We analyzed counts, vital rates, and limiting factors for northern Yellowstone elk (Cervus elaphus) before and after wolf (Canis lupus) restoration in 1995–1996 to evaluate predictions that elk numbers would move to a lower equilibrium point with corresponding density-related changes in vital rates. Elk counts decreased from approximately 17,000 in 1995 to 8,335 in 2004. Pregnancy rates for prime-age females (3–15 years) during 2000–2003 were high (0.90) and similar to those during 1950–1967 when elk density was 30% lower (5–9 elk/km²). The survival rate for prime-aged females was 0.85 (95% CI = 0.81–0.87) compared to 0.99 when harvests were low and wolves absent. The proportions of elk harvested each year increased as elk numbers decreased during 1990–2002 but departed from this anti-regulatory trend as permit levels were reduced in 2003–2004. Snow pack strongly influenced elk vulnerability to hunting by increasing migration to lower elevations. Thus, harvests removed a relatively constant proportion (27±5%) of animals that migrated out of the park each year, primarily prime-aged females with high reproductive value. Conservative estimates of wolf off-take (>1,000 elk) exceeded harvests by 2003, with wolves primarily selecting calves and older elk with lower reproductive value. Recruitment decreased as the ratio of wolves to elk increased, and wolves maintained high kill rates and rapid population growth despite a 50% decrease in elk counts. Elk numbers likely will continue to decrease until 1) levels of harvest and predation decrease sufficiently, 2) there is sufficient time for recruitment of calves to prime breeding age, and 3) there is a numerical response of wolves to fewer elk. We recommend that managers quickly adjust antlerless permit quotas to population size in a density-dependent manner so that harvests do not accelerate the decrease in elk numbers.