Completed Agricultural production which is dependant on the clearing of forest land (usually logged natural forest or long-term forest fallow) is common to many countries of the humid tropics. This shifting (slash and burn or swidden) agriculture results in the reduction of forest cover as only a proportion of the cleared land is returned to forest. In Bolivia, and in other countries, rise in populations through immigration and natural increase from 1-5 people per sq.km. to 5-50 per sq.km over the last 40 years has resulted in rapid reduction of forested cover and the degradation of siol, water and bio-diversity resources. In order to conserve resources and to provide farmers with long-term stability and economic well being it is necessary to develop farming systems that allow farmers to produce sufficient food and/or incomes from permanent holdings on a sustainable basis. This accords with the CIAT (Bolivia) strategic plan for 1990-95 which states as a main priority the development of sustainable agricultural production systems which permit the definitive establishment of farmers on their land. To address the destruction of the following problems: (i) moist tropical forest by expansion of the agricultural frontier.(ii) Degradation of natural resource (principally soil, water and biodiversity) through unsustainable agriculture practice.(iii) Lack of locally verified, sustainable agricultural income - generating technologies. The research successfully developed and promoted crop and agroforestry systems, and management practices that enhance stability and sustainability in forest margin agriculture in Bolivia. The impact of the project was greatly increased by operating as part of a larger research effort which addressed three technical issues (soil fertility, weed management and income-generation) simultaneously. Participatory research methodologies were, to a large extent, institutionalised in the main collaborating institution. Recommendations for improved weed management were developed for annual (rice-based) and perennial systems, principally through the use of legume covers. Results from 158 on-farm trials (supported by 5 formal trials) validated cropping systems which integrate technologies for weed control, soil fertility maintenance, and improved income generating opportunities. The slash-and-burn colonist farmers targetted by the project are recognised to be among the core poverty groups in Bolivia. In recognition of the lack of financial capital, most cropping systems tested by the project had a low requirement for external inputs. Higher-capital systems were made more accessible by using crop sequences that provide income in the short and medium term. Considerable effort was put into improved rice systems, and the integration of rice into perennial systems, in recognition of its critical importance as a cash and subsistence crop for the poorest farming families. Development and promotion of improved management strategies for the integration of crop/livestock/agroforestry systems.Improved farmkng systems for maintaining or enhancing soil fertility: controlling weeds and diversifying farm income, developed and promoted.Methods for farmer participatory research developed and documented. Improved farming systems for maintaining or enhancing soil fertility; controlling weeds and diversifying farm income, developed and promoted. The NRSP-SPP-CIAT project validated the suitability of perennial species for the project areas (including citrus, palm hearts (Bactris gasipaes), macadamia, bananas) and identified viable cropping systems and sequences. The association of perennials with annuals, semi-perennials and legume covers proved how small-scale farmers, with very limited financial capital, can diversify into perennial systems. The annuals and semi-perennials offset establishment costs and provide a source of income in the short and medium term. The legume covers prevent weed build-up and reduce labour requirements for weeding, both during establishment (in association with annuals and semi-perennials) and in mature plantations. Suitable legume species and management were identified for both phases. The association of citrus with pineapple, annual crops and legume covers provided the most popular. In a survey of collaborating farmers at the end of the project, 54% of those testing citrus had already adopted and expanded these systems on their farms, and most were not (as previously) planting monocrops, but had adopted the principle of cropping systems.Interim guidelines for improving bush-fallow productivity developed. Two novel agroforestry systems were tested with 20 farmers. One of the systems enriches existing bush fallow with native fruit and timber species, leading to a conversion (without burning) to a productive and income-generating system. The second system starts with cleared land and takes it through a sequence of annual, semi-perennial and perennial species to a permanent and productive tree-rich system. The native fruits grow well within the bush fallow, but are slow to produce fruit. There is a role for research to select provenances that are quick bearers, and methods such as grafting that might speed fruiting. One of the timber species tested grows fast, but mahogany is a slow grower. All are long-term crops, and it is important to include short and medium-term income sources in the farmng system, while protecting establishing trees from cattle. Adoption is limited by an underlying attitude that trees are not a valuable resource. Markets are often inaccessible and small, and are poorly understood by farmers. Seedlings are often unavailable, and technical assistance for agroforestry has been weak. CIAT should continue to monitor the performance of the agroforestry plots, in collaboration with farmers.Methods for farmer participatory research (FPR) developed and documented. Participatory methodologies were successfully used and developed by the project. It is now an accepted methodology in CIAT (who previously had little experience outside on-station, researcher managed trials) as well as other target institutes. Significant advances were made (and documented) in developing FPR methodologies, particularly in terms of data management, analysis and interpretation.