Stocking fish is considered to be critical to prevent extinction in most conservation programs. This is because it is often the only aspect of management can be conducted in a timely manner. Unfortunately, augmentation has had little success in recovering endangered species. In some cases, augmentation has made things worse. Obviously, a new approach is needed if aquaculture-assisted fisheries programs are to be successful. That new approach is conservation aquaculture. Raising fish in environmentally enriched environments has been shown to produce fish that have increased post-augmentation fitness. To maximize post-augmentation fitness, fish used in aquaculture-assisted fisheries programs need to be raised in an environment that resembles the wild environment, using conservation aquaculture. In order to properly evaluate the fish culture component of an aquaculture-assisted fisheries program, a new set of hatchery evaluation metrics are needed. Most aquaculture-assisted fisheries programs use fish produced in traditional, barren culture units, using intensive production management. Because of that, the program uses food fish hatchery evaluation metrics: survival rate, yield, feed conversion, and profit (cost/fish). These metrics are important in food fish farming, but using them in an aquaculture-assisted fisheries program is part of the problem. In conservation aquaculture, the hatchery evaluation metrics are: genetic variance, domestication, epigenetics, and behavior. These evaluation metrics produce an evolution-based evaluation of the augmented fish. One critical difference is survival rate. In food fish farming, you need a high survival rate to make a profit. But in an aquaculture-assisted fisheries program, a high hatchery survival rate lowers post-augmentation survival and can lower the fitness of the wild stock following introgression. This occurs because in intensive aquaculture, you circumvent natural selection by feeding fish, excluding predators, etc. This enables fish with sub-viable genotypes to survive in the hatchery. Consequently, sub-viable genotypes exist in the fish that are augmented, and that places a genetic (evolutionary) burden on the wild population, which lowers its fitness following introgression. In conservation aquaculture, you are trying to mimic the wild environment, where survival rates are low because natural selection ruthlessly culls sub-viable genotypes. By not feeding fish but forcing then to forage and compete with their conspecifics for food and by stocking predators, selection in a conservation aquaculture program will cull fish with sub-viable genotypes, mimicking the natural selection that occurs in the wild. The difference in pre-augmentation selection pressures between those produced by traditional aquaculture and conservation aquaculture explains why conservation aquaculture produces fish for aquaculture-assisted fisheries programs with better evolutionary fitness. Conservation aquaculture accomplishes this by producing a lower hatchery survival rate, and that’s why hatchery survival rate is not an important metric in conservation programs. The ideas that have been discussed throughout the book were used to design and manage the world’s first large-scale, purpose-built conservation aquaculture mesocosm, the “Refugium” at the Los Lunas Silvery Minnow Refugium. The design of the Refugium and the conservation aquaculture management that was developed to produce an endangered species is presented. In addition to producing fish for augmentation, a purpose-built conservation aquaculture mesocosm can be used to learn about a species’ behavior, which can then be used to improve recovery management. Behavior studies that were conducted in the Refugium and, how this knowledge can be used in recovery management, are described.