a1 Instituto Nacional de Pesquisas da Amazônia-INPA, Av. André Araújo, 2936 cp 478 69.011-970, Manaus-Amazonas, Brazil Tel: +55 92 3643 1822 Fax: +55 92 3642 8909 e-mail: email@example.com
Wright and Muller-Landau (2006) proposed that rural–urban migration could reduce tropical forest loss worldwide and allow secondary forest to occupy abandoned clearings, thereby reducing the expected magnitude of tropical species extinctions. However, the usefulness of this global generalization is highly dependent on its being correct for Brazilian Amazonia, which has the world's largest remaining area of tropical forest. Among the features of the deforestation process that make urbanization an unlikely mechanism for land abandonment in Brazil is the fact that most deforestation is done by ranchers rather than by shifting cultivators. A family of shifting cultivators is only capable of farming a limited area and cannot greatly expand the area it cultivates if additional land becomes available. Ranchers, by contrast, can occupy huge areas and expand their holdings when the opportunity arises. Were part of the population to leave for the cities their land would usually be bought by neighbouring ranchers who would maintain and expand the deforested landscape. As Wright and Muller-Landau (2006) predicted, their interpretations have generated considerable controversy (Brook et al. 2006; Gardner et al. 2007; Laurance 2006; Sloan 2007). Only part of this controversy will be considered here, namely the expectation that rural population in tropical areas will decline, allowing secondary forests to regrow in large areas of abandoned farmland. Specifically, the present paper will consider the applicability of the Wright/Muller-Landau hypothesis to Brazilian Amazonia.