To develop further phenomenological understanding of load interaction effects in fatigue, examinations of the influences of stress intensity (K) level, plate thickness and chemical environment on fatigue crack growth response following a single high-load excursion (overload) were carried out on a 2219-T851 aluminum alloy. An overload ratio (that is, the ratio between the magnitude of the overload and the maximum in the subsequent constant-amplitude fatigue loading) of 2.0 was used. Experiments were carried out in dehumidified argon, air (30 to 70 percent relative humidity), and3.5percent NaCl solution at room temperature. The results showed that delay (as measured by the duration of overload affected crack growth) increased with increasing K level and with decreasing plate thickness, and decreased with increasing aggressiveness of the chemical environment. The high-load excursion (overload) affected crack growth through a region of material ahead of the crack tip. Within this overload affected zone, crack growth rate first increased (sometimes), followed by fairly rapid decrease to a minimum value (delayed retardation), and then increased gradually to its steady-state value. The overload affected zone size was found to depend on K-level, and on crack-tip constraint, and to be independent of chemical environment, and was found to be equal to the appropriate (plane-strain or plane-stress) plastic zone size for the overload. Identification of a delayed retardation zone was made, and identification of this zone with the cyclic plastic zone size for the preceding fatigue loading was suggested. The effects of K level, plate thickness and chemical environment on delay were considered in relation to their respective influences on the overload-affected-zone and delayed-retardation-zone sizes, and on the rate of fatigue crack growth. A residual stress intensity concept for describing fatigue crack growth response within the overload affected zone was considered. With suitable modifications, reasonable estimates of crack growth response could be obtained. Further need for verification and understanding of these modifications are discussed.