Globally, conservation translocations are common often using island, or ‘virtual island’ populations as a source site. We investigated demographic and genetic consequences of using a single source island population for multiple translocations. Our harvest selection protocol preserves the source population whilst providing individuals that are representative of the source population. Our study species, Tasmanian devils, are endangered due to an infectious cancer with an 80 % reduction in population size across their range. In 2012, an island population was established to protect the species. The island population was harvested three times with devils translocated to wild location across Tasmania (2016, N=16; 2017, N=33; 2018, N=30). Efforts to harvest individuals demographically representative of the source population was achievable in 2017 and 2018, with a large male bias occurring in 2016 due to logistic constraints. Using internal relatedness (IR; proxy for heterozygosity), we showed that harvesting animals in equal-sized groups of high, medium and low IR, succeeded in maintaining genetic profiles of the source population over repeated harvests. There was little within-year variation in observed heterozygosity pre- and post-harvest compared to the harvest group, and only a small reduction (0.001) in heterozygosity across years. We used simulations to show that using a random selection would likely have eroded diversity relative to our structured approach (86.6% of iterations showed greater loss). Our method maintained a viable source population over multiple harvests, providing conservation managers with a tool for conserving single source populations used in threatened species translocations.