The setting of conservation priorities within species requires explicit goals and identification of the appropriate targets for conservation. I suggest that the conservation goal should be to conserve ecological and evolutionary processes, rather than to preserve specific phenotypic variants–-the products of those processes. From this perspective, we should seek to conserve historically isolated, and thus independently evolving, sets of populations (i.e., Evolutionarily Significant Units, ESU). This can require manipulation of the component Management Units (MUs), some of which may be phenotypically distinct. Here I explore the application of this approach to the design of translocations for conservation management. A process-oriented approach suggests that individuals should not be translocated between ESUs, but under some circumstances it is appropriate to mix individuals from different MUs within an ESU. These circumstances include augmentation of remnant populations that are showing signs of inbreeding depression or increased fragmentation and the use of mixed stocks for reintroductions into modified or changing environments or for introductions into novel environments. These actions are consistent with the goal of maintaining processes, but the extent to which differences in adaptation or coadaptation constrain the viability of populations subject to translocation needs further exploration. Both theory and limited experimental evidence suggests that these constraints can be overcome if sufficient genetic variation is present and evolutionary processes can operate without substantially reducing population viability. One remaining question is the extent to which genetic coadaptation, and the resulting outbreeding depression, develops along environmental gradients and how this compares to the build-up of genetic incompatibility between historically isolated populations.