Ilha Grande, an island in the south of Rio de Janeiro state, Brazil, has faced a long series of historical boom-and-bust cycles. Yet, during the last century, the island has recuperated most of its original forest cover, compared to rapid and continuous deforestation of the continental Atlantic forest. This paper investigates the historical determinants of forest cover changes, and relates them to the contemporary tourism boom from the two mega-cities, Rio de Janeiro and Sao Paulo. A case study of rising backpacker tourism in a traditional fishermen's village is presented, and findings are related to other parts of the island. The common hypothesis that 'backpackers' generate no income on the island is rejected: they spend little per PAX, but in places where visitor numbers are large, tourism income is amazingly high, compared to any traditional productive activity, producing a considerable local poverty-alleviation impact. The study also confirms the high economic potential of forest-based services - in this case, recreation and landscape-contemplation benefits - in forests close to urban areas of middle-income developing countries. The large local tourism cash inflows have been used for residential construction, to buy consumer durables, and to buy additional leisure time. Although local labour requirements for tourism are generally low, rising income has implied momentous changes in the structure of local production. In addition, it is argued that negative environmental impacts from tourism are under-researched, and probably have been over-stated in previous assessments. The recommendation in these studies of implementing 'carrying-capacity' based tourist-access restrictions does not seem justified. A well-designed entrance-fee scheme to finance necessary investments in the tourism sector may be a much more feasible alternative, in order to embark on a development path of the sector which is beneficial for both tourists and residents, while conserving the integrity of the forested landscape.