Trends and yield constraints in the production systems of beans and cassava in Latin America with emphasis on research design implications

The trends and yield constraints in production systems of beans and cassava are analyzed; examples are given of the role of 3 types of information (macro stage, micro stage, and critical inference stage) in the design of research for beans and cassava at the Centro Internacional de Agricultura Tropical (CIAT, Colombia). It is indicated that the principal focus of information processing in international centers is to achieve a more solid empirical support (or rejection) of the critical inferences, which are made explicit and are then evaluated with a more systematic data collection. In commodity programs there is a natural evolution to farm level data collection. In the evolution of crop technologies, the consequences new technologies may have on income distribution are discussed; cassava and beans are essentially produced by small farmers outside prime agricultural areas. The choice of breeding strategy that involves the relationship among several disciplines is also analyzed and it is noted that the different constraints and the different crop characteristics require different strategies for cassava and beans to produce the same output, a stable high- yielding var. not dependent on high levels of purchased inputs. (CIAT