This paper addresses the question, when are threatened or endangered species really recovered? The US Endangered Species Act enables the de-listing of species once demographic criteria are met. In the Greater Yellowstone ecosystem, two protected apex carnivores, grizzly bears ( Ursus arctos) and wolves ( Canis lupus), face removal from federal government protection due to population increases, a point at which they are expected to be integrated components of this ecosystem. We tested the assumption that these two carnivores are playing normative ecological roles in the Yellowstone ecosystem by comparing the extent to which wolves and bears have re-instilled anti-predator responses in a primary prey species, moose ( Alces alces), within wolf and bear recovery zones. As a type of control, we contrasted female moose from two areas in Alaska with different predator regimes to those in Wyoming. Populations from mainland Alaska, a region with a relatively intact carnivore assemblage, responded significantly more to odors of both carnivores. In contrast, a basic anti-predator reaction was lacking in Wyoming; and responses to grizzly bear odor only nominally increased after dependent young experienced heightened mortality. Additionally, the level of response among Alaskan moose living under virtual predator-free conditions for 25+ years closely resembled that of conspecifics in Wyoming. That such striking variation in prey responses exists re-enforces critical ecological differences between predator-intact and -defunct systems. Thus, although grizzly bears and wolves in the Yellowstone area will most likely be de-listed within the next few years, whether such action would be ecologically defensible is arguable. At this point in the recovery process, these predators may currently have limited ecological impacts in large portions of this region, at least as gauged by one potentially important prey species, moose. Although our data suggest ecologically incomplete conditions, other indices of carnivore recovery that include responses of other important prey species such as elk ( Cervus elaphus), may be more in tune with carnivore activities. We recommend that different types of ecological data available throughout recovery zones be used in consort with demographic criteria to evaluate when endangered carnivores are more fully integrated into their ecosystems. And, in the event of a disparity between these criteria, we also encourage a dialogue focusing on approaches towards bringing ecological conditions in concordance with demographic criteria, irrespective of whether one considers increasing population levels beyond the current target levels required for de-listing,and/or simply, additional time for the recovery process.