Abstract/Summary

Documentations of patterns in traditional knowledge is of paramount importance if we are to fully understand the relationship between people and their local environment. Investigations of medicinal plant use also contribute to discoveries of new drugs, by highlighting taxa which might have bioactivity of interest, and preservation of traditional knowledge, by identifying species important to local communities and describing how traditional knowledge is transmitted to future generations. Increasingly methods borrowed from ecology and community phylogenetics are used to understand patterns in medicinal plant use and inform bioprospecting. These methods are best applied to information-rich, large data sets, yet despite the growth in ethnobotanical research, data are scattered. This thesis exploits traditional knowledge documented in literature and herbaria, using newly compiled data for Brazil to explore patterns in medicinal plant use, in documentation of ethnomedicinal species, pharmacological investigation and distribution of plant use. Comparison of data from literature and herbarium sources, based on 1400 use reports of medicinal use of species of the family Leguminosae, showed herbaria to be a valuable source of data for ethnobotanical research. Despite the lower proportion of herbarium records (compared to literature records) with information on therapeutic application, modes of use and parts used, herbarium records validated data in the literature and added new information. These and additional data were used to demonstrate Ewe, a web-based database developed here, with tools to store and visualise ethnobotanical data. Research questions relating to pharmacological investigation of Brazilian plants used in traditional medicine, and to the spatial distribution of traditional knowledge were addressed using these and other data sets compiled for this study. Phylogenetic methods were used to explore the diversity of Leguminosae species with medicinal use, revealing phylogenetic diversity of species used despite of lineages rich in species with medicinal use (hot nodes) accounting for 45% of the species used in local medicine. Hot nodes are better characterized pharmacologically. Finally, the spatial distribution of traditional knowledge in Brazil was investigated, using a novel generic-level phylogeny for the Brazilian flora and species-level herbarium data for Angiosperms in Brazil with medicinal uses. According to taxonomic and phylogenetic measures of beta diversity, spatial heterogeneity in medicinal plant use was identified, highlighting local adaptation of medicinal floras despite the preference for some plant groups revealed by clustering in deep nodes.