Socio-agronomic features define the structure of socioecological seed exchange network and “on farm” conservation of agrobiodiversity in quilombola communities in Brazil - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Official websites use .gov A .gov website belongs to an official government organization in the United States. Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites. Search Log in Dashboard Publications Account settings Log out Search… Search NCBI Primary site navigation Search Logged in as: Dashboard Publications Account settings Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide PERMALINK Copy As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice J Ethnobiol Ethnomed . 2026 Mar 11;22:38. doi: 10.1186/s13002-025-00847-4 Search in PMC Search in PubMed View in NLM Catalog Add to search Socio-agronomic features define the structure of socioecological seed exchange network and “on farm” conservation of agrobiodiversity in quilombola communities in Brazil Isabella Fernandes Fantini Isabella Fernandes Fantini 1 Departamento de Botânica, Laboratório de Sociobiodiversidade (SociobioLab), Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil 2 Programa de Pós-Graduação em Biodiversidade e Conservação da Natureza, Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil Find articles by Isabella Fernandes Fantini 1, 2, ✉ , Gustavo Taboada Soldati Gustavo Taboada Soldati 1 Departamento de Botânica, Laboratório de Sociobiodiversidade (SociobioLab), Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil 2 Programa de Pós-Graduação em Biodiversidade e Conservação da Natureza, Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil Find articles by Gustavo Taboada Soldati 1, 2 , Fernanda Vieira da Costa Fernanda Vieira da Costa 3 Departamento de Ecologia, Laboratório de Biodiversidade e Interações Ecológicas (LaBIE), Instituto de Ciências Biológicas, Universidade de Brasília (UnB), Brasília, CEP 70910-900 Distrito Federal Brazil Find articles by Fernanda Vieira da Costa 3 , Thiago da Silva Novato Thiago da Silva Novato 4 Departamento de Ciências Florestais Laboratório de Ecologia, Manejo e Conservação da Fauna Silvestre (LEMaC), Escola Superior de Agricultura Luiz de Queiroz da Universidade Estadual de São Paulo (Esalq-USP), Piracicaba, CEP 13418-900 Brazil Find articles by Thiago da Silva Novato 4 , Fátima Regina Gonçalves Salimena Fátima Regina Gonçalves Salimena 2 Programa de Pós-Graduação em Biodiversidade e Conservação da Natureza, Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil 5 Departamento de Botânica, Herbário Leopoldo Krieger (CESJ), Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil Find articles by Fátima Regina Gonçalves Salimena 2, 5 Author information Article notes Copyright and License information 1 Departamento de Botânica, Laboratório de Sociobiodiversidade (SociobioLab), Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil 2 Programa de Pós-Graduação em Biodiversidade e Conservação da Natureza, Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil 3 Departamento de Ecologia, Laboratório de Biodiversidade e Interações Ecológicas (LaBIE), Instituto de Ciências Biológicas, Universidade de Brasília (UnB), Brasília, CEP 70910-900 Distrito Federal Brazil 4 Departamento de Ciências Florestais Laboratório de Ecologia, Manejo e Conservação da Fauna Silvestre (LEMaC), Escola Superior de Agricultura Luiz de Queiroz da Universidade Estadual de São Paulo (Esalq-USP), Piracicaba, CEP 13418-900 Brazil 5 Departamento de Botânica, Herbário Leopoldo Krieger (CESJ), Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora, CEP 36036-900 Minas Gerais Brazil ✉ Corresponding author. Received 2025 Aug 14; Accepted 2025 Dec 31; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13094044 PMID: 41814307 Abstract Background Traditional agricultural systems are rooted in the local management, selection, and conservation of agrobiodiversity. Understanding the socioecological dynamics that sustain these systems is essential for developing sustainable practices that ensure food security and sovereignty in the territories of traditional and Indigenous peoples. This study assessed the role of seed exchange networks in on-farm agrobiodiversity conservation in quilombola communities in Brazil that face environmental and political threats. We emphasize the role of socioecological networks and socio-agronomic variables in shaping how agrobiodiversity is maintained, shared, and regenerated across time and space. Methods We conducted semi-structured interviews, free listing, participant observation, and guided tours with 48 agrobiodiversity management units (AMUs) from five communities, documenting socio-agronomic variables and ethnovariety richness with botanical identification in the field and literature. We recorded all ethnovarieties shared among internal AMUS - living in the quilombola communities, and external AMUs - outside the territory. We then sorted 15 ethnovarieties per internal AMUs to collect data on seed exchange interactions. Further, we registered data on socio-agronomic variables, ethnovarieties richness, and seed flows (donation and reception) to analyse the properties of AMUs (nodes in the network) and seed exchange patterns in the network, and assess their potential for conserving agrobiodiversity. Results and discussion We documented a total of 359 ethnovarieties. The complete and open network was formed by 185 AMUs − 48 internal and 137 external - which realized 424 events of seed exchanges. Agro-environmental diversity, cultivated area, and the period living in the community were positively associated with AMUs’ richness and its centrality in the network, highlighting their role as agrobiodiversity guardians and network bridges. The seed exchange network displayed low nestedness, low connectance, and high modularity, indicating the formation of cohesive subgroups of AMUs with strong exchanges among specific partners and limited intergroup seed flows. These findings reflect social segregation and reveal vulnerabilities, as varieties unevenly distributed across modules may not circulate widely, reducing agrobiodiversity resilience. Conclusions We argue that historical and material conditions are critical for sustaining on-farm agrobiodiversity conservation in quilombola territories. Land tenure security and territorial rights are essential for maintaining traditional agroecosystems that integrate ecological knowledge, cultural heritage, and biodiversity management. Strengthening seed exchange connectivity, fostering collaboration across groups - from inside and outside territories - are urgent actions to enhance resilience, safeguard traditional knowledge, and ensure long-term biocultural justice. Supplementary Information The online version contains supplementary material available at 10.1186/s13002-025-00847-4. Keywords: Biocultural heritage, Ethnobiology, In situ conservation, Landrace seeds, Political ecology, Traditional ecological knowledge Background Agrobiodiversity encompasses all biological diversity in agricultural ecosystems, including the inter- and intraspecific richness of domesticated, wild, ruderal, and spontaneous plants, along with the associated popular knowledge and management practices [ 1 ]. Indigenous peoples and traditional communities are the primary actors responsible for generating, caring for, and maintaining agrobiodiversity (Fig. 1 ) through popular genetic improvement systems within their territories [ 2 , 3 ]. These practices are recognized as “on-farm” conservation strategies, embodying agricultural practices’ cultural and continuous aspects. They promote the dynamic and equitable establishment of agroecosystems in a decentralized manner, are low-cost, engage local guardians in decision-making processes, preserve high genetic diversity, and contribute to agricultural, food, and nutritional security and sovereignty within traditional territories [ 4 , 5 ]. In this context, promoting and strengthening on-farm conservation represents a fundamental decolonial strategy to address ongoing socio-environmental crises [ 1 , 4 , 6 , 7 ]. Fig. 1. Open in a new tab Ethnovarieties of corn with blue ears and rosinha beans generated, managed, and maintained by quilombola communities in Mariana, Minas Gerais, Brazil (pat42, 76 years old). Author: Isabella F. Fantini. Date: March 2022 One of the key processes associated with agrobiodiversity is the concept of “seed exchange networks,” where “seeds” refer to any type of plant propagule [ 1 , 8 , 9 ]. These networks operate as socioecological systems, characterized by the dynamic interplay between humans and the environment [ 10 ]. In these systems, humans may act as “biological corridors” linking distinct components of biodiversity in tangled networks [ 11 ]. Socioecological systems, thus, may be very diverse, encompassing ethnovarieties in entangle networks of sociobiodiversity. As follows, ethnovarieties can be viewed as biological species that embody socio-biocultural traits resultanting from the selection and management practices of traditional farmers [ 12 ]. Then, incorporates the socio-environmental context of the studied systems. The ethnovariety concept is useful in studies that aim to assess the sociobiodiversity, as it is broader than the scientific nomenclature of species and varieties [ 13 ]. Scholarly works on seed exchange networks have frequently aimed to identify the roles or positions of individuals (i.e., network nodes) within these systems, emphasizing structural properties such as node centrality, which represents their importance in the system [ 14 – 22 ]. For instance, Abizaid et al. [ 14 ] identified highly connected “bridge” farmers in Amazonian Indigenous villages, while Ricciardi [ 16 ] applied measures of degree and betweenness to reveal which families ensured equitable access to seeds in Guatemala. Some contributions, aligned with agribusiness interests, have used centrality to facilitate the introduction of formal market species [ 14 , 17 , 23 – 25 ]. For example, Otieno et al. [ 26 ] employed network analysis to defend a policy of inserting commercial seed varieties in East Africa. However, the vast majority of publications have linked network analysis to agrobiodiversity conservation [ 8 , 11 , 13 , 15 , 16 , 18 , 19 , 21 , 22 , 27 – 31 ]. So, the aforementioned survey has sought to understand how farmers access agrobiodiversity, to design more efficient governance and management strategies to increase global biodiversity, assuming that network structure (nodes and edges) influences the access, maintenance, improvement, and local distribution of ethnovarieties [ 1 ]. This body of work usually highlights three key approaches to assess nodes importance in socio-agronomic systems: (1) nodal farmers or agrobiodiversity management units (AMUs) with high degree centrality, typically those donating or receiving the most ethnovarieties [ 11 , 15 , 17 , 18 , 22 , 24 , 26 , 29 ]; (2) connector farmers, identified by high betweenness centrality, who act as bridges linking different modules or groups [ 11 , 15 ], 17, 18, 22, 24, 26, 29]; and (3) AMUs that distribute seeds to more peripheral nodes, characterized by harmonic closeness centrality, facilitating access for more vulnerable AMUs [ 15 ]. Building on these insights, various authors have sought to identify the sociobiocultural, that is, social and cultural aspects linked to nature traits that distinguish AMUs that are more central in exchange relationships [ 11 , 13 – 19 , 21 , 22 , 24 , 26 – 29 ]. The concept of homophily, a fundamental principle in social network theory, suggests that actors tend to interact with socially or spatially similar peers; in this context, homophily helps explain why farmers exchange seeds primarily within kinship or neighborhood circles [ 15 , 16 ]. Consequently, several analyses have demonstrated positive correlations between territorial and familial proximity and seed exchange, with AMUs more likely to exchange ethnovarieties with neighbors and kin [ 8 , 11 , 14 , 15 , 17 , 22 , 29 ]. This growing literature has consistently highlighted the role of trust in seed-sharing relationships, noting that social ties, such as kinship, territorial proximity, or cultural similarities, among central AMUs help sustain crop diversity, while social barriers, such as linguistic divergence, can hinder it [ 9 , 14 – 19 , 26 , 27 ]. Central AMUs are also often associated with greater social prestige [ 21 ]. Furthermore, they tend to: (1) have greater crop richness [ 11 , 13 , 14 , 18 , 26 , 29 ]; (2) be recognized for their managers’ knowledge of agrobiodiversity [ 11 , 14 , 16 – 19 , 21 , 22 , 24 , 27 ]; and (3) be known for introducing traditional seeds from outside the territory or for having better access to local markets or seed banks [ 22 , 32 ]. Although the age of individual managers does not appear to directly influence centrality, families with longer histories in the territory tend to occupy more central positions [ 18 ]. Central AMUs are also often associated with greater social prestige [ 21 ]. Additionally, they tend to display higher crop richness, be recognized for their managers’ agrobiodiversity knowledge, and be known for introducing heirloom seeds from outside the territory or having better access to local markets or seed banks [ 14 , 15 , 19 , 22 , 26 , 27 , 32 ]. Although the age of individual managers does not appear to directly influence their centrality, families with longer histories in the territory tend to hold more central positions [ 14 , 15 , 21 ]. Evidence also suggests that economically vulnerable groups rely more heavily on seed networks to remain in their territories [ 22 , 26 ]. Despite the current knowledge, several socio-agronomic dimensions remain underexplored, such as territorial extension, agroenvironmental diversity, and the economic activity levels of AMUs. Further, there is a significant gap in understanding how emergent properties derived from seed exchange networks can indicate the resilience of socioecological systems to perturbations and other stressors from external environments. In other words, the knowledge on how such structural metrics reflect a system’s capacity to absorb or withstand interventions remains overlooked. Understanding the organization of interactions between AMUs is essential to inform collective management strategies for agrobiodiversity maintenance and conservation [ 1 , 33 ]. Some studies on seed exchange networks have utilized structural metrics involving interaction nestedness [ 13 , 24 ], connectance [ 26 ], and interaction density [ 10 , 22 , 26 ] to explore how they can determine contexts of threats or conservation. It is suggested that low values of nestedness and connectance can negatively impact socioecological systems by undermining trust among peers and eroding traditional knowledge, thus weakening the collective memory of territories. As follows, low connectance in seed exchange networks can also demonstrate the possibility and necessity of incorporating contexts that may promote better relationships between AMUs and external agents, which may foster greater diversity of Otieethnovarieties within agricultural systems. Thus, it is suggested that future studies should further investigate the real effects of network structure on the resilience and conservation of agrobiodiversity in socioecological systems [ 26 ]. Similarly, a significant gap remains in understanding how network modularity influences the functioning of these systems. Specifically, we may expect that the formation of subgroups of AMUs that interact more frequently with each other than with the rest of the network can hinder or limit agrobiodiversity conservation. Therefore, assessing the influence of modularity in seed exchange networks helps to better understand how crop diversity is maintained and shared [ 1 , 33 ]. Given the above, studies of agrobiodiversity encompassing a seed exchange network approach indicate that access to ethnovarieties is closely tied to the centrality of AMUs involved [ 16 – 19 , 21 , 23 , 24 , 26 , 29 , 31 ]. However, the influence of socio-agronomic variables on AMUs centrality remains unclear, as well as the understanding of how network architecture (modularity, nestedness, connectance, among other metrics) relates to agrobiodiversity conservation. In this scenario, our study evaluates the role of seed exchange networks for assessing “on-farm” agrobiodiversity conservation. To do that, we hypothesize that (H1) socio-agronomic characteristics influence the richness of ethnovarieties managed by AMUs. We predict that the lower per capita monetary income, alongside longer residence in the community and household, greater number of economic activities, total managed agro-environments, and total cultivated area, will lead to a greater number of managed ethnovarieties. We expect to identify what socio-agronomic characteristics of the AMUs influence their capacity to contribute to “on-farm” agrobiodiversity conservation by providing the necessary conditions and resources for the cultivation and management of ethnovarieties. Our second hypothesis (H2) evaluates the influence of socio-agronomic variables and the richness of ethnovarieties managed by AMUs on their centrality in the network. We predict that the longer residence in the community, a greater number of managed agro-environments, larger cultivated areas, and higher ethnovariety richness will result in a greater AMUs centrality in the network. Here, we aim to analyze the role of AMUs in network dynamics, understanding their influence on ethnovariety flows and distribution. Finally, our third hypothesis (H3) attempts to test the role of network structure in predicting agrobiodiversity conservation. Assuming that the potential for conserving agrobiodiversity is tied to the diversity and dynamics of ethnovariety within the network, we predict that lower modularity, coupled with higher nestedness and connectance, will enhance the potential of the system in conserving agrobiodiversity. Materials and methods Study area The study area lies within the phytogeographic domain of the Atlantic Forest, transitioning into the Cerrado, and is characterized by a local phytophysiognomy type known as Semideciduous Seasonal Forest [ 34 ]. The main economic activities in the region include family farming and artisanal mining. The study was conducted in five quilombola communities: Castro, Embaúbas, Engenho Queimado, Vila Santa Efigênia I and Vila Santa Efigênia II, all located in the rural area of the Mariana city, Minas Gerais, Brazil. Quilombola communities are rural Afro-Brazilian settlements whose inhabitants descend from African and/or indigenous peoples of different origins who took refuge in the territories as a form of resistance to the violence and persecution of European colonizers. They embrace specific ways of life and are nationally recognized as traditional communities since the promulgation of the Brazilian Constitution in 1988 [ 35 ]. The studied communities were officially recognized as “traditional quilombola communities” by the Palmares Cultural Foundation in 2010. However, like many traditional peoples in Brazil, they have not yet received collective land ownership [ 35 ]. The territory is heavily dominated by mining operations like Vale and BHP Billiton companies, facing constant threats, including the collapse of the Fundão dam in 2015, considered the largest environmental crime in Brazil [ 36 ]. This disaster led to the physical and chemical contamination of the Rio Doce Basin by mining waste, which affected the Gualaxo River [ 37 ], a tributary that flows through the study area. As a result, it caused significant damage to agriculture, fishing, and recreational activities for the local residents. Participatory data construction After obtaining approval from the Research Ethics Committee (CEP) of the Universidade Federal de Juiz de Fora (CAAE: 51210421.10000.5147) and registration in the Sistema Nacional de Gestão do Patrimônio Genético e do Conhecimento Tradicional Associado (Sisgen, code: A31BD79), the project was presented to the communities during a meeting with the territorial association, and all participants were invited to sign the Free and Informed Consent Term. The interviews, conducted between December 2021 and March 2022, were realized in the Portuguese quilombolas’ native language, and targeted 48 individuals who were selected for being agro-environmental managers, over 18 years of age, and residents of the territories for at least five years. We used semi-structured interviews [ 38 , 39 ] to access socio-agronomic data, i.e., gender, age, total income, per capita income, total number of people, time living in the community, time in the household, total economic activities, total agricultural activities, total cultivated area, diversity of agro-environments, and diversity and total cultivated ethnovarieties. Additionally, the interviews aimed to understand ethnovariety conservation strategies and the management practices of productive backyards. All interviews were recorded, and a field notebook was used for additional documentation. Subsequently, we applied the free listing methodology [ 38 , 40 ] to detail the richness of plant ethnovarieties related to human nutrition present within the AMUs. Immediately after these two stages, we proceeded to the guided tour [ 38 ]. While walking through their agro-environments, the interviewees were able to (i) supplement the free-listing, and the researcher (I.F.F.) can (ii) gain insights and differentiate agro-environments from the perspective of backyard managers; (iii) map the agro-environments. Thus, using a GPS, we identified each agro-environment and estimated the area managed and cultivated by each family. In this sense, participant observation was also used to enrich the research data [ 41 , 42 ]. Botanical identification was conducted during the guided tour. Ethnovarieties that presented identification challenges were thoroughly photographed (vegetative and reproductive parts) and later identified through bibliographic consultation [ 43 , 44 ], following the APG IV classification system. Specific identification keys [ 45 , 46 ] were also used when necessary. A community guide - Ms. Magda Nascimento - was present throughout all the field research. Her participation was fundamental in describing agro-environments and verifying plant identifications—particularly in cases where distinct local names might refer to the same ethnovariety. As the data collection progressed, she became increasingly involved as a community researcher, playing a key role in describing and characterizing agro-environments and providing a holistic understanding of multi-species relationships [ 47 ]. His data triangulation is essential for dynamically understanding the complexity of socioecological systems [ 9 ]. To characterize the structure and dynamics of seed exchange networks we employed an “open network approach” [ 15 , 22 ] as we included external AMUs entities in the exchange relations, i.e., entities outside the quilombola communities. We denominated the system as ‘seed exchange network’ as it is a commonly used term to couple with the donation and receipt of distinct plant propagules (seed, seedlings, cuttings, and so on). It is important to note that ethnobiological research in traditional communities should not involve exhaustive data collection [ 38 ]. In our case, assessing all seed flows would lead to a very exhaustive survey for both parties, making the methodology logistically difficult and even unethical. To couple with this challenge, after consolidating the free listing of all ethnovarieties cultivated by each internal AMU, we sorted 15 ethnovarieties from each AMU to assess the relations of seed exchanges. After that, we used semi-structured interviews [ 38 , 39 ] to document the social origins of the selected ethnovarieties, understand their qualities and challenges, record exchange events, as well as the AMUs involved in receiving and donating ethnovarieties. Hence, we assessed a socioecological network that encompasses an open and complete system of seed flows. All data collected and analyzed in this manuscript derive from a portion of the first author’s master’s thesis [ 48 ]. Data analysis The data collected in the field were transcribed and categorized into electronic spreadsheets, allowing for the creation of a database that recorded the explanatory variables and responses necessary to test all hypotheses. To test (H1) whether socio-agronomic variables (total income, per capita income, time in the community, time in residence, total economic activities, total agricultural activities, total agro-environments, and total cultivated area) determine the richness of ethnovarieties cultivated by an AMU, a generalized linear model (GLM) was constructed, considering for multicollinearity among the explanatory variables [ 49 – 51 ]. Variables with high correlation indices (greater than 0.6) were excluded, remaning seven of them: total income, per capita income, time in the community, time in residence, total economic activities, total agro-environments, and total cultivated area. We built a GLM with a Poisson distribution, however, due to overdispersion of the residuals, the negative binomial distribution was selected as a better fit for the data [ 49 , 50 ]. The model was then adjusted by progressively removing variables with no significant influence, based on the Akaike Information Criterion (AIC) [ 49 – 51 ]. It was compared to a null model to determine whether the adjusted model was statistically significant [ 49 ]. The open seed exchange weighted network was built considering AMUs as nodes and the seed exchange relations as edges connecting them. The frequency of interactions was represented by distinct exchanges between two AMUs. We employed metrics from ecological and social networks to assess AMUs’ (nodes in the network) importance in the network, i.e., their centrality. We computed degree centrality, betweenness centrality, and harmonic proximity of the AMUs [ 16 ]. To test whether socio-agronomic variables define the centrality of AMUs in the network (H2), we considered as explanatory variables those that were significant when testing H1, along with the richness of cited species. To couple with the high number of predictive variables, we performed a Principal Component Analysis (PCA) with data standardization and Euclidean distance [ 50 , 52 ] to reduce data dimensionality. The analysis revealed that the first two principal components were able to summarize the variables, explaining 75% of the data variance. Consequently, these two components were used as the explanatory variables [ 50 , 52 ] and centrality metrics as response variables (Gaussian distribution) in GLMs models built to test H2. Using the explanatory variables and estimated responses, we developed three GLMs, one for each response variable. These models met the assumptions of variance homogeneity and residual normality, with no issues of over- or under-dispersion of residuals. Additionally, all three models were tested against a null model [ 50 , 52 ]. To test whether the structure of the seed exchange network predicts the local agrobiodiversity conservation (H3), we calculated three network metrics commonly used to characterize social and ecological interaction patterns in complex systems: weighted nestedness, weighted connectance, and modularity. Nestedness refers to a topology in which the network exhibits a hierarchical structure across different organizational levels [ 53 ]. We specifically calculated weighted nestedness with the WNODF metric [ 54 ], which in our study, could indicate that the network is mainly structured by a central cohesive group [ 35 , 55 ] of AMUs that facilitate the majority of seed exchanges, thus suggesting network stability under potential factors that may segregate the communities. Connectance is a widely used metric to assess network connectivity, representing the proportion of actual interactions relative to all possible interactions (the size of the network) [ 56 ]. We computed weighted connectance, wherein a high value would suggest frequent seed flows among the majority of AMUs. To end, we calculate modularity using the Q algorithm with the Dormann-Strauss method [ 57 ]. Modularity measures the extent to which interactions are structured in a modular structure [ 58 ], meaning that subsets of AMUs interact more frequently with each other than with the rest of the network. In our study, high modularity indicates groups of AMUs that share more ethnovarieties among themselves (modules), weakening the sharing among distinct modules and thus challenging ethnovarieties conservation within the network [ 1 , 33 ]. The values of observed nestedness, connectance, and modularity were contrasted with null models; significance was obtained by standardizing the metrics in standard deviations (Z-standardized metric) [ 59 ]. All metrics were calculated using the Bipartite [ 60 ] and Igraph [ 61 ] packages in R software, with functions wherein some metrics are computed as one-mode networks by default (e.g., centrality metrics) and others can be modeled accordingly [ 60 ]. Likewise, all statistical analyses were performed using R version 2022.07.1 [ 62 ]. To visualize the structure of the seed exchange network, we created a graph in Gephi software [ 14 , 17 , 26 ], where the AMUs were represented as nodes, and the edges represent seed exchange relationships. The thickness of the edges reflects the exchanges — i.e., the more exchanges between two AMUs, the thicker the edge. We applied the degree centrality metric to determine the size of the nodes and modularity to color the modules. The ForceAtlas 2 algorithm was used to spatially organize the nodes. Results Socioagronomic variables that influence ethnospecies variety “ I plant for myself , for others , and for the animals ” (par45, 73 years old). The 48 internal AMUs encompass participants whose age ranged from 35 to 86 years (average = 59.4 years ± 12.1 years). On average, they have lived 55 ± 18.3 years in their community and 32 ± 14.7 years in their current residence. Each AMU consists of an average of 1.4 people, with a per capita income of R$ 1,448.15. In Brazil, the average for 2021 was 2.79 people per household and a per capita income of R$1,367.00 [ 63 ]. The family economy is supported by various non-agricultural activities (an average of 2.85 non-agricultural activities per household) and government benefits (received by 43 participants 85.4%). The AMUs engage in diverse agricultural systems, encompassing an average of 5.77 activities, including: (1) Agriculture proper: managing gardens, backyards, roçados fields (“roçados”), swine fields, areas dedicated to the production of annual crops such as corn and cassava or semi-perennial crops such as sugarcane; (2) Raising pigs; (3) Raising chickens; (4) Dairy production; (5) Beef cattle; (6) Seedling production; (7) Field weeding services; (8) Exchange relationships and sales of surplus production. These activities take place across six main agro-environments (Fig. 2 ), namely: (A) “Terrace”: The area near the residences where delicate ornamental and food plants, such as herbs, are concentrated, requiring more care; (B) “Orchard”: A space for cultivating citrus and fruit-bearing shrubs, which holds significant dietary, social, and recreational importance; (C) “Agroforestry backyard”: Humid, shaded soil covered by leaf litter, where secondary forest can develop, used to grow economically important plants such as bananas ( Musa x paradisiaca L.), taro ( Colocasia esculenta L.), and mangoes ( Mangifera indica L.); (D) “Roça or chácara”: Red latosol, typically fertilized with cow manure, comprising larger areas further from residences, used for intercropping beans, corn, squash, as well as sugarcane and cassava- essential crops for the inhabitants’ food security (E) “Garden”: Soil rich in organic matter, closer to the house and fenced, designated for the continuous cultivation of vegetables, medicinal plants, and herbs; (F)”Piçarra land”: Stony, yellowish, dystrophic, and acidic soil that supports limited crop diversity, primarily peanuts ( Arachis hypogaea L.) and pineapples ( Ananas comosus (L.) Merril). Fig. 2. Open in a new tab Systematization of Agro-environments - ( A ) Terreiro, ( B ) Orchard, ( C ) Agroforestry Backyard, ( D ) Roça or Chácara, ( E ) Garden, and ( F ) Piçarra Land - managed by the quilombola communities of Castro, Embaúbas, Engenho Queimado, Vila Santa Efigênia I, and Vila Santa Efigênia II in the municipality of Mariana, Minas Gerais, Brazil Each family manages, on average, 4.12 agro-environments and has approximately 0.225 hectares (sd = 0.19) of arable land. Despite the high standard deviation (95.8%) — indicating that some AMUs possess significantly more land than others — our research partners expressed satisfaction with the size of their lands. The agricultural calendar is closely aligned with the water cycle, with the planting of “roças” and fruit trees during the rainy season, specifically from September to November. By the end of this period, in March, as the dry season sets in, garden management is carried out until the next rainy season. The products harvested from the agro-ecosystems are used for family self-consumption, with any surplus typically sold in the city of Mariana. In the studied territories, the selection of food plants for management is based on their nutritional value, economic interest, ease of cultivation, taste memory, and emotional connections to the ethnovarieties. From this agricultural system, 1,919 use citations were recorded across 359 ethnovarieties that represent 134 species from 44 botanical families. The most prominent families were Brassicaceae (35 varieties), Poaceae (28), Rutaceae (26), and Fabaceae (25). Seventeen botanical families had only one or two cultivated varieties. The species with the greatest variety were kale ( Brassica oleracea L.) (24 varieties), sugarcane ( Saccharum officinarum L.) (17 varieties), banana ( Musa x paradisiaca L.) (17 varieties), orange ( Citrus sinensis L.), and avocado ( Persea americana Mill.) (10 varieties). Sixty-six species had only one or two varieties. When considering use citations, banana was the most frequently cited (182 citations), followed by orange (96 citations), kale (94 citations), mango ( Mangifera indica L.) (79 citations), yam ( Colocasia esculenta L.) (69 citations), and cassava ( Manihot esculenta L.) (67 citations). Thirty-nine species were cited by only one or two research participants. When considering the total number of citations for specific varieties, the most notable were coquinho mango ( Mangifera indica L.) (41 citations), lobrobó ( Pereskia aculeata Mill.) (40 citations), purple-skinned cassava ( Manihot esculenta L.) (37 citations), campista orange ( Citrus sinensis L.) (35 citations), apple banana ( Musa x paradisiaca L.) (34 citations), devil’s lime ( Citrus limonia Osbeck) (32 citations), silver banana ( Musa x paradisiaca L.) (32 citations), red acerola ( Malpighia emarginata DC.) (31 citations), and jabuticaba ( Plinia peruviana (Poir.) Govaerts) (31 citations). Notably, 195 ethnovarieties were cited by only one or two research partners. On average, we found a richness of 40 cultivated ethnovarieties per AMU. Of the 359 ethnovarieties, 53 are cultivated but not maintained by an AMU for replanting in the next agricultural cycle. Instead, when the time comes to plant again, these ethnovarieties are shared by a nearby AMU, recognized as the seed keeper. Additionally, 39 ethnovarieties are maintained by only one or two AMUs. None of the cited varieties were cultivated across all six recorded agro-environments. The most versatile varieties, cultivated in five agro-environments, were purple-skinned sweet potato ( Ipomoea batatas (L.) Lam.), purple-skinned cassava ( Manihot esculenta L.), and common and purple sugarcane ( Saccharum officinarum L.), all cultivated in gardens, backyards, orchards, fields, or piçarra land. White guava also stood out, being cultivated in the same environments, though replacing the garden with the yard. A total of 213 varieties (~ 59%) were cultivated in only one agro-environment. Regarding potentially arable land, i.e., the total area declared by all research participants where a variety can be cultivated, coquinho mango ( M. indica ) stood out with 0.058 km², followed by devil’s lime (0.046 km²), jabuticaba (0.041 km²), white and red guava (0.041 km²), and silver banana (0.044 km²). The varieties most highly regarded for their attributed qualities (e.g., high productivity and organoleptic quality) were purple-skinned cassava (21 citations), acerola (16 citations), coquinho mango (15 citations), apple banana (13 citations), lobrobó (13 citations), and yam (12 citations). On the other hand, the varieties with the most challenges, due to less favorable traits such as low productivity, high water demand, or bitter taste, were lettuce ( Lactuca sativa L.), yam, candogueira tangerine ( Citrus unshiu (Mak.) Marcov.), and pitanga ( Eugenia uniflora L.), each with four citations. We have found that total agroenvironments (estimate = 3.057e-013; p = 0.0001), total cultivated area (estimate = 6.201e-05; p = 0.0173), and time in the community (estimate = −6.390e-03; p = 0.0466) positivelly influenced species richness (AIC = 385.82 and R²=0.79). Thus, as the diversity of agroenvironments, cultivated area, and time of residence in the community increase, so does the richness of managed ethnovarieties (Fig. 3 ). The generalized linear model resulted in the following equation: Variety richness (y) = 2.558e + 00–6.390e-03 * time in community (x1) + 3.057e-013 * total agroenvironments (x2) + 6.201e-05 * total cultivated area (x3). Therefore, our first hypothesis was corroborated. The socio-agronomic context — especially agro-environmental diversity, cultivated area, and length of stay in the community — positively influenced the richness of cultivated ethnovarieties. These results confirm that diversified management, over a larger area, and longer stays in the territory favor the conservation of agrobiodiversity on the property. Fig. 3. Open in a new tab Generalized linear models showing the effect of ( a ) total managed agroenvironments, ( b ) total cultivated area (cul), ( c ) and time in the community on the richness of agricultural ethnovarieties cultivated in five quilombola communities in the municipality of Mariana, Minas Gerais, Brazil Socio-agronomic variables and the richness of cultivated varieties influence the importance of AMUs The PCA with the variables of time living in the community, total of managed agroenvironments, total cultivated area, and species richness explained 74.44% of data variation (Fig. 4 a). The first principal component accounted for 50.06% of the variation, with the most representative variables being total agroenvironments and total varieties. The second principal component accounted for 24.38% of the variation, with time in the community and total varieties as the most representative variables. The degree centrality of the AMUs (Fig. 4 b) was positively influenced by the socio-agronomic variables of time living in the community, total agroenvironments managed, cultivated area, and the richness of cultivated ethnovarieties ( p = 0.0002, AIC = 293.05). Likewise, betweenness centrality ( p = 0.0109, AIC = 777.16; Fig. 4 c) and harmonic closeness centrality ( p = 0.00779, AIC = 365.24; Fig. 4 d) were positively explained by the first principal component. Fig. 4. Open in a new tab Models used to assess the relationship between socio-agronomic variables and the richness of crop varieties on the centrality of agrobiodiversity management units (AMUs) in five quilombola communities in Mariana, Minas Gerais, Brazil. ( a ) Principal Component Analysis (PCA) incorporating total cultivated area, total managed agroenvironments, time living in the community, and richness of agricultural varieties cultivated; ( b ) Generalized linear models demonstrating the positive effect of the first axis of PCA on degree centrality, ( c ) betweenness centrality, and ( d ) harmonic closeness centrality Thus, our second hypothesis was also corroborated. We found that socio-agronomic variables and ethnovariety richness, together, determined the structural position of the AMUs within the seed exchange network. The AMUs that manage larger and more diverse areas occupied more central and connecting roles, reinforcing their importance as bridges for the circulation of seeds and knowledge within and between communities. The structure of the seed exchange network predicts the potential of agrobiodiversity conservation A total of 137 external (outside the communities) and 48 internal AMUs (individuals and families from the quilombola communities), formed a complete seed exchange network with 185 AMUs (Fig. 5 ). Among the studied quilombola territories, Vila Santa Efigênia II accounted for the highest number of exchange events (23.7%), followed by Castro (18.3%), Vila Santa Efigênia I (11.4%), Embaúbas (7.4%), and Engenho Queimado (7.15%). Regarding external municipalities or districts with the higher rate of seed exchanges with the quilombola communities, Mariana accounted for 8.3% (distant 25,2 km from the quilombola territory), Furquim with 3.5% (distant 4,7 km), and Acaiaca with 3.1% (distant 16,8 km). Through the interviews, we identified 359 ethnovarieties generated and maintained by the quilombola territories. However, the 424 seed exchange events were documented during the second analytical phase, when 15 ethnovarieties per AMU were randomly selected for detailed discussion in the interviews. This sample represented a balanced subset of the agrobiodiversity managed in the territories and was designed to capture the main exchange pathways while maintaining a participatory and non-exhaustive approach, respecting the ethical limits of field research with traditional communities. Fig. 5. Open in a new tab Seed exchange network among the quilombola communities of Castro, Embaúbas, Engenho Queimado, and Vila Santa Efigênia (I and II) in the municipality of Mariana, Minas Gerais, Brazil. The nodes represent agrobiodiversity managing units (AMUs), with node size indicating degree centrality and color representing the network’s emerging modules. Each edge corresponds to the frequency of exchange between partners The high richness of ethnovariety formed a network with low and significant degree of nestedness (WNODF = −2.20, z-score = −12.29). Similarly, connectance was extremely low and significant (C = 0.022; z-score = −2.73). On the other hand, we observed a modular structure that was significantly greater than expected by chance (Q = 0.60; z-score = 42.65). Thus, our third hypothesis was not corroborated. Although the network demonstrated a high potential for agrobiodiversity conservation, the emerging properties were contrary to what was expected. Discussion The studied quilombola communities encompass in their territories a high richness of food ethnovarieties managed within their agroenvironments, which are organized into a complex seed exchange network. We found that as the number of agroenvironments, cultivated areas, and duration of the residency in the community increases, so does the richness of ethnovarieties maintained by the AMUs. Socio-agronomic variables and the richness of cultivated ethnovarieties positively influenced the importance of AMUs in the network, i.e., their centrality. Further, we observed that the seed exchange network displayed low nestedness, low connectance, and high modularity. This structure suggests a segregated pattern of seed flow with potential consequences for the genetic pool of agrobiodiversity, reflecting the relationships among communities and their inhabitants in terms of life histories and socio-environmental contexts. The high richness of 359 ethnovarieties reflects the diverse life stories and idiosyncratic characteristics of research partners. Also, the edaphoclimatic complexity of the agroenvironments seems to influence the heterogeneity of agrobiodiversity cultivated, which encompasses plant varieties that are highly adapted to the local environmental niches. This pattern of high plant richness was similarly observed in other Brazilian quilombola communities [ 35 , 64 ]. Our research also demonstrated that 39 ethnovarieties (ca. 11%) are maintained by only one or two research partners. This result is linked to the social prestige associated with agrobiodiversity guardians and the biocultural value of these plants, indicating that the distribution of ethnovarieties across the territory is not homogeneous. In this sense, possessing, and conserving rare varieties symbolizes specialized knowledge, access to differentiated exchange networks and the ability to maintain ancient agricultural lineages [ 13 , 21 ]. We also found that 53 ethnovarieties (ca. 59%) have only one or two guardians responsible for maintaining propagules for future plantings across all AMUs in the territory. These findings highlight the fact that AMUs in the territory already depend on exchange relationships to preserve a great portion of their ethnovarieties within their agroenvironments. While it is not possible to say with certainty whether all rare ethnovarieties are actually being exchanged within the network, some of them were mentioned during the interviews as being donated or shared among neighbors. When testing the effect of socio-agronomic variables on the richness of crop varieties cultivated by an AMU, we found that the diversity of agroenvironments positively influences the diversity of managed species. Specifically, the three AMUs that had six agroenvironments, yard, orchard, agroforestry backyard, roça or chacara, garden, and piçarra land, managed an average of 90 varieties (SD = 26). This result demonstrates that the heterogeneity of edaphoclimatic conditions enables the maintenance of crops by increasing the potential for experimentation and conservation, which in turn fosters plant diversification [ 65 ]. These findings provide a practical illustration of the ongoing coevolutionary process within traditional agroenvironments [ 66 ]. We also observed a positive influence of total cultivated area and the richness of managed ethnovarieties, suggesting that larger areas offer more opportunities for experimentation. In fact, staple crops - essential for food and nutritional security - such as beans, corn, and squash, require certain spaces within the system. As follows, larger properties have the opportunity to introduce new ethnovarieties; for example, with sufficient space to test a new bean variety without the risk of cross-fertilization or to explore a new crop that has never been produced before. So even if the new ethnovariety yields zero, the AMU doesn’t lose what is essential for its food security. Therefore, larger land areas create surplus spaces for cultivating and managing additional food species. Additionally, larger areas increase the potential for interaction with neighboring households, facilitating seed exchanges between families. As follows, the size of the AMUs associated with closer proximity to neighbors is determinant for conserving agrobiodiversity, as it reduces the time needed to acquire specific varieties, making seed procurement more economically feasible while strengthening affective and trust-based relationships [ 29 ]. Another relationship identified was the positive influence of the time of residence in the territory on the total number of crop varieties managed. This result should be interpreted with caution, since only a few households have been residing in the area for less than 30 years. However, this pattern does not appear to result from sampling bias, as the analyses were structured to minimize causal interferences and to control for multicollinearity among socio-agronomic variables. This result suggests that the time in the community plays a key role in increasing the genetic pool richness of crop varieties, which are accumulated, tested, and adapted to specific agroenvironments [ 65 ]. Although farming practices may become more challenging with age, the AMUs led by older farmers carefully select which varieties to manage, prioritizing those most suited to their needs and conditions. These older farmers become key references for crop species that are easier to manage (e.g., “chifre de veado” okra, which is easier to harvest), tied to gustatory memory (e.g., “manteiga” collard greens, a variety consumed since early childhood), and/or associated with emotional connections (e.g., “puteca” beans, a variety that helped the community during times of food scarcity). Together, these factors build trust and elevate the status of older residents in their territories, making them vital guardians of agrobiodiversity [ 21 ]. Conversely, the total number of economic and agricultural activities, as well as total and per capita income, did not influence the management of agrobiodiversity. This finding may be due to the low variation in income and financial activities by the families. Additionally, the period of time living in the current residence did not impact the richness of crop varieties cultivated. This is likely because quilombola community members tend to carry their crop varieties when they relocate, underscoring the socioecological importance of these varieties to the community. Lastly, the total number of people in the household did not influence the richness of managed crop varieties, indicating that, regardless of household size, the AMU diversity is maintained by a few integrants, though the quantity of food produced may vary. Another key finding of this study is that some socio-agronomic variables, such as the total number of agroenvironments, total cultivated area, time of residence in the community, and richness of managed crop varieties, positively affect the importance of AMUs in the territory, i.e., their centrality in the network. These results point out which socio-agronomic variables are prioritary to define the importance of AMUs to unveil the structure and dynamics of the seed exchange network. Although we did not explicitly examine whether the most central AMUs also engage in seed exchanges over longer geographical distances or between distinct villages, qualitative observations suggest that some of these highly connected households often participate in inter-community exchanges, especially during fairs, visits, or kinship-related events. This study, therefore, goes beyond the current knowledge, demonstrating that AMUs that act as agrobiodiversity guardians are those that share the most crop varieties and serve as bridges between different territories, facilitating the flow of both genetic resources and knowledge within the territory [ 14 , 18 , 19 , 21 , 26 , 29 ]. Only one institution outside the communities stood out in the methods used to identify key nodes in the network. This was the Paulo Freire Agricultural Family School in the municipality of Acaiaca, which welcomes students from the communities and runs several workshops on socio-environmental issues, promoting seed exchanges within the communities. This demonstrates the importance of events that foster social mobilization for agrobiodiversity conservation. When examining the influence of the emergent proprieties of the seed exchange relations on the potential for agrobiodiversity conservation, we found that the network exhibited low nestedness, low connectance, and high modularity. This structural pattern suggests that AMUs form cohesive subgroups that exchange seeds more frequently with specific partners than with others outside their group. This segregated pattern indicates a centralization of the genetic pool of agrobiodiversity. This segregation in seed exchanges can be explained by factors such as geographic proximity and sociobiocultural relationships, which align with the principle of homophily. The finding that the communities of Vila Santa Efigênia II and Castro account for the majority of exchange events (23.7% and 18.3%, respectively) can be attributed to the higher concentration of AMUs that are geographically and kinship-wise close to one another. In contrast, the communities of Embaúbas and Engenho Queimado (which account for 7.4% and 7.15% of seed exchange relationships) exhibit more spatial segregation between their AMUs, despite familial ties. Thus, this structural pattern suggests that AMUs form cohesive subgroups interacting more frequently within their group than with other AMUs in the system. We have found that external connections with cities and districts outside quilombola territories (e.g., Mariana and Acaiaca municipalities) contributed to almost half of seed flows (47% of events). This result is quite impressive as it emphasizes the socio-ecological service provided by the quilombola communities in maintaining and distributing the agrobiodiversity to the whole region. Through this role, they contribute to agricultural, food, and nutritional security and sovereignty over an extensive geographic region. In contrast, the seed exchange network exhibits emergent properties that suggest some fragility. Although there is a high potential for connections among various actors from inside and outside quilombola territory (i.e., the high richness of crop ethnovarieties), the low connectance of the network increases the risk of agrobiodiversity loss. Low levels of nestedness and connectance in socio-ecological networks may suggest low resilience of the system, for instance, its capability to absorb and adapt to climate changes, compromising peer trust and the preservation of traditional knowledge [ 26 , 50 ]. As a result, the collective memory of the territories is weakened [ 55 , 67 ]. In fact, high modularity in socioecological networks indicates that exchange dynamics depend on distinct subgroups that are poorly connected. As a consequence, the network becomes more vulnerable to disturbances, as crop varieties are unevenly distributed in the territory. The loss of a variety within a particular group (module) can have a significant impact on the whole system, given the lower likelihood of exchanges between distinct parts of the territory (i.e., between modules) [ 1 , 33 ]. This segmented structure seems to reflect geographic fragmentation of territories that likely limits the introduction of new crop varieties into the system, thereby restricting innovation and socio-environmental adaptations [ 33 ]. Conclusion We can conclude that historical and material conditions play a critical role in the “on-farm” conservation of agrobiodiversity. These findings underscore the importance of land ownership and territorial security for traditional communities to continue preserving their way of life, which is rooted in the ecological capital, i.e., non-economic exchanges between humans and nature, which generates coevolution [ 6 ]. These features involving land recognition are especially relevant given that their territories are situated in a region experiencing intense anthropogenic impacts, primarily through mining. Therefore, it is crucial to protect and recognize quilombola communities as an ontologically sustainable system, essential for preserving their socio-ecological knowledge [ 68 ]. Based on our results, we propose three actions to improve quilombola agrobiodiversity management. First, local seed exchange relations should be formally recognized as legitimate conservation mechanisms, and their guardians acknowledged as rights-bearing agents who safeguard and maintain agrobiodiversity. Second, policies must support the creation and strengthening of community seed banks, fairs, and exchange events to enhance autonomy and access to diverse genetic resources. Third, educational programs in rural and quilombola schools — linked to local food distribution chains that should be supported by institutional procurement of quilombola products — can strengthen local economies and youth engagement in agriculture. To our knowledge, this is the first study that provides practical and feasible initiatives to be adopted by traditional communities and their legal regulations, to promote agrobiodiversity and food sovereignty. Supplementary Information Supplementary Material 1 (34.6KB, docx) Acknowledgements Our deepest gratitude goes to all those who welcomed us so generously into their homes, to the Quilombola Association of Vila Santa Efigênia and Adjacent Areas, and to the Saberes do Território Collective, whose support was essential for the development of this research. Author contributions The conception and design of the study were contributed by GTS, IFF and FVC. IFF conducted the data collection and cleaning. FVC and GTS performed the data analysis. IFF wrote the first draft of the manuscript, TSN and FRG developed a critical review of the article, and both authors commented on subsequent versions. All authors reviewed and approved the final manuscript. Funding The fieldwork activities were funded by the research project “Medicinal and Useful Plants of the Rio Doce Basin” (CAPES 881.118042/2016-01). Authors IFF, FVC, and TSN received scholarships from CAPES. Data availability The datasets used and/or analysed during the current study are available on the open science platform Zenodo (DOI: 10.5281/zenodo.18166370). Code Availability The data collected from the interviews were organized in Microsoft ® Excel and analyzed using R software. The codes (custom code) are available from the authors upon reasonable request. Declarations Ethics approval and consent to participate The project was approved by the Ethics Committee for Research (CEP) of the Federal University of Juiz de Fora (CAAE: 51210421.10000.5147) and registered in Sisgen (code: A31BD79). Consent for publication Each farmer signed a Free and Informed Consent Term (TCLE), required by Brazilian Law 13.123/2015, which addresses access to traditional knowledge associated with biodiversity. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References 1. Labeyrie V, et al. Networking agrobiodiversity management to foster biodiversity-based agriculture: a review. Agron Sustain Dev. 2021. 10.1007/s13593-020-00662-z. [ Google Scholar ] 2. Eloy L, et al. Os Sistemas agrícolas tradicionais Nos interstícios Da Soja no brasil: processos e limites Da conservação Da agrobiodiversidade. Confins. 2020;45. 10.4000/confins.28182. 3. Emperaire L. Agrobiodiversidade e roças. In: Cunha MC, Magalhães SB, Adams C, eds. Povos tradicionais e biodiversidade no Brasil: contribuições dos povos indígenas, quilombolas e comunidades tradicionais para a biodiversidade, políticas e ameaças. São Paulo: SBPC; 2021. pp. 14–55. Accessed 26 Sep 2022. Available from: http://portal.sbpcnet.org.br/livro/povostradicionais7.pdf 4. Rodríguez CS. El Despojo de La Riqueza biológica: de patrimonio de La Humanidad a recurso Bajo soberanía Del Estado. 1st ed. Heredia: EUNA; 2013. [ Google Scholar ] 5. Santonieri L, Bustamante PG. Conservação ex situ e on farm de recursos genéticos: desafios Para promover sinergias e complementaridades. Bol Mus Para Emílio Goeldi Cienc Hum. 2016;11(3):677–90. 10.1590/1981.81222016000300008. [ Google Scholar ] 6. van der Ploeg JD. Sete teses sobre a agricultura camponesa. In: Petersen P, org. Agricultura familiar camponesa na construção do futuro. Rio de Janeiro: AS-PTA; 2009. pp. 17–32. ISBN: 978-85-87116-14-7. 7. Santilli J. Agrobiodiversidade e direitos Dos agricultores. 1st ed. São Paulo: Peirópolis; 2009. [ Google Scholar ] 8. Badstue LB, et al. Examining the role of collective action in an informal seed system: a case study from the Central Valleys of Oaxaca, Mexico. Hum Ecol. 2006;34(2):249–73. 10.1007/s10745-006-9016-2. [ Google Scholar ] 9. Pautasso M, et al. Seed exchange networks for agrobiodiversity conservation: a review. Agron Sustain Dev. 2013;33(1):151–75. 10.1007/s13593-012-0089-6. [ Google Scholar ] 10. Kluger LC, Gorris P, Kochalski S, Mueller MS, Romagnoni G. Studying human–nature relationships through a network lens: a systematic review. People Nat. 2020;2(4):1100–16. 10.1002/pan3.10136. [ Google Scholar ] 11. Calvet-Mir L, Calvet-Mir M, Molina JL, Reyes-García V. Seed exchange as an agrobiodiversity conservation mechanism: a case study in Vall Fosca, Catalan Pyrenees, Iberian Peninsula. Ecol Soc. 2012. 10.5751/ES-04682-170129. [ Google Scholar ] 12. Martins PS. Biodiversity and agriculture: patterns of domestication of Brazilian native plant species. Acad Bras Cienc. 1994;66:26–219. [ Google Scholar ] 13. Emperaire L, Eloy L, Seixas AC. Redes e observatórios Da agrobiodiversidade: Como e Para quem? Uma abordagem exploratória Na região de Cruzeiro do Sul, acre. Bol Mus Para Emílio Goeldi Cienc Hum. 2016;11(1):159–92. 10.1590/1981.81222016000100009. [ Google Scholar ] 14. Abizaid C, Coomes OT, Perrault-Archambault M. Seed sharing in Amazonian Indigenous rainforest communities: a social network analysis in three Achuar villages, Peru. Hum Ecol. 2016;44(5):577–94. 10.1007/s10745-016-9852-7. [ Google Scholar ] 15. Labeyrie V, Thomas M, Muthamia ZK, Leclerc C. Seed exchange networks, ethnicity, and sorghum diversity. Proc Natl Acad Sci U S A. 2016;113(1):98–103. 10.1073/pnas.1513238112. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 16. Ricciardi V. Social seed networks: identifying central farmers for equitable seed access. Agric Syst. 2015;139:110–21. 10.1016/j.agsy.2015.07.002. [ Google Scholar ] 17. Rodier C, Struik PC. Nodal farmers’ motivations for exchanging sorghum seeds in northwestern Ethiopia. Sustainability. 2018. 10.3390/su10103708. [ Google Scholar ] 18. Song Y, et al. Network analysis of seed flow, a traditional method for conserving tartary buckwheat ( Fagopyrum tataricum ) landraces in Liangshan, Southwest China. Sustainability. 2019. 10.3390/su11164263. [ Google Scholar ] 19. Subedi A, et al. Who maintains crop genetic diversity and how? Implications for on-farm conservation and utilization. Culture & Agriculture. 2003;25:41–50. [ Google Scholar ] 20. Subedi A, Garforth C. Gender, information and communication networks: implications for extension. Eur J Agric Educ Ext. 1996;3(2):63–74. 10.1080/13892249685300201. [ Google Scholar ] 21. Thomas M, Caillon S. Effects of farmer social status and plant biocultural value on seed circulation networks in Vanuatu. Ecol Soc. 2016. 10.5751/ES-08378-210213. [ Google Scholar ] 22. Violon C, Thomas M, Garine E. Good year, bad year: changing strategies, changing networks? A two-year study on seed acquisition in Northern Cameroon. Ecol Soc. 2016. 10.5751/ES-08376-210234. [ Google Scholar ] 23. Abay F, De Boef W, Bjørnstad Å. Network analysis of barley seed flows in Tigray, Ethiopia: supporting the design of strategies that contribute to on-farm management of plant genetic resources. Plant Genet Resour. 2011;9(4):495–505. 10.1017/S1479262111000773. [ Google Scholar ] 24. Otieno GA, Lacasse H, Fadda C, Reynolds TW, Recha JW. Social seed networks for climate change adaptation in Western Kenya: results from a study to better understand farmers’ primary sources of seed information in the Nyando climate-smart villages. 2018. Available from: https://hdl.handle.net/10568/93210 25. Seboka B, Deressa A. Validating farmers’ Indigenous social networks for local seed supply in central rift Valley of Ethiopia. J Agric Educ Ext. 1999;6(4):245–54. 10.1080/13892240085300071. [ Google Scholar ] 26. Otieno G, Zebrowski WM, Recha J, Reynolds TW. Gender and social seed networks for climate change adaptation: evidence from bean, finger millet, and sorghum seed systems in East Africa. Sustainability. 2021;13(4):1–24. 10.3390/su13042074. [ Google Scholar ] 27. Fenzi M, Rogé P, Cruz-Estrada A, Tuxill J, Jarvis D. Community seed network in an era of climate change: dynamics of maize diversity in Yucatán, Mexico. Agric Hum Values. 2022;39(1):339–56. 10.1007/s10460-021-10249-3. [ Google Scholar ] 28. Kawa NC, McCarty C, Clement CR. Manioc varietal diversity, social networks, and distribution constraints in rural Amazonia. Curr Anthropol. 2013;54(6):764–70. 10.1086/673528. [ Google Scholar ] 29. Llamas-Guzmán LP, Lazos Chavero E, Perales Rivera HR, Casas A. Seed exchange networks of native maize, beans, and squash in San Juan Ixtenco and San Luis Huamantla, Tlaxcala, Mexico. Sustainability. 2022. 10.3390/su14073779. [ Google Scholar ] 30. Pinto MFC. Caminhos Da agrobiodiversidade: Redes de Trocas de Sementes Em Sistemas agroecológicos Na Serra Catarinense, Alto Vale do Rio Tijucas, Santa Catarina. Dissertação. Universidade Federal de Santa Catarina; 2014. 31. Poudel D, Sthapit B, Shrestha P. An analysis of social seed network and its contribution to on-farm conservation of crop genetic diversity in Nepal. Int J Biodivers. 2015;2015:1–13. 10.1155/2015/312621. [ Google Scholar ] 32. Pautasso M. Network simulations to study seed exchange for agrobiodiversity conservation. Agron Sustain Dev. 2015;35(1):145–50. 10.1007/s13593-014-0222-9. [ Google Scholar ] 33. Barbillon P, Thomas M, Goldringer I, Hospital F, Robin S. Network impact on persistence in a finite population dynamic diffusion model: application to an emergent seed exchange network. J Theor Biol. 2015;365:365–76. 10.1016/j.jtbi.2014.10.032. [ DOI ] [ PubMed ] [ Google Scholar ] 34. IBAMA. Laudo técnico preliminar: impactos ambientais decorrentes do desastre envolvendo o rompimento da barragem de Fundão, em Mariana, Minas Gerais. Sinapse Múltipla. 2015;(1):38. Available from: https://www.ibama.gov.br/phocadownload/barragemdefundao/laudos/laudo_tecnico_preliminar_Ibama.pdf 35. Gonçalves MC, et al. Traditional agriculture and food sovereignty: Quilombola knowledge and management of food crops. J Ethnobiol. 2022;42(2):241–60. 10.2993/0278-0771-42.2.241. [ Google Scholar ] 36. Fabrício SA, Ferreira DDM, Borba JA. A panorama of Mariana and Brumadinho disasters: what do we know so far? REAd Revista Eletrônica de Administração (Porto Alegre). 2021;27(1):128–52. 10.1590/1413-2311.310.102806. [ Google Scholar ] 37. Santos EM, Nalini HA, Abreu AT, Nogueira LB. Mobilization of heavy metals in river sediments from the region impacted by the Fundão dam rupture, Brazil. Environ Earth Sci. 2021. 10.1007/s12665-021-10107-9. [ Google Scholar ] 38. Albuquerque UP, Ramos MA, Lucena RFP, Alencar NL. Methods and techniques used to collect ethnobiological data. In: Albuquerque UP, Cruz LVF, Lucena RFP, Alves RRN, editors. Methods and techniques in ethnobiology and ethnoecology. 1st ed. New York: Springer; 2014. pp. 15–38. [ Google Scholar ] 39. Alexiades MN. Collecting ethnobotanical data: an introduction to basic concepts and techniques. In: Alexiades MN, ed. Selected guidelines for ethnobotanical research: a field manual. New York: The New York Botanical Garden; 1996. pp. 53–94. Available from: https://www.researchgate.net/publication/304496311 40. Thompson EC, Juan Z. Comparative cultural salience: measures using free-list data. Field Methods. 2006;18(4):398–412. 10.1177/1525822X06293128. [ Google Scholar ] 41. Amorozo MCM, Viertler RB. A abordagem qualitativa Na Coleta e análise de Dados Em etnobiologia e etnoecologia. In: Albuquerque UP, Lucena RFP, Cunha LVFCC, editors. Métodos e técnicas Na pesquisa etnobotânica. 3rd ed. Recife: NUPPEA; 2010. pp. 65–82. [ Google Scholar ] 42. Combessie JC. O método Em sociologia: o Que é, Como se Faz. São Paulo: Edições Loyola; 2004. [ Google Scholar ] 43. Brasil. Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Atenção Básica. Alimentos regionais brasileiros. 2nd ed. Brasília: Ministério da Saúde; 2015. 484 p. ISBN: 978-85-334-2145-5. 44. Lorenzi H, Lacerda MTC, Bacher LB. Frutas no Brasil Nativas e exóticas: de Consumo in natura. São Paulo: Plantarum; 2015. [ Google Scholar ] 45. Carvalho SZ, Mendes L, Bonini DM. Etnoconhecimento de plantas de uso medicinal por benzedeiras/benzedores e rezadeiras/rezadores de Anhembi e Mogi das Cruzes–SP. Rev Eletrônica Correlativo. 2017;16(2):133–52. 10.15603/1677-2644/CORRELATIO.V16N2P133-152. [ Google Scholar ] 46. Heiden G, Barbieri L, Neitzke RSN. Chave para a identificação das espécies de abóboras (Cucurbita, Cucurbitaceae) cultivadas no Brasil. Embrapa Clima Temperado; 2007. Available from: https://www.infoteca.cnptia.embrapa.br/infoteca/handle/doc/745868 47. Kirksey SE, Helmreich S. The emergence of multispecies ethnography. Cult Anthropol. 2010;25(4):545–76. 10.1111/j.1548-1360.2010.01069.x. [ Google Scholar ] 48. Fantini IF, Soldati GT, Salimena FRG. Rede de Trocas de Sementes e conservação Da agrobiodiversidade: Um Estudo Em comunidades quilombolas no município de Mariana, Minas Gerais. Dissertação. Universidade Federal de Juiz de Fora; 2023. 49. Diniz SE, Thiele J. Modelos de regressão Em R. Clube de Autores; 2021. 50. Da Silva F, Gonçalves-Souza T, Paterno G, Provete D, Vancine M. Análises ecológicas no R. Recife: NUPPEA; 2022. [ Google Scholar ] 51. Zuur A, Leno EN, Smith GM. Analyzing ecological data: Springer; 2007. 52. Borcard D, Gillet F, Legendre P. Numerical ecology with R. 2011. [ Google Scholar ] 53. Girvan M, Newman MEJ. Community structure in social and biological networks. Proc Natl Acad Sci U S A. 2002;99(12):7821–6. 10.1073/pnas.122653799. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 54. Almeida-Neto M, et al. A consistent metric for nestedness analysis in ecological systems: reconciling concept and measurement. Oikos. 2008;117:1227–39. 10.1111/j.2008.0030-1299.16644.x. [ Google Scholar ] 55. Bascompte J, Jordano P, Melian CJ, Olesen JM. The nested assembly of plant-animal mutualistic networks. Proc Natl Acad Sci U S A. 2003;100(16):9383–7. 10.1073/pnas.1633576100. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 56. Poisot T, Gravel D. When is an ecological network complex? Connectance drives degree distribution and emerging network properties. PeerJ. 2014;2:e251. 10.7717/peerj.251. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 57. Beckett SJ. Improved community detection in weighted bipartite networks. R Soc Open Sci. 2016;3(1):140536. 10.1098/rsos.140536. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Newman MEJ. Modularity and community structure in networks. Proc Natl Acad Sci U S A. 2006;103(23):8577–82. 10.1073/pnas.0601602103. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 59. Miyauchi A, Kawase Y. Z-score-based modularity for community detection in networks. PLoS One. 2016;11(1):e0147805. 10.1371/journal.pone.0147805. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 60. Dormann C et al. bipartite: visualising bipartite networks and calculating some (ecological) indices. CRAN. 2025. Accessed 29 Dec 2025. Available from: https://github.com/biometry/bipartite 61. Csárdi G, Nepusz T, Traag V, Horvát S, Zanini F, Noom D. igraph reference manual. Accessed 29 Dec 2025. Available from: https://igraph.org 62. R Core Team. R: a language and environment for statistical computing. Accessed 29 Dec 2025. Available from: https://www.R-project.org 63. Instituto Brasileiro de Geografia e Estatística (IBGE). Censo demográfico 2022. Rio de Janeiro. 2022. Accessed 29 Dec 2025. Available from: https://censo2022.ibge.gov.br/ 64. Duarte Almada ED. Entre as serras: etnoecologia de duas comunidades quilombolas no Sudeste brasileiro. Tese (Doutorado). Universidade Estadual de Campinas (Unicamp); Campinas, SP; 2012. 239 p. Available from: https://hdl.handle.net/20.500.12733/1619271 65. Berkes F, Colding J, Folke C. Rediscovery of traditional ecological knowledge as adaptive management. Ecol Appl. 2000;10(5):1251–62. 10.1890/1051-0761(2000)010[1251:ROTEKA]2.0.CO;2. [ Google Scholar ] 66. Altieri M. Agroecologia: bases científicas Para agricultura sustentável. 3rd ed. São Paulo/Rio de Janeiro: Expressão Popular; 2012. [ Google Scholar ] 67. da Costa FV, Guimarães MFM, Messias MCTB. Gender differences in traditional knowledge of useful plants in a Brazilian community. PLoS One. 2021;16(7):e0253820. 10.1371/journal.pone.0253820. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 68. Soldati GT, Almada ED. Political ethnobiology. Ethnobiol Conserv. 2024;13. 10.15451/ec2024-07-13.20-1-1. Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Material 1 (34.6KB, docx) Data Availability Statement The datasets used and/or analysed during the current study are available on the open science platform Zenodo (DOI: 10.5281/zenodo.18166370). The data collected from the interviews were organized in Microsoft ® Excel and analyzed using R software. The codes (custom code) are available from the authors upon reasonable request. Articles from Journal of Ethnobiology and Ethnomedicine are provided here courtesy of BMC ACTIONS View on publisher site PDF (2.9 MB) Cite Collections Permalink PERMALINK Copy RESOURCES Similar articles Cited by other articles Links to NCBI Databases Cite Copy Download .nbib .nbib Format: AMA APA MLA NLM Add to Collections Create a new collection Add to an existing collection Name your collection * Choose a collection Unable to load your collection due to an error Please try again Add Cancel Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top