It is proposed in this document that the selectivity of migration, in terms of both people and places become a more imposing influence in urbanization as the role of natural increase as a source of urban growth diminishes. Recent U.S. growth policy proposals have frequently been marked by a simplistic view of how urban growth works, compounded by an exaggerated sense of what policy has in its power to do. On the one hand, policymakers fail to perceive the full momentum of demographic change that could be guided; but they also underestimate the force of demographic constraints on their intended actions. This paper considers five aspects of the migration process related to these points: (1) the concentration of migratory growth in only a few metropolitan areas; (2) migration's apparently one-sided economic wisdom, arising out of the weakness of economic 'push'; (3) the potential for return migration, which derives from latent migratory predispositions coupled with new sources of retirement income; (4) wide local variations among metropolitan areas in the rate of migratory circulation; and (5) the intensifying phenomenon of urban population decline, now a characteristic of entire metropolitan areas. It is possible for a national growth policy to be development-oriented and for it to build on the processes of change. (Author/JM)