This paper develops a theoretical foundation for the notion of differential urbanization, in which groups of large, intermediate-sized, and small cities go through successive periods of fast and slow growth in a continuum of development that spans the evolution of urban systems in developed and less developed countries. A model depicting net migration patterns over time for major metropolitan, intermediate-sized, and small urban areas identifies six stages of differential urbanization. Data from three countries that span the development spectrum are used to test the accuracy of this model. A distinction between mainstream and substream migration flows provides an indicator of the concurrent concentration and deconcentration forces shaping urban systems. Couiiterurbaiiitatio~i represents the final phase in the first cycle of urban development, and is followed by a second cycle in which urbanization and spatial concentration dominate once again. At advanced levels of urbanization, the model can b e used t o characterize the degree of development within regions or subregions of a country. Also, precise definitions are suggested for the "clean break," the end of urbanization, and the beginnings and ends of polarization reversal and counterurbanization.

Of the eighteen countries studied in this paper, eleven (Japan, Sweden, Norway, Italy, Denmark, New Zealand, Belgium, France, West Germany, East Germany, and The Netherlands) show either a reversal in the direction of net population flow from their sparsely populated, peripheral regions to their densely populated core regions or a drastic reduction in the level of this net flow. In the first seven of these eleven countries, this reversal or reduction became evident only in the 1970s; in the last four, its onset was recorded m the 1960s. Six countries (Hungary, Spain, Finland, Poland, South Korea, and Taiwan) have yet to show an attenuation in the movement of persons into their most densely populated regions. Some possibly unreliable British data likewise fail to reveal a slackening in the "drift south" of the British population. Three additional discoveries described in this paper are the following: (1) Migration continues strong into the capital regions of the three Eastern European countries studied here (Poland, Hungary, and East Germany). However, the low natural increase of these regions has blunted their expansion. (2) Though domestic migration into the capital regions of France, Sweden, and Norway has declined dramatically, foreign immigration into these regions remains at a high level. (3) Net domestic migration into the core regions of Sweden, Japan, and Italy, countries separated by vast distances, fluctuate from year to year in a remarkably similar manner.

Changes in the spatial distribution of minority populations and factors responsible for such changes form an important research topic in the study of the contemporary immigrant societies of Europe and North America. This paper clarifies both the trends and determinants of the spatial redistribution of mainly Russian‐speaking ethnic minorities in Estonia by focusing on out‐migration from the gateway cities or the main minority concentration areas. We use individual data from the 2000 census. The main results show that the dominant flow of migration among members of ethnic minorities is out of the main gateway cities in Estonia, and this trend is very different from the situation during the Soviet period. All of the personal characteristics that measure intentional ties with the majority population and the host country, such as proficiency in Estonian, having an Estonian partner, and Estonian citizenship, exert a positive influence on both leaving these areas and settling in non‐ethnic destinations, while the mere passage of time (generation replacement) has no straightforward influence on minority spatial redistribution. Copyright © 2010 John Wiley & Sons, Ltd.

The results of this study indicate that during the 1990s urbanization was the dominant process in inter-regional migration, and residential suburbanization prevailed at the intra-metropolitan level.These tendencies are new and different from the trend existing during the Soviet era.These urbanization and suburbanization trends correspond closely with the expectations of the city life cycle and the differential urbanization models, indicating that patterns in Estonia during the 1990s follow those experienced by capitalist countries. The evaluation of urbanization and migration trends by age, education, gender and ethnicity, using 2000 census data, indicate that young people, females, migrants with a higher education, and Estonians accounted for the national urbanization trends. Also, in contrast to the Soviet era, Estonians participated fully in the urbanization trends of the 1990s. Suburbanization was the predominant direction of movement within metropolitan areas, and this trend was similar for all population subgroups. Yet people with a higher education and those in the family ages were most likely to suburbanize. Estonian migration trends for the total population were similar to many countries of the former Soviet Union, but different from East Central European countries.We believe that the main reason for such a contradiction stems from the compositional differences of the populations in these two groups of countries. Migration trends of population subgroups,however,seem to be similar across countries in transition. In addition, this study develops a Migration Concentration Index (MCI) to measure concentration and deconcentration processes in inter-regional migration.

This paper summarises the usefulness of the differential urbanisation model, to characterise regional urban development in terms of a temporal sequence of stages from urbanisation through polarisation reversal to counter-urbanisation, as revealed in nine empirical tests (Britain, Estonia, Finland, Western Germany, India, Italy, Russia, South Africa, and Turkey). First, results of the testing are summarised. Based on the results of these nine empirical tests, we attempt to answer the question 'Did the differential urbanisation model fit reality?' Also discussed are modifications to the model suggested during the course of empirical testing. Then directions for future testing and improvements to the model are suggested.