching, residual compressive stresses and fracture face micro-roughness. It is argued that the retarded crack advance is effectiveLy governed by the micromechamisms of Stage I growth although nominaLLy Stage II conditions exist in the post-overload zone.The impl i cations of the present arguments are shown to be consi stent wi th a number of typi cal post-overload phenomena cited in the 1 iterature .= crack length = total retarded crack growth distance = delay distance (retarded crack advance up to the lowest growth ra te ) = baseline crack propagation rate = retarded crack propagation rate = average asperity height = height of the specimen = effective stress intensity factor for branched crack = Mode I stress intensity factor for branched crack = Mode I stress intensity factor for main crack = Mode II stress intensi ty factor for branched crack = Mode II stress intensity factor for main crack = closure stress intensity factor = closure stress intensity factor accounting for asperity contact = baseline maximum stress intensity factor -overload maximum stress intensity factor = baseline minimum stress intensity factor = stress intensity factor due to residual stress field = value of KIR during baseline (pre-overload) crack growth = value of KIRfollowing an overload = number of cyl ces for retarded crack growth