Agroforestry performance on small farms in Amazonia: Findings from the Rondonia agroforestry pilot project

Experiences from not only 'success stories' but also 'failed' agroforestry projects provide potentially useful lessons for future agroforestry-project designers. Experimental one-hectare agroforestry plots were established on 50 small-scale farms in the western Brazilian Amazon State of Rondonia from 1993 to 1995. Drawing from a menu of 25 different species (10 tropical hardwoods and softwoods and 15 fruits and palms), this species trial shows encouraging survival and growth performance for most species under wide ranging plot management regimes. Tropical hardwood survival rates (after 18 months) ranged from 65% for Cerejeira (Torresea acreana) to 88% for mahogany (Swietenia macrophylla). Survival rates for commercial fruit and palm species were even higher. A comparison of attributes of two sub-groups (successful and unsuccessful planters) suggests that previous experience with perennial monocultural cropping, greater social participation, land use history, and soil chemistry are positively associated with successful agroforestry species performance, while no significant differences exist between successful and unsuccessful planters in household size, area deforested, area in pasture, and land tenure security. A closer analysis of 'failed' agroforestry plots indicates the primary importance of social factors originating at the household-level (e.g. inadequate plot maintenance, improper planting techniques, illness, etc.). Twelve different causes of plot failure were cited, falling into three classes. Of the total number of reasons given for plot failure, household level factors represented 54% of all causes cited. Project design and implementation factors (inappropriate plot design, defective planting material, etc.) were cited 25% of the times and environmental factors (soil fertility constraints and pasture grass invasion) were cited 21% of the times.