The creation of new populations in conservation reserves plays an increasingly important role in reducing the extinction risk of endangered plants. We experimentally introduced seedlings of the federally endangered forb Astragalus bibullatus into protected limestone cedar glades in central Tennessee and tested how source population, transplant season, and glade site affected the demographic performance of transplants over 7 years. Plants derived from seed from three source populations survived at similar rates at introduced sites, but exhibited different germination rates in greenhouse cultivation trials. Even though glade sites were seemingly similar in abiotic conditions and in close proximity (<400 m), survival rates varied widely among sites. Plants introduced in the spring never produced seed and were nearly four times more likely to die than plants introduced in the fall. Within 3 years of outplanting, 20% of plants introduced in the fall produced seed, yet sexual reproduction ceased after the third year and only one seedling recruit was observed after 7 years of monitoring. This resulted in small subpopulations that were not self-sustaining probably due to demographic stochasticity. In addition to highlighting the need for long-term monitoring of introduced populations to draw appropriate conclusions, our study emphasizes the importance of properly timing outplantings with the appropriate seasons and having multiple introduction sites as a hedge against environmental variability. However, overcoming seed limitation during the early stages of population growth and determining “safe sites” for seedling recruitment remains a central challenge in introducing endemic plants into seemingly suitable but unoccupied habitat.