The well-being of the global human population rests on provisioning services delivered by 12% of the Earth's 400,000 plant species1. Plant utilization by humans is influenced by species traits2,​3,​4, but it is not well understood which traits underpin different human needs5. Here, we focus on palms (Arecaceae), one of the most economically important plant groups globally6, and demonstrate that provisioning services related to basic needs, such as food and medicine, show a strong link to fundamental functional and geographic traits. We integrate data from 2,201 interviews on plant utilization from three biomes in South America—spanning 68 communities, 43 ethnic groups and 2,221 plant uses—with a dataset of 4 traits (leaf length, stem volume, fruit volume, geographic range size) and a species-level phylogeny7. For all 208 palm species occurring in our study area, we test for relations between their traits and perceived value. We find that people preferentially use large, widespread species rather than small, narrow-ranged species, and that different traits are linked to different uses. Further, plant size and geographic range size are stronger predictors of ecosystem service realization for palm services related to basic human needs than less-basic needs (for example, ritual). These findings suggest that reliance on plant size and availability may have prevented our optimal realization of wild-plant services, since ecologically rare yet functionally important (for example, chemically) clades may have been overlooked. Beyond expanding our understanding of how local people use biodiversity in mega-diverse regions, our trait- and phylogeny-based approach helps to understand the processes that underpin ecosystem service realization, a necessary step to meet societal needs in a changing world with a growing human population5,8.

Plant-based ecosystem services are crucial for satisfying human needs5,9. These services, including provisioning services such as food and medicine, can be associated with particular species' traits that capture plant form and function10. For example, humans have selected plant species with traits that maximize crop yield, such as large fruits or height3,4. Understanding which species' traits are linked to particular human needs is important for our efforts to maximize the resilience of ecosystem services in a changing world. Despite previous studies which address this question11, the absence of large-scale approaches—spanning a variety of localities, biomes, cultures and biodiversity uses—has precluded our understanding of how species' traits govern the delivery of ecosystem services and fulfil human needs.

Here, we explore the association of species' traits with human plant utilization by integrating three large datasets on ethnobotany, traits and phylogeny. Our study area encompasses the western Amazon, and the Andes and Chocó biodiversity hotspots, which together also support a large ethnic diversity (Fig. 1). We focus on palms because they are one of the most useful and economically important plant families in the tropics6. The study area is exceptionally rich in palms and ranks second globally in palm diversity12. For the ethnobotanical dataset, we interviewed forest inhabitants about the uses of palms following a standard protocol13 during 18 months of fieldwork, spanning a latitudinal gradient of 22°. This is the largest ethnobotanical documentation effort in the world for palms. We focus on the most important use categories that are key for the livelihoods of the rural population in our study area: Food, Construction, Culture and Medicine, as well as their respective use subcategories14 (Supplementary Table 1). For each use category and subcategory, we calculated species' relative frequency of citation (RFC)15, an index that quantitatively assesses people's consensus about a species' perceived value across our study region (Supplementary Table 2). Additionally, for all species in our study area, we compiled three functional traits that encapsulate the leaf–height–seed (LHS) strategy scheme16 (leaf length, stem volume and fruit volume), as well as a key geographic trait (range size), describing a species' availability.

Figure 1: Distribution of study communities.
Figure 1

Map of northwestern South America showing the location of the 68 study communities and their ethnicity. Amerindian: 1, Bará; 2, Barasana; 3, Baré; 4, Bora; 5, Carijona; 6, Cubeo; 7, Desano; 8, Itano; 9, Macuna; 10, Matapi; 11, Miraña; 12, Tanimuca; 13, Tatuyo; 14, Ticuna; 15, Uitoto; 16, Yahuna; 17, Yucuna; 18, Camsá; 19, Inga; 20, Quillasinga; 21, Emberá; 22, Achuar; 23, Cofán; 24, Quichua; 25, Chachi; 26, Tsa'chila; 27, Aguaruna; 28, Amarakaeri; 29, Cocama; 30, Ese Eja; 31, Shipibo; 32, Urarina; 33, Sapiteri; 34, Chanka; 35, Cavineña; 36, Chácobo; 37, Pacahuara; 38, Tacana; 39, Yaminahua; 40, Yuracaré; 41, Leco. Non-Amerindian: A, Afro-American; M, Mestizo.

We tested the relation between 208 palm species' traits and their RFC value for the different use categories using generalized linear modelling (GLM). We found that different traits correlate with different uses (Supplementary Table 3a). For instance, important species used for Culture have big leaves and fruits, whereas species used for Medicine have big leaves and are widespread. However, because each use category integrates information from various subcategories (Supplementary Table 1), and subcategories often contain different sets of species with different traits (for example, in Construction, not all species in the subcategory Thatch occur in the subcategory Houses), we further explored correlations between traits and the most important subcategories (Supplementary Fig. 1). We found that use subcategories were always related to fewer combinations of traits than use categories (for example, of the three traits significantly associated with the Construction category, only one trait was significantly related to either subcategory). Further, different subcategories within the same use category were often related to different traits (for example, Houses–geographic range versus Thatchfruit volume in Construction), and many subcategories were never related to any trait (for example, Cosmetic and Ritual in Culture) (Supplementary Table 3a). The Thatch–fruit volume relationship is difficult to interpret from an ethnobotanical viewpoint (because thatching requires leaves, not fruits), unless palm fruit volume is significantly inter-related with palm leaf length. Indeed, both traits were significantly inter-related (Supplementary Fig. 3), and we account for this in the analysis below by applying a principal component analysis (PCA). Overall, these results illustrate the need for future trait-based studies to disaggregate ecosystem service categories to their constituent subcategories (where true biological meaning resides), which is not common practice.

The correlations we observed between species' traits and utility could arise from phylogenetic autocorrelation. That is, traits of utilized species might be more similar than expected by chance because closely related species are similar in both utility17,18 and traits19,20 (Fig. 2). To explore this, we used a new phylogeny of palm species7 and calculated the degree of phylogenetic clustering in plant utility and traits with Pagel's lambda (λ)19 for all species in our study region. We found that phylogenetic clustering was strong for species used for Construction (λ = 0.91), but weak for species used for Food, Culture and Medicine (λ = 0.18, λ = 0.09 and λ = 0.00, respectively) (Supplementary Table 3b). Specifically, some palm tribes/subtribes (for example, Phytelepheae, sensu Dransfield et al.21) had a consistently higher RFC than average in all use categories, whereas others (for example, Cocoseae/Bactrinidae sensu Dransfield et al.21) had a RFC below average in all use categories (Supplementary Fig. 2). Additionally, we assessed the degree of phylogenetic clustering in each subcategory, and found that nine of the 14 subcategories (64%) showed strong phylogenetic clustering, including five of the seven Medicine subcategories (Supplementary Table 3b). We also recovered a degree of phylogenetic clustering for the species' traits: phylogenetic clustering was strong for leaf length and fruit volume (λ = 0.89 and λ = 0.89, respectively), moderate for stem volume (λ = 0.46) and weak for geographic range (λ = 6.6 × 10–5). Given the phylogenetic clustering of both species' utility and traits, we performed a phylogenetic generalized least-squares (PGLS) analysis22 to account for any effect of phylogenetic autocorrelation on our analyses. In the majority of cases, the PGLS and GLM analyses yielded qualitatively similar results (Supplementary Table 3) as expected because of the absence of phylogenetic clustering in 50% of the uses in the dataset. However, the PGLS recovered different relations than the GLM between traits and utility in certain uses with strong phylogenetic clustering (for example, Construction, Thatch–Construction, and Infections–Medicine), which underscores the need to account for phylogenetic autocorrelation in ethnobotanical studies.

Figure 2: Phylogenetic distribution of species' traits and of the contribution of 208 palm species to different ecosystem services in northwestern South America.
Figure 2

Species' contribution to different ecosystem services is based on the RFC index (see Methods). All columns were standardized to have the same variance. Numbers 1–10 show palm tribes (subtribes in parentheses).

Research in functional ecology has shown that plant functional traits are often inter-related23, which is also the case between palm stem volume and leaf length (R= 0.36, P = 6.12 × 10–8) and between palm leaf length and fruit volume (R= 0.42, P = 2.36 × 10–10) (Supplementary Fig. 3). Because trait multicollinearity can obscure trait–use relationships, we performed a PCA on the three functional traits (leaf length, fruit volume, stem volume), selecting the first axis (PC1) to represent them jointly (accounting for 54% of their variability). As all three traits load positively on PC1, it represents overall plant size. We did not include geographic range size in the PCA because it represents a fundamentally different species characteristic and affects a different aspect of species' perceived utility, namely availability. In general, large and widely distributed palm species are the most conspicuous, more people and cultures are exposed to them, and they show the highest probability of human–plant interactions and cross-cultural knowledge transmission of perceived utility24. We explored the association of PC1 (‘plant size’) and geographic range with different uses by means of GLM and PGLS analyses. Overall, we found that plant size was a stronger predictor than geographic range of a species' importance for Food and Culture (and their respective subcategories), whereas geographic range was a stronger predictor for Construction and Medicine (Supplementary Table 4). Some uses were not significantly related to either plant size and/or geographic range (for example, Ritual–Culture, Cosmetic–Culture, Respiratory–Medicine).

What do these findings reveal about the processes that underpin ecosystem service realization? Psychologists have ranked human needs—based on importance for survival—from basic human needs (for example, food and health) to less-basic needs (for example, self-esteem and self-actualization)25 (Fig. 3a). Typically, one would expect basic needs to show a strong link to species' traits because species that satisfy basic physiological needs must possess particular traits and be cost-effective to gather. In contrast, species used for non-basic needs could possess diverse traits because of idiosyncratic cultural preferences. We investigated how different plant uses position along the plant size and geographic range axes (PGLS coefficients), and how this relates to the hierarchy of human needs (Fig. 3b). Our findings show that different combinations of species' traits drive which benefits people obtain from biodiversity. We hypothesize that the strong coupling of plant size and Food is consistent with a need for constantly high yields in subsistence livelihoods that rely on local resources for food security. The weaker link between Medicine and plant size and a strong link to geographic range is consistent with availability as key for the selection of medicinal resources. This suggests that the traits selected for do not really reflect a specific need for efficacy because general and specific medical conditions correlate with availability. Efficacy as medicine should be the underlying trait, but the pharmacological tests needed to verify this are largely inexistent. In principle, a high Pagel's λ would be expected for working medicines because the concentration (and existence of) any chemical compound is expected to be phylogenetically conserved. That certain palm tribes (Manicarieae, Chamaedoreeae and Euterpeae) had very high RFC scores for Medicine (Supplementary Fig. 2) and that five of seven medical subcategories were strongly phylogenetically clustered is a noteworthy lead for future ethnopharmacological work. Construction's intermediate position along the plant size and geographic range axes highlights that both quantity (for example, leaf or stem biomass) and especially availability influence people's selection of species as housing material. At the other extreme, higher-level human needs such as Ritual–Culture uses do not necessarily reflect physical or chemical properties of plants, and hence need not be strongly linked to plant traits. Ritual–Culture uses in South America include far more uses than psychotropic ones (for example, uses related to myth–religious aspects, festivals and feasts, construction of coffins, sorcery), and we propose that this is why Ritual–Culture uses showed no clear relation to phylogeny, and potentially underlying plant chemistry. This would be consistent with numerous cases in Africa and Asia where idiosyncratic cultural preferences for various traits, and not plant chemistry, strongly determine human selection of plants for Ritual–Culture26.

Figure 3: The hierarchy of human needs, species' traits, and ecosystem services.
Figure 3

a, Human needs can be organized hierarchically (sensu Maslow25), from basic needs at the bottom of the pyramid to more existential or self-fulfilment needs at the top of the pyramid. b, Different combinations of plant size and geographic range (PGLS coefficients) underpin the benefits that people derive from nature and the hierarchy of human needs. Plant size is represented as axis one of a principal component analysis on the three functional traits—leaf length, fruit volume, and stem volume. Uses related to basic human needs (for example, Food) show a significant positive link with both plant size and geographic range (yellow symbols); uses related to psychological needs (for example, Culture) only show a significant link to one of the two traits (blue or red symbols); and uses related to self-fulfilment needs (for example, Ritual, Cosmetic) never show significant links to traits (white symbols).

By integrating local stakeholders' views, ecology and phylogeny, we provide a novel interdisciplinary perspective on the linkages between ecosystem services, human needs and species' traits. Typically, these relationships are explored at local scales27, but taking advantage of large-scale, cross-cultural ethnobotanical datasets can shed new light on these linkages. Our study highlights the importance of ethnobotanical research documenting how cultural heritage is linked to biodiversity resources, particularly in areas where this is understudied, such as South America28. We demonstrate that this research can help us better understand the processes that underpin ecosystem service realization, which will be crucial to meet current and future societal needs under strong human population growth and environmental changes. Our findings on New World palms may not reflect global trends across other societies and plant families, and further trait-based research on the full spectrum of human-associated plants is therefore urgently needed. Given our reliance on a limited number of food crops29, the use of similar interdisciplinary approaches to predict which lineages of the tree of life are likely to yield novel food crops and medicines could be highly rewarding. Still, the traditional importance of plant size and availability may also have prevented our complete understanding of wild-plant services. This may be so because ecologically rare yet functionally (for example, chemically) important clades may have been overlooked, even by local communities with great knowledge about biodiversity. Previous studies have shown that imminent global change may cause a shift in community functional traits30 and our findings indicate that such shifts are bound to directly affect provisioning services—particularly covering food—that is directly linked to species' traits. Hence, efforts looking to enhance the resilience of ecosystem services for human well-being8 should consider the linkages between species' traits, evolutionary history and human needs.


Species' traits

For all species in our study area we obtained five morphological measurements: maximum stem height, maximum stem diameter, maximum leaf length, average fruit length and average fruit diameter from Henderson31, and complemented it with data from recent taxonomic revisions32,​33,​34,​35,​36,​37,​38,​39. For the few species lacking some measurements (stem height, N = 8 species; maximum stem diameter, N = 4; maximum leaf length, N = 2; maximum rachis length, N = 1; maximum petiole length, N = 4; average fruit length, N = 7; average fruit diameter, N = 6), we assigned the average value of all congeneric species. For our analyses, we selected three functional traits following Westoby's LHS plant ecology strategy scheme16: maximum leaf length, stem volume (Vstem = π × r2 × h; r, stem radius, derived from maximum stem diameter; h, maximum stem height) and fruit volume (Vfruit = 4/3×π×a×b×c; a, average fruit length/2; b, fruit height/2; c, average fruit width/2). Leaf length and fruit volume reflect specific leaf area and seed size in Westoby's scheme, respectively. Further, seed size and fruit volume are closely related in palms (many palm genera are one-seeded)40. Together, leaf length and fruit volume represent the best available data for New World palms40. We used stem volume instead of maximum stem height to better correct for the size of the relatively small overall size of climbing palms, which are tall but thin.

Geographic range

We derived a list of all palm species growing in our study area from Balslev et al.41 and followed The Plant List42 to unify nomenclature. This resulted in a total regional pool of 291 native palm species, 208 of which were present in our study communities, and which we subsequently analysed. For these 208 species, we compiled 1° grid square resolution distribution maps from Bjorholm et al.43 Because the distribution maps of Bjorholm et al. were based on taxonomic knowledge from 199544, we made additional distribution maps for newly described species and for genera with more recent taxonomic treatments (for example, Aiphanes45, Astrocaryum46,47, Attalea (J.-C. Pintaud, personal communication), Bactris48, Calyptrogyne33, Ceroxylon39, Desmoncus35, Geonoma34, Hyospathe32, Pholidostachys36 and Welfia37). We first mapped known collection localities into a grid with unit cells of 1° in latitude and longitude, and subsequently added new grid cells where we expect species to occur based on our field experience in the region, or that of other experts (Jean-Christophe Pintaud, Attalea). Thus, these new distribution maps are expert geographic range maps, as those that Bjorholm et al.43 derived from Henderson et al.44. For species with updated geographic ranges in Colombia, we geo-referenced the expected geographic range maps of Galeano & Bernal38. In combination, these distribution maps represent the most up-to-date geographic ranges available for all species in our study area. All distribution maps were made using ArcGIS 10.1 (ESRI Inc.). From these maps, we quantified the geographic range size of each of the palm species in our study area by counting the number of 1° grid squares in which it is present. We refer to this number as ‘geographic range’ in our study.


Approval for this study was granted by the Committee for Ethical Research of the Autonomous University of Madrid (no. 48-922; Principal Investigator M.J. Macía). Before initiating data collection in situ, we obtained oral informed consent at the community level and then from the individual before each interview. Informants were made aware of their right to discontinue the interviews at any time and that all of the information provided would be anonymized. We collected ethnobotanical information about palms making 2,201 interviews over 18 months of fieldwork (May 2010 to December 2011) using a standardized method13,49, developed to address the void in large-scale ethnobotanical efforts and comparative analyses14 (Fig. 1). We collected ethnobotanical data with two types of participants: expert informants and general informants. Experts were selected through consensus during community meetings. In communities too large for gathering all villagers, we asked general informants to recommend their most knowledgeable peers. Walks in the field with each expert were performed to identify all palm species growing in the surrounding forest, register their vernacular names and document their uses. Once experts were interviewed, we used the list of compiled vernacular names as the basis for interviews with general informants. General informants were selected in each community in a stratified manner to have a representative sample of age classes (18–30 years, 28%; 31–40 years, 23%; 41–50 years, 20%; 51–60 years, 13%; >60 years, 16%) and gender (women, n = 1107; men, n = 1094). Interviews were generally conducted in Spanish or with a local interpreter when needed. Palm species were identified in the field using Galeano & Bernal38, Borchsenius et al.50 and Moraes51, and specimens collected when our field identification needed confirmation. Palm-collecting permits were obtained through the following authorities: Instituto Amazónico de Investigaciones Cieníficas Sinchi (Colombia); the Ministry of Environment (Ecuador); the Instituto Nacional de Recursos Naturales (Peru); and the Dirección General de Biodiversidad y Áreas Protegidas (Bolivia). Field studies did not involve endangered or protected species. Voucher specimens (N = 203) are deposited in the herbaria of AAU, AMAZ, COL, LPB, QCA and USM, acronyms according to Thiers52. We classified use reports from interviews into one of ten use categories following the Economic Botany Data Collection Standard53, with recently proposed modifications14. Our analyses focus on the most important palm-based ecosystem services in our study area, which correspond to the use categories Food (‘Human food’ in ref. 14), Construction, Culture (‘Cultural’ in ref. 14) and Medicine (‘Medicinal and veterinary’ in ref. 14) and their respective subcategories14,54 (for a description refer to Supplementary Table 1). In addition, we include a broad measure of usefulness in our analyses, which we here term ‘Any service’. This measure contains information on whether a species has use reports associated to the four important palm-based ecosystem services or to any of the other less-important14,54 services (Animal food, Environmental, Fuel, Toxic, Utensils and tools, and Other uses) (Supplementary Table 1). For each of these classifications of palm-based ecosystem services (four use categories, 14 subcategories and Any service), we calculated a modified version of the RFC, an index that has been recommended for comparative studies that assess the perceived value of species15. To obtain the RFC we first counted the number of informants that mentioned a given species in a use category (or subcategory, or any service) and then counted the number of informants across the geographic range of each species, by counting the number of interviews performed per 1° grid cell in our study region. The RFC is thus the number of informants that cite a given species for a particular use category or subcategory divided by the total number of informants interviewed across the species' geographic range. Given our large-scale analyses, our approach to compute the RFC limits the number of informants for computing the RFC to those interviewed in the grid cells where the particular species occurs. This was done to ensure that widespread species, to which more informants are exposed, did not necessarily receive higher RFC values.

Data analysis

For each use classification we performed separate GLM analyses, selecting the RFC as the response variable and species' geographic range and traits as the explanatory variables. Before analyses, we normalized all species' morphological traits and geographic range to a mean of 0 and a standard deviation of 1. The response variable was arcsin transformed and normalized to a mean of 0 and a standard deviation of 1. Plant morphological traits are often significantly correlated23, and this was also the case in our trait dataset (Supplementary Fig. 2). To ease interpretation and avoid potential multicollinearity problems in our analyses, we performed a PCA on the three morphological traits (leaf length, fruit volume and stem volume). We selected PC1 (accounting for 54% of the variability in the data) and ran additional GLM analyses with the RFC values as response variables and PC1 and geographic range as explanatory variables (Supplementary Table 3).

We also reanalysed all data incorporating phylogenetic autocorrelation into our analysis. All phylogenetic analyses were based on the 1,000 constraint trees from the species level phylogeny of palms of Faurby et al.7. The trees were trimmed to only include species present in the study region. Before analyses, we normalized all species' morphological traits and geographic range to a mean of 0 and a standard deviation of 1. The response variable was arcsin transformed and normalized to a mean of 0 and a standard deviation of 1. To calculate the degree of phylogenetic clustering of uses and traits, we used Pagel's lambda19 (hereafter λ). Values of λ close to 0 are found for data without phylogenetic dependence, whereas values close to 1 indicate strong phylogenetic clustering. We assessed correlations between species' traits and uses while accounting for phylogenetic autocorrelation using a PGLS22 analysis. The results from 1,000 trees were then combined as the weighted mean (by Akaike information criterion weight) of the point estimates, standard errors and P-values from the analysis of each tree. Additional PGLS analyses were made to assess the correlation between PC1 and geographic range values with uses. All analyses were performed in R55 using commands from the libraries phylolm56 phytools57, ape58 and quantitative PCR (qPCR)59.

Data availability

The data that support the findings of this study are available from the corresponding author upon request.

Additional information

How to cite this article: Cámara-Leret, R. et al. Fundamental species traits explain provisioning services of tropical American palms. Nat. Plants 3, 16220 (2017).


  1. 1.

    The State of the World's Plants Report (Royal Botanic Gardens, Kew, 2016).

  2. 2.

    et al. Functional traits, the phylogeny of function, and ecosystem service vulnerability. Ecol. Evol. 3, 2958–2975 (2013).

  3. 3.

    Evolution, consequences and future of plant and animal domestication. Nature 418, 700–707 (2002).

  4. 4.

    & The nature of selection during plant domestication. Nature 457, 843–848 (2009).

  5. 5.

    et al. Linking functional diversity and social actor strategies in a framework for interdisciplinary analysis of nature's benefits to society. Proc. Natl Acad. Sci. USA 108, 895–902 (2011).

  6. 6.

    Tropical Palms (Food and Agriculture Organization of the United Nations, 2011).

  7. 7.

    , , & An all-evidence species-level supertree for the palms (Arecaceae). Mol. Phylogenet. Evol. 100, 57–69 (2016).

  8. 8.

    , , & Domesticated nature: shaping landscapes and ecosystems for human welfare. Science 316, 1866–1869 (2007).

  9. 9.

    , , & Paleoindian cave dwellers in the Amazon: the peopling of the Americas. Science 272, 373 (1996).

  10. 10.

    , , & Biodiversity loss threatens human well-being. PLoS Biol. 4, e277 (2006).

  11. 11.

    et al. Towards an assessment of multiple ecosystem processes and services via functional traits. Biodivers. Conserv. 19, 2873–2893 (2010).

  12. 12.

    et al. Quaternary and pre-Quaternary historical legacies in the global distribution of a major tropical plant lineage. Glob. Ecol. Biogeog. 21, 909–921 (2012).

  13. 13.

    , & in Proceedings of the Botany 2011 Symposium Honoring Richard E. Schultes (eds Ponman, B. & Bussmann, R. W.) 41–71 (Missouri Botanical Garden, 2012).

  14. 14.

    et al. Palm uses in northwestern South America: a quantitative review. Bot. Rev. 77, 462–570 (2011).

  15. 15.

    & Cultural importance indices: a comparative analysis based on the useful wild plants of southern Cantabria (northern Spain). Econ. Bot. 62, 24–39 (2008).

  16. 16.

    A leaf–height–seed (LHS) plant ecology strategy scheme. Plant Soil 199, 213–227 (1998).

  17. 17.

    et al. Preserving the evolutionary potential of floras in biodiversity hotspots. Nature 445, 757–760 (2007).

  18. 18.

    et al. Phylogenies reveal predictive power of traditional medicine in bioprospecting. Proc. Natl Acad. Sci. USA 109, 15835–15840 (2012).

  19. 19.

    Inferring the historical patterns of biological evolution. Nature 401, 877–884 (1999).

  20. 20.

    & Comparative methods for explaining adaptations. Nature 351, 619–624 (1991).

  21. 21.

    et al. Genera Palmarum: The Evolution and Classification of Palms (Kew Publishing, 2008).

  22. 22.

    & Phylogenies and the comparative method: a general approach to incorporating phylogenetic information into the analysis of interspecific data. Am. Nat. 149, 646–667 (1997).

  23. 23.

    et al. The global spectrum of plant form and function. Nature 529, 167–171 (2016).

  24. 24.

    , , , & Geospatial patterns in traditional knowledge serve in assessing intellectual property rights and benefit-sharing in northwest South America. J. Ethnopharmacol. 158, 58–65 (2014).

  25. 25.

    A theory of human motivation. Psychol. Rev. 50, 370 (1943).

  26. 26.

    , , & New categories for traditional medicine in the economic botany data collection standard. J. Ethnopharmacol. 155, 1388–1392 (2014).

  27. 27.

    & Biodiversity governance: a tower of Babel of scales and cultures. PLoS Biol. 13, e1002108 (2015).

  28. 28.

    , , & Ethnobotanical knowledge is vastly under-documented in northwestern South America. PLoS ONE 9, e85794 (2014).

  29. 29.

    , & Patterns and processes in crop domestication: an historical review and quantitative analysis of 203 global food crops. New Phytol. 196, 29–48 (2012).

  30. 30.

    et al. Plant phenotypic plasticity in a changing climate. Trends Plant Sci. 15, 684–692 (2010).

  31. 31.

    Evolution and Ecology of Palms (The New York Botanical Garden Press, 2002).

  32. 32.

    A multivariate analysis of Hyospathe (Palmae). Am. J. Bot. 91, 953–965 (2004).

  33. 33.

    A multivariate study of Calyptrogyne (Palmae). Syst. Bot. 30, 60–83 (2005).

  34. 34.

    A revision of Geonoma (Arecaceae). Phytotaxa 17, 1–271 (2011).

  35. 35.

    A revision of Desmoncus (Arecaceae). Phytotaxa 35, 1–88 (2011).

  36. 36.

    A revision of Pholidostachys (Arecaceae). Phytotaxa 43, 1–48 (2012).

  37. 37.

    & A revision of Welfia (Arecaceae). Phytotaxa 119, 33–44 (2013).

  38. 38.

    & Palmas de Colombia: Guía de Campo (Editorial Universidad Nacional de Colombia, 2010).

  39. 39.

    & A revision of the Andean wax palms, Ceroxylon (Arecaceae). Phytotaxa 34, 1–64 (2011).

  40. 40.

    , & Geographical variation and environmental correlates of functional trait distributions in palms (Arecaceae) across the New World. Bot. J. Linn. Soc. 179, 602–617 (2015).

  41. 41.

    , & Cosecha de Palmas en el Noroeste de Suramérica: Bases Científicas Para Su Manejo y Conservación (Pontificia Universidad Católica del Ecuador, 2015).

  42. 42.

    The Plant List (Royal Botanical Garden, Kew & Missouri Botanical Garden);

  43. 43.

    , , & Environmental and spatial controls of palm (Arecaceae) species richness across the Americas. Glob. Ecol. Biogeogr. 14, 423–429 (2005).

  44. 44.

    , & Field Guide to the Palms of the Americas (Princeton Univ. Press, 1995).

  45. 45.

    & Aiphanes (Palmae) (New York Botanical Garden Press on behalf of Organization for Flora Neotropica, 1996).

  46. 46.

    The genus Astrocaryum (Arecaceae). Rev. Peru. Biol. 15, 31–48 (2008).

  47. 47.

    , , & Detailed assessment of the distribution of Astrocaryum sect. Huicungo (Arecaceae) in Peru. Rev. Peru. Biol. 18, 279–282 (2011).

  48. 48.

    Bactris (Palmae) (New York Botanical Garden Press on behalf of Organization for Flora Neotropica, 2000).

  49. 49.

    , & Toma de datos etnobotánicos de palmeras y variables socioeconómicas en comunidades rurales. Ecol. Bolivia 45, 44–68 (2010).

  50. 50.

    , & Manual to the Palms of Ecuador (Aarhus Univ. Press, 1998).

  51. 51.

    Flora de palmeras de Bolivia (Plural Editores, 2004).

  52. 52.

    Index Herbariorum: A Global Directory of Public Herbaria and Associated Staff. New York Botanical Garden's Virtual Herbarium (New York Botanical Garden, 2016);

  53. 53.

    Economic Botany Data Collection Standard. Prepared for the International Working Group on Taxonomic Databases for Plant Sciences (TDWG) (Royal Botanic Gardens Kew, 1995).

  54. 54.

    et al. Ecological community traits and traditional knowledge shape palm ecosystem services in northwestern South America. Forest Ecol. Manag. 334, 28–42 (2014).

  55. 55.

    R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2015).

  56. 56.

    & phylolm: Phylogenetic Linear Regression. R package v.2.5 (2015);

  57. 57.

    Phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 3, 217–223 (2012).

  58. 58.

    , & APE: analyses of phylogenetics and evolution in R language. Bioinformatics 20, 289–290 (2004).

  59. 59.

    & qpcR: Modelling and analysis of real-time PCR data (R Foundation for Statistical Computing, 2014).

Download references


We thank the 68 study communities and participants of our fieldwork interviews. We are also grateful to our partner institutions and their research teams who dedicated resources to facilitate our research. We thank G. Galeano, R. Bernal, J.-C. Pintaud, R. Valencia, H. Navarrete, L. de La Torre, M. Moraes, B. Millán and R. Carrillo for their support and discussions. We extend our gratitude to J.C. Copete, M. Soto Gomez, N. Paniagua, L. Camelo, R. Bussmann and M. Jaimes for assistance in fieldwork, to D. Warren for helpful discussions, to S. Cámara-Leret for contributing palm illustrations, and to I. Cámara-Leret for assistance with design and layout of figures. This study was funded by the European Union, 7th Framework Programme (FP7-PALMS-Contract no. 212631, to H.B.), and also supported by the Russell E. Train Education for Nature Program of the WWF, Anne S. Chatham Fellowship of the Garden Club of America, William L. Brown Center, Universidad Autónoma de Madrid travel grants programme, and a GSST fellowship of Aarhus University. J.-C.S. and B.G. were supported by the European Research Council (ERC-2012-StG-310886-HISTFUNC); C.H.S.-L. and N.R. by the People Programme (Marie Curie Actions) of the European Union's 7th Framework programme (FP7-PEOPLE-2012-IEF-328637 - BiodiversityAltitude); W.D.K. by the Netherlands Organization for Scientific Research (824.15.007) and the University of Amsterdam (starting grant); and H.B. by the Danish National Science Research Council (272-06-0476).

Author information

Author notes

    • Rodrigo Cámara-Leret

    Present address: Royal Botanic Gardens, Kew, Richmond TW9 3AB, UK.


  1. Section for Ecoinformatics and Biodiversity, Department of Bioscience, Aarhus University, Ny Munkegade 114, 8000 Aarhus C, Denmark

    • Rodrigo Cámara-Leret
    • , Henrik Balslev
    • , Bastian Göldel
    •  & Jens-Christian Svenning
  2. Department of Biogeography and Global Change, Museo Nacional de Ciencias Naturales, CSIC, Calle José Gutiérrez Abascal 2, 28006 Madrid, Spain

    • Søren Faurby
  3. Department of Biological and Environmental Sciences, University of Gothenburg, Box 461, SE 405 30 Göteborg, Sweden

    • Søren Faurby
  4. Departamento de Biología, Área de Botánica, Universidad Autónoma de Madrid, Calle Darwin 2, 28049 Madrid, Spain

    • Manuel J. Macía
  5. Institute for Biodiversity and Ecosystem Dynamics (IBED), University of Amsterdam, PO Box 94248, 1090 GE Amsterdam, The Netherlands

    • W. Daniel Kissling
  6. Evolutionary Genomics Section, Natural History Museum of Denmark, Faculty of Science, University of Copenhagen, Sølvgade 83S, 1307 Copenhagen K, Denmark

    • Nina Rønsted
    •  & C. Haris Saslis-Lagoudakis


  1. Search for Rodrigo Cámara-Leret in:

  2. Search for Søren Faurby in:

  3. Search for Manuel J. Macía in:

  4. Search for Henrik Balslev in:

  5. Search for Bastian Göldel in:

  6. Search for Jens-Christian Svenning in:

  7. Search for W. Daniel Kissling in:

  8. Search for Nina Rønsted in:

  9. Search for C. Haris Saslis-Lagoudakis in:


R.C.-L. and C.H.S.-L. conceived and designed the study. R.C.-L. and M.J.M. did the ethnobotanical fieldwork. R.C.-L., M.J.M., H.B., S.F., B.G. and W.D.K. provided data. R.C.-L., B.G. and H.B. built new geographic range maps. R.C.-L., C.H.S.-L. and S.F. analysed the data. R.C.-L. and C.H.S.-L. wrote the paper. All authors discussed the results and commented on the manuscript.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to Rodrigo Cámara-Leret.

Supplementary information

PDF files

  1. 1.

    Supplementary Information

    Supplementary Figures 1–3, Supplementary Tables 1–4.

About this article

Publication history






Rights and permissions

To obtain permission to re-use content from this article visit RightsLink.