Stress can facilitate memory formation and retention via glucocorticoids (GCs), which are released at low baseline concentrations, but increase in response to a stressor. Exposure to increased levels of circulating GCs shortly after learning typically enhances long-term memory formation. Long-term memories last hours to weeks, whereas memories that persist for months to years are known as long-lasting memories. Despite considerable knowledge of the effects of GCs on short- and long-term memory, understanding of the effects of GCs on long-lasting memory is essentially nonexistent. The goal of this study was to understand the effects of GCs on the acquisition and retention of long-lasting memory. We asked: do circulating glucocorticoids affect the long-lasting memory of an emotionally arousing event? We exposed free-living Florida scrub-jays (Aphelocoma coerulescens; FSJs) to a modified fear conditioning protocol that used a novel ‘predator’ as the conditioned stimulus. Endogenous and manipulated levels of corticosterone (CORT; the primary avian GC) immediately after conditioning were compared to the FSJs’ long-lasting response to the novel ‘predator’ using flight initiation distance (FID). Our results indicate that CORT has an inverted U-shaped relationship with the retention of long-lasting memory with low and high levels impairing memory formation or retention. Additionally, FSJs with a moderate stress-response phenotype display an intensified conditioned response with time (i.e., FID increases over time), superficially similar to sensitization, but occurring after associative learning in the absence of reinforcement and over a much longer time frame. This suggests that FSJs can experience a type of memory modification that amplifies their original perception, which challenges the conventional wisdom that memory tends to either remain static or decay with time.