We assessed the relative importance of wilderness to gray wolf (Canis lupus) population dynamics over 50 years in a population that 1) was long extant (i.e., not reintroduced or recolonized), 2) was not subject to harvest in our study area until recently, and 3) used both wilderness and adjacent, mainly public, non-wilderness. We analyzed the survival of radiocollared wolves (n = 756 collared-wolf tenures) during 1968–2018 in the Superior National Forest, Minnesota, USA, including the Boundary Waters Canoe Area Wilderness. Over 50 years, adult annual survival was 78%. Wolves captured in wilderness tended to exhibit higher survival than those captured in nonwilderness, but the difference was more pronounced during harvest years and post-harvest years when wilderness wolf survival remained relatively high and non-wilderness wolf survival dropped (relative to pre-harvest). During Nov–Apr of pre-harvest years for adults, the natural mortality rate was similar for non-wilderness wolves and wilderness wolves (both 6%), but the anthropogenic mortality rate was higher for non-wilderness wolves than wilderness wolves (7% versus 1%), as was the illegal mortality rate (5% versus 1%). During Nov–Apr of preharvest years, wilderness wolves were less likely to die than non-wilderness wolves (p = 0.042; hazard ratio = 0.59), pups were more likely to die than adults (p = 0.002; hazard ratio = 1.84), and males were less likely to die than females (p = 0.053; hazard ratio = 0.73). Our long-term wolf survival, cause-specific mortality, and hazard results will inform management agencies whenever wolves are delisted, and jurisdiction for them passes to states.