Initial movements of re- introduced wildlife populations can determine short-term restoration success. Managers need ways to encourage release site fidelity to mitigate suboptimal breeding, reduce mortality rates, and minimize human-wildlife conflicts. We studied initial movement ecology of elk (Cervus elaphus) fitted with GPS collars and re- introduced to the Missouri Ozarks in 2011 (n = 32), 2012 (n = 21), and 2013 (n = 31) for 6 mo post release. We assessed maximum displacement from the release site, range shifts, and range size across four sequential time frames (0–10 d, 11–31 d, 32–61 d, and 62–183 d). Compared to other elk restorations in eastern North America, site fidelity was high, with maximum distance from the release site 62–183 d post release ≤10 km for 94% of 2011 animals, 57% of 2012 animals, and 97% of 2013 animals. Elk range sizes were similar during the first 61 d post release but doubled in size 62–183 d post release to an average of 26.2 km² (range: 4.0–218.8 km²). The average range overlap for individual elk in sequential time periods was between 23–26% across years, indicating elk used different areas over time. Release site had the greatest influence on initial movements; one site used in 2012 was associated with greater release site displacement and range sizes. Maternal cows also demonstrated higher site fidelity to the release site than nonmaternal cows. High site fidelity and small home ranges in elk recently restored to Missouri may be attributed to soft release, minimal human disturbance, quality habitat, and release groups of mature females.