This paper examines Cameroon's model of forest-management decentralisation by characterising its organisational infrastructure and by assessing - and anticipating - the ecological effects of those policy changes. The essay is based on nvironmental governance research conducted in Cameroon during the last three years. Five Community Forests, one Council Forest and nine forestry fee management committees were covered by the study. Methods of data collection included participant observation, analysis of historical trends, semi-structured interviews at the regional level, focus group meetings, historical transects of landscape and future scenarios. The first section describes the key reforms put in place by the new forest management regime, including the institutional and socio-organisational choices legally prescribed for outlying actors in the decentralised management of forests and their revenues. The mechanisms connecting these institutional and socio-organisational choices to implementation - that is, the management of Community Forests, Council Forests and forestry revenues - are examined in the second part of the essay. The third section assesses the social outcomes of these processes. Ecological risks and uncertainties due to the way decentralised management is conducted at both the local and the regional levels are addressed in the fourth section. The Cameroonian decentralisation model, already registered some positive infra-outcomes. After all, decentralisation is not mechanical. It is not only explainable by laws and institutional arrangements: its implementation depends on many variables. It requires sufficient time to develop due to the complexity of human and institutional behaviour and because of unpredictability, variability, contingency and change, as well as many other stimuli. In sum, it is a ‘story' of adaptation, with successes and failures. Policy innovations and reforms are in themselves experiments and should not be perceived as victories or crises, but, instead, as arenas of ongoing lessons and progressive learning.