Intraspecific variation, knowledge and local management of cassava (Manihot esculenta Crantz) in the semiarid region of Pernambuco, Northeast Brazil

Historically, the species Manihot esculenta Crantz (cassava) is an important crop of great livelihood value for most rural farming communities of Brazil. Cassava ethnovarieties are selected for cultivation over generations based on farmers' interests, leading to the maintenance and a very particular local diversity evolution. Although cassava has become a model species in studies that seek to understand how the knowledge and the strategies of use adopted by human populations influence cassava diversity throughout the world, factors that influence cassava diversity in semiarid regions are not well documented. Accordingly, this study examines the distribution of cassava ethnovarieties and whether socioeconomic characteristics influence the knowledge and use of these cassava ethnovarieties, among seven rural communities located in the semiarid region of Pernambuco, Northeast Brazil. Social exchange networks play an important role in the maintenance and distribution of cassava diversity. The structure of network depicting farmers and the ethnovarieties they cultivate was investigated. The methodological approach was based on classical methods in ethnobiology, such as semistructured interviews, field visits and discussions with farmers. Through snowball sampling, 50 farmers were identified and interviewed. The results show that socioeconomic factors affect the knowledge and use of cassava ethnovarieties and can be important predictors in studies that aim to evaluate the knowledge and use of food plants. It was clear that farmers use traditional knowledge and their preferences to select and cultivate cassava ethnovarieties, such as high yields, good quality of flour, good quality of cooking, that are important to take decisions to maintain or abandon them. Network analysis indicated that most farmers shared a common ethnodiversity core, while some also cultivated a smaller group of lower occurrence among the communities studied, which may have resulted from individual farmers' preferences.