The need to obtain more bio-physical agroforestry data from on-farm situations is emphasized. Two methods are discussed: studies of single, established multipurpose trees and on-farm experimentation. Some of the complexities of the former are described and briefly illustrated. A new approach to on-farm experimentation is proposed which involves randomly sampling natural conditions on-farm which are selected to be relevant and useful comparisons ("ecological" treatments), and/or with added manipulations ("interference" treatments). The advantages of this ecological approach, using any "quadrats" designs, are discussed vis-a-vis classical conventionally designed experiments, which may not be perceived by the farmer as so relevant, nor lend themselves to his active participation so readily
Publication year
1988
Authors
Language
English
Keywords
agroforestry, experimentation, multipurpose trees, on-farm research, surveys
