Differences in income distributions have long characterized regions within the United States. The US South has historically had both lower incomes and greater inequality than the non-South. The magnitude and persistence of the differences have attracted intense scrutiny by a number of able social scientists.'The size, persistence, and seem-ing intractability of the regional differences have been of particular interest to economists because, with freely flowing labor and capital, the lack of convergence seems to contradict the neoclassical model of economic growth. 2 It is now apparent to even casual observers that fundamental changes have occurred in the South relative to the non-South. For example, migration of skilled and highly educated manpower and industrial relocations to the South have narrowed regional differences. The impact of these changes on income distributions in the South and non-South in the decade of the 1970's is the subject of this paper. The convergence of US regional income distributions is investigated by considering some fundamental questions. What is meant by convergence? What hypothesis should be tested to confirm or reject the claim that convergence has occurred? Has regional convergence actually occurred? To address these questions, we apply recent contribu-tions to the theory and measurement of income distributions to the South and non-South. Developments in applied welfare theory establish an important relation between dominance of one income distribution by another and relative levels of economic well-being. We use micro data from the public-use samples of the 1970 and 1980 censuses to compare the South and non-South income distributions across time and to test for rank dominance, which is equivalent to first-degree stochastic dominance. We also investigate convergence by testing for generalized Lorenz dominance, which is equivalent to second-degree stochastic dominance. In addition to changes in comparative well-being, changes in the relative degree of income inequality are also of interest. To investigate this issue, we examine convergence of South and non-South Lorenz curves. Recently developed inference-testing procedures are applied to measure changes, to test for convergence, and to rank the 1969 and 1979 income distributions.