Many large species have declined worldwide due to habitat fragmentation and poaching. The defaunation of large frugivores and the consequent reductions of seed dispersal services may have immediate effects on plant demography. Yet, the lasting effects of frugivore defaunation on microevolutionary processes of the plants they disperse remain understudied. We tested if the loss of large seed dispersers can lead to microevolutionary changes of a tropical palm. We show that frugivore defaunation is the main driver of changes in allelic frequency among populations. Turnover of alleles accounted for 100% of dissimilarity in allelic frequencies of individuals between defaunated and non-defaunated forests; and individuals from defaunated sites are 1.5 times more similar genetically than those found in pristine sites. Given that sizeable fractions of the palm fruit crops remain undispersed in defaunated sites due to lack of large-bodied frugivores, this distinct pattern of gene pool composition of early recruits may reveal strong dispersal limitation for specific genotypes, or collapses of gene flow between fragmented areas, or both. Because most of tropical tree species rely on seed dispersal by vertebrates, our results show that defaunation has a lasting effect on microevolutionary processes, with potential consequences for persistence under scenarios of environmental change.
Many species and populations of animals are facing a dramatic decline worldwide, a phenomenon named “Anthropocene defaunation”1. The decline of animals may have serious negative ecological and evolutionary consequences, especially if these animals play important roles in mutualistic processes, such as pollination and seed dispersal2,3,4. Vertebrate defaunation frequently targets large-bodied species1 and several studies have found that the extinction of seed dispersers may have immediate effects on plant populations, such as significant reductions of fruit removal success5, collapse of seed dispersal distance6,7 and recruitment8. Size-biased defaunation of seed dispersers may also impose marked changes in the selective pressures on plants because smaller species cannot provide the same dispersal services, entailing rapid evolutionary changes in seed size9. A lasting consequence of these contemporary, human-driven, effects would be changes in the genetic pool caused by the loss of connectivity among fragmented populations due to the reduction of long-distance dispersal (LDD) events6,10,11 and by filtering-out specific genotypes from the gene pool due to the selective foraging behavior of dispersal agents12.
Given the negative effects of defaunation on plant populations, we tested the hypothesis that the functional loss of large seed dispersers may lead to microevolutionary changes of the plants that rely on their dispersal. In addition to anthropogenic impacts, the distribution of genetic variability among plant populations over large spatial scales may be also due to a combination of historic and ongoing landscape effects that ultimately influence the joint action of gene flow, selection and genetic drift13,14,15. Therefore, we explored alternative hypotheses (Table 1) defined from specific landscape attributes and used replicated areas in a diverse array of landscape conditions to test the relative importance of alternative drivers of genetic variability distribution16.
Here, we analyzed recruits (seedlings) of the palm Euterpe edulis in 19 populations in the Brazilian Atlantic forest. This palm is a dominant species distributed across the Atlantic forest of South America17. Euterpe edulis produces round fleshy fruits ranging from 8.3 to 14.1 mm in diameter18 and they are dispersed by large (e.g., cotingas Procnias nudicollis, toucans Ramphastos spp. and guans Penelope spp. and Aburria jacutinga) and medium-sized avian frugivores (Turdus spp.)9. We found previously that the functional loss of the large seed dispersers due to defaunation resulted in phenotypic differentiation in seed traits among E. edulis populations and has been a driver for a rapid evolutionary reduction of seed size in defaunated palm populations9.
We tested four different hypothetical scenarios to assess variation in local patterns of genetic variability, i.e., to determine which scenario best fits the observed pattern of genetic variation. We thus contrasted genetic variability patterns across replicated local palm populations in different situations according with these scenarios.
Our working hypothesis is framed on the effects of loss of large-bodied frugivores and the lasting effects of the associated loss of long-distance seed dispersal and reduced recruitment success in defaunated areas6,9,10. Several studies have pointed out the role of seed dispersers in determining the spatial genetic structure of animal-dispersed plants7,13,19,20,21,22. These empirical studies provide convincing evidence of negative impacts of defaunation on tree populations, showing that the foraging behavior and movement pattern of the seed dispersers may affect seed dispersal distance, and consequently, have a lasting signal on the spatial genetic structure of plants at local scales. Large frugivores consume many fruits and a wide range of seed sizes, while smaller birds eat fewer fruits and smaller seeds (<12 mm wide)9. Consistent directional selection against large seed size in defaunated areas9 may also drive changes in allelic frequencies, with the gene pools of the populations that lost the large seed dispersers being more similar when contrasted with “control” stands in pristine areas, where the full assemblage of mutualistic avian frugivore species is present. Thus, the loss of large seed dispersers may drive microevolutionary changes among E. edulis stands due to a selection against large seed size and to the collapse of the long-distance seed dispersal events, impairing effective gene flow (see e.g., Karubian et al.22, documenting small-scale genetic signatures of variable foraging patterns by frugivores). Simulation models demonstrate that it is possible to find effects of divergent selection on neutral markers, mainly with the combined action of strong selection and very low migration rate23. We expect this selection to be strong in defaunated landscapes where the local extinction of large frugivores leads to dramatic decrease in fruit removal rates and seed dispersal effectiveness10,24.
Neotropical forests have been severely affected by habitat loss and fragmentation, resulting in biologically impoverished patches25,26. Forest fragmentation may lead to drastic reductions in population size and may increase the spatial isolation of populations27. Population isolation may raise inbreeding levels by increasing the probability of mating between closely related individuals and self-pollination due to, for example, changes in the composition and behavior of pollinators28. Isolation may also limit the dispersal among populations, reducing gene flow and population connectivity28. Therefore, we predicted that forest fragmentation, measured by percentage of forest cover, would lead to microevolutionary changes among E. edulis populations.
The Atlantic forest of South America is heterogeneous in climate, relief and vegetation type29 and can be divided into several biogeographical regions30. Biogeographical region may also influence the distribution of genetic variability through evolutionary time (phylogeographic effects) and impose distinct selective pressures leading to microevolutionary changes among populations with distinct biogeographical origins. For example, at the community level, plant species composition differs between biogeographical regions composed by rain forest and by semideciduous forest, and plant composition may further include species that are able to cope with a longer dry season31. Thus, individuals from the same biogeographical region (here rain forest or semideciduous forest) may share similar gene pool features due to dispersal limitation combined to local adaptation, and we would thus expect microevolutionary changes among individuals from rain forest to semideciduous forest.
We sampled the E. edulis individuals in delimited and isolated populations with adequate replication across sample groups (strata) defined by the previous hypotheses. We also tested the effects of sampling design as a potential influence for the observed genetic variability distribution, i.e., a situation where differences would emerge by chance effects related to the specific populations sampled.
We found that the hypotheses based on defaunation and biogeographic regions lead to significant microevolutionary changes among populations of Euterpe edulis (Table 1, Fig. 1). Both hypotheses had the greatest discrimination power and correct assignment of individual palm genotypes to the a priori groupings defined by the hypotheses (represented by kappa index), followed by forest fragmentation and sampling design (Table 1 and Fig. 1). As expected, adscription to biogeographic areas had a significant effect on the differentiation of the genetic pools (kappa estimator of correct assignment, 0.92 [0.88–0.96]). However, the kappa value for the defaunation hypothesis was very similar (0.92 [0.88–0.95]) despite the fact of including populations from either biogeographic origin in each of the groups being compared. The cluster comprising individuals from defaunated populations showed more variability in discriminant function scores when compared to the cluster including individuals from non-defaunated sites and to the clusters related to the hypothesis of biogeographical regions effects (Supplementary Information Figs S1 and S2).
We assessed how defaunation and biogeographical regions have influenced genetic variability and contemporary effective population sizes across the sites. We found that microevolutionary changes between defaunated and non-defaunated sites and between rain forest and semideciduous forest were not due to changes in genetic variability or contemporary effective population size (see Table S4 for genetic variability of each site). Models containing defaunation and biogeographical regions as explanatory variables did not predict variation in genetic diversity (Defaunation: df = 1, F = 0.02, p-value = 0.90; Biogeographical regions: df = 1, F = 1.00, p-value = 0.33), allelic richness (Defaunation: df = 1, F = 2.29, p-value = 0.15; Biogeographical regions: df = 1, F = 0.05, p-value = 0.82), inbreeding coefficient (Defaunation: df = 1, F = 1.14, p-value = 0.30; Biogeographical regions: df = 1, F = 0.57, p-value = 0.46) and contemporary effective population size (Defaunation: df = 1, F = 1.60, p-value = 0.22; Biogeographical regions: df = 1, F = 0.48, p-value = 0.50).
We analyzed whether differences in allelic richness (β-allelic diversity) among defaunated and non defaunated sites or among sites from distinct biogeographic regions is due to alleles replacement (turnover, βsim) or richness differences driven to passive allelic loss (nestedness-resultant dissimilarity, βsne)32. The turnover component of allelic richness measures the replacement of alleles in one site by different ones in other site. On the other hand, the nestedness component measures allelic gain or loss; for instance, the alleles of one site with low number of alleles are a subset of alleles of other site containing a high number of alleles32. A nested pattern may thus represent two main situations: the alleles missing in populations with lower allelic richness were lost by drift, or the population with higher allelic richness presents ancient alleles that were spread in all populations and new mutations which have not yet widespread. We found that 100% of dissimilarity in allelic richness among palm recruits in defaunated and non defaunated sites (Mean βdiversity = 0.64, 95% CI = 0.63 to 0.64; Mean βsim = 0.64, 95% CI = 0.63 to 0.64; Mean βsne = 0.00, 95% CI = 0.00 to 0.00, Supplementary Information Table S5) and among rain forest and semideciduous forest (Mean βdiversity = 0.61, 95% CI = 0.60 to 0.62; Mean βsim = 0.61, 95% CI = 0.60 to 0.62; Mean βsne = 0.00, 95% CI = 0.00 to 0.00, Supplementary Information Table S5) is due to allele turnover. Moreover, we found that palm recruits from defaunated sites are genetically more similar than those from non-defaunated sites; and recruits from rain forests sites are genetically more similar than those from semideciduous forests. The β-allelic diversity and its components are lower among palm recruits in defaunated sites (Mean βsim = 0.49, SD = 0.10, Mean βsne = 0.00, SD = 0.00, Table 1) compared to non defaunated sites (Mean βsim = 0.64, SD = 0.09, Mean βsne = 0.00, SD = 0.00, Table 1) and lower among individuals in rain forest (Mean βsim = 0.53, SD = 0.11, Mean βsne = 0.00, SD = 0.00, Table 1) compared to semideciduous forest (Mean βsim = 0.58, SD = 0.10, Mean βsne = 0.00, SD = 0.00, Table 1).
Finally, we estimated the contribution of spatial effects on genetic differentiation of E. edulis. We performed Redundancy Analysis (RDA) using the forward selection method and found that the model including the effects of defaunation and biogeographical regions, but not spatial effects, was the best to explain variation in allele frequencies among sites (R2adj = 0.12, p-value = 0.01).
Our results indicate that the defaunation of large seed dispersers has a distinct signal on large-scale genetic variability, potentially arising from microevolutionary changes in the palm populations. This supports and extends previous findings9 providing evidence for a deeper far-reaching consequence than previously thought. There has been multiple recent evidence demonstrating that the extinction of seed dispersers leads to a decline in recruitment success8,33 and an increase in plant extinction risk34. In addition, previous research has associated seed dispersal limitation with genetic consequences consistent with the loss of gene flow among populations11. By assessing realized dispersal (early seedlings established) our study demonstrates for the first time that defaunation may lead to changes in the frequency of alleles among populations, potentially driving long-term homogenization of genetic pools within plant populations from defaunated sites.
We found that defaunation leads to genetic changes in Euterpe edulis populations. This could be a synergistic effect from a higher frequency of mating among closely-related individuals, a reduction of effective population size, limited long distance dispersal among populations11,15,35, and rapid evolutionary changes in the seed size of E. edulis populations in defaunated areas9. Our results suggest that the observed genetic changes were not due to a higher frequency of mating among closely-related individuals or reduction of population size. Instead, the results suggest that genetic changes among E. edulis populations were consistently associated with scenarios of strong seed dispersal limitation9 and/or collapse of gene flow among fragmented areas11. We found that the allelic richness differences among defaunated and non-defaunated sites are largely attributable to allele replacement. Moreover, we found that palm recruits from defaunated sites are genetically more similar than recruits from non-defaunated sites. The altered seed rain composition due to reduction of fruit removal success in defaunated sites9 might be determined by distinct combinations of recruited genotypes in these disturbed populations, creating a lasting signal in the composition of the gene pool12,36. This suggests that strong selection pressure imposed by defaunation could result in changes of allelic frequencies among defaunated and non defaunated sites. Moreover, our results suggest that a similar selective pressure (lack of large-bodied seed dispersers) has led to a homogenization in terms of allelic frequency across palm populations in defaunated sites37.
Our results also showed that biogeographical origin has a signature on the genetic structure of E. edulis populations, with allele replacement being the main driver of allelic richness differences between rainforest and semideciduous forests. These two biogeographical regions show marked differences in annual rainfall and predominant soil types38. There are currently few studies that compare differences in biodiversity composition (genotype, phenotype and species level diversities) between these two forest types31,39,40. Nevertheless, the few available studies show ample differences in plant community composition31 among these biogeographical regions. For example, species in semidecidous forest are able to cope with longer dry seasons31. Thus, we would expect differences in genetic composition among plant population between rain and semidecidous forest, given the differences among plant community composition between the two biogeographical regions and the potential for a marked phylogeographic signature on genetic diversity. These signals of biogeographical regions on genetic variability distribution can be a result of dispersal limitation within each of the two forest types, genetic drift, and local adaptation of E. edulis to these different domains of the Atlantic forest35. These historical influences would be “overlaid” on top of the defaunation effects: the DAPC directly contrasting the replicated populations in defaunated and non-defaunated scenarios (i.e., ignoring biogeographic region) indicated that the loss of large dispersers had a distinct signal on genetic structure of E. edulis populations, remarkably differentiating the two scenarios despite the variation due to forest physiognomy. Remarkably, the magnitude of this effect appears as large as the magnitude of differentiation when comparing populations from the two biogeographic regions. The large-scale fragmentation and defaunation in Atlantic forest is a recent anthropogenic impact. Despite that, our results reveal that the strong selective pressure imposed by defaunation on the genetic pool of E. edulis can be comparable in terms of its effects size with signals left by biogeographical process. If we assume that smaller values of β-allelic diversity represent a higher homogenization of genetic pools among populations, ongoing defaunation is driving a stronger genetic homogenization than we would expect just from historical, phylogeographic dynamics.
Our results provide a rigorous test, within an adequately replicated setup of distinct landscape types, of the hypothesis that contemporary defaunation is affecting the diversity of genetic pools of animal-dispersed tropical plants. This supports previous findings10,19,33,41,42 and strengthens the evidence for pervasive alterations of forest recruitment in disturbed tropical landscapes. Moreover, our results contradict predictions43 that rapid genetic changes are not as important as ecological degradation in human-disturbed habitats15. Our previous studies showed that the loss of large-bodied frugivores and different biogeographical origin are causing subtle changes in the selection regimes that drive rapid phenotypic evolution in E. edulis9. We found that this rapid evolutionary change in seed size and limited long distance seed dispersal due to the functional extinction of large frugivores are strong enough to have imprinted the gene pool characteristics of plant populations. These results have broad implications because most tree species are animal dispersed in tropical forests44 and defaunation of large frugivores is becoming omnipresent in fragmented and non-fragmented forests9. Therefore, the lack of seed dispersers is not only leading to phenotypic changes but also genotypic diversity and gene pool structure, with unknown effects on the long-term persistence of plant species and entire communities. This genetic erosion due to defaunation may be crucial to plants facing ongoing climate change scenarios.
Material and Methods
Study sites and species
The Brazilian Atlantic forest is an ideal system to test the effects of antrophocene defaunation on ecological and evolutionary processes. The Atlantic forest originally covered from northeast to south of the Brazilian territory and parts of Argentina and Paraguay45, but today it has been reduced to 12% of its original 150 million ha46. This reduction entailed a sustained and dramatic process of habitat fragmentation during the last ~500 yr resulting in islands of wild habitat surrounded by crops, pastures and urban matrix and an alarming loss of biodiversity45. The Brazilian Atlantic forest can be divided into several biogeographical regions based on climate, altitude and biodiversity composition30. Our study covers two of these biogeographical regions: Serra do Mar (here after rain forest) and Interior (here after semideciduous forest). Rain forests occur along the Brazilian Atlantic coast and receive high annual rainfall, whereas semideciduous forests have marked rainfall seasonality and occur in the inland Atlantic forest47.
The heart of palm (Euterpe edulis, Arecaceae), a threatened Atlantic forest species, was once one of the dominant palms in this ecosystem17. The species is present in rain forests and semideciduous forests, where it is restricted to wet microhabitats. Although once abundant, this palm species is currently endangered and locally extinct in many areas owing to illegal harvesting of the edible meristem (heart of palm;48). E. edulis is a self-compatible monoecious species, but with predominant outcrossed reproduction49 and pollination performed mainly by small-sized bees (e.g., Trigona spinipes). Their fruits are eaten by more than 58 birds and 20 mammalian species but are dispersed mostly by a reduced subset of large frugivorous birds and thrushes (Turdus spp.;9,50).
Sampling design and hypotheses for genetic differentiation
We sampled 19 sites from two different biogeographical regions, rain forest and semideciduous forest. The sites had a distinct percentage of forest cover (Supplementary Information Table S1) and well documented bird community composition (Supplementary Information Table S6). Historically, all sites shared a similar assemblage of seed dispersers, but forest fragmentation and hunting have impoverished the assemblage of large frugivores in many sites9. We sampled 30 seedlings (total of 545 individuals) of E. edulis in each site. To avoid or minimize the effects of spatial genetic autocorrelation, we spaced the sampled seedlings at least 10 m apart. This sampling design allowed us to test distinct hypotheses, once we sampled small and large patches sizes with distinct bird compositions and biogeographical regions (Fig. 1A, see details in Supplementary Information Table S1).
We isolated genomic DNA by using the CTAB extraction procedure and genotyped all individuals using eight microsatellite loci (EE5, EE8, EE25, EE43, EE45, EE47, EE52, EE63), following PCR protocol described by Gaiotto et al.51. We sized DNA fragments on an ABI Prism 3100 automated DNA sequencer (Applied Biosystems, CA) with the GeneScan ROX 500 size standard (Applied Biosystems, CA), and scored alleles with GeneMapper v4.1 software (Applied Biosystems, CA). We genotyped ten percent of all individuals two times in independent PCR amplifications to check for genotyping error and, we also checked for alleles dropout and null alleles (see details on Supplementary Information Table S2).
Testing hypotheses for genetic differentiation
We used Discriminant Analysis of Principal Component (DAPC52) to find out which hypothesis (defined as a priori groupings of the local palm populations according to defaunation status, biogegraphical region, forest fragmentation status and sampling) maximized the differences among groups (clusters). The sampled palm populations were defined according to a replicated design across the different strata defined by each hypothesis (see Fig. 1A). DAPC relies on data transformation using Principal Component Analysis as an a priori step to a Discriminant Analysis52. DAPC optimizes the separation of individuals into the pre-defined groups52. We defined these groups of sampled genotypes according to each of the different hypotheses and ran separate DAPC analyses for each. We used the DAPC to estimate the percentage of correct assignment (proportion of genotypes correctly assigned to their actual sample group) to test how well each hypothesis could discriminate among palm recruits based on grouped genotypes52. Thus, we are not using DAPC as an exploratory analysis to find k-clusters, but as a statistical method to test alternative hypotheses about differences among pre-defined groups of samples. To carry out the analysis, we first classified the individuals into groups according to each hypothesis (e.g., samples coming from defaunated sites vs. samples from pristine sites). Then, we conducted a DAPC and calculated the proportion of correct assignment of individuals into each hypothesis (defaunation, biogegraphical regions, forest fragmentation and sample design). To avoid over-fitting, we retained 50 principal components, corresponding to 82% of the total genetic information. We carried out the DAPC using the R package adegenet53. To verify whether the proportion of assignment of each hypothesis differed from a random distribution, we constructed a null model with 1000 permutation and a 95% confidence interval. We also calculated the Kappa estimator, which estimates the mean proportion of corrected assignment after removing correct assignment by chance54. We interpreted the hypothesis that had highest proportion of corrected assignment as the most likely hypothesis for distribution of genetic variation.
Genetic variability and contemporary effective population size
We assessed the influence of defaunation and biogegraphical regions on genetic variability and contemporary effective population size across the study sites. We estimated genetic diversity (He - expected heterozygosity under Hardy-Weinberg equilibrium, following Nei55), allelic richness based on rarefaction analysis (AR56) and inbreeding coefficient (f, obtained from analysis of variance of allelic frequency57) using the software FSTAT 220.127.116.118. We also estimated contemporary effective population size (Ne), using the software NeEstimator59. To test whether defaunation and biogegraphical regions are correlated with variation in genetic variability and contemporary effective population size, we conducted a generalized linear model (GLM), after testing for spatial autocorrelation (see details on Supplementary Information Table S3).
Turnover component of allelic richness
The difference in allelic richness among sites may be mainly due to allele replacement (turnover) or richness differences (nestedness-resultant dissimilarity32). We estimated each component using the R package betapart60, which partitions allelic richness dissimilarity into turnover and nestedness components. In contrast to Diniz-Filho et al.32 that estimated these components using presence/absence of alleles, we carried out the analysis using allelic frequency. We used the Bray-Curtis’s dissimilarity index to estimate β-allelic diversity and its components, and we also calculated the relative importance of the turnover component for the overall allelic richness dissimilarity. We carried out the β-allelic diversity analysis in two ways. First, we randomly sampled all palm recruits from one site of each group of each hypothesis (i.e. one from defaunated and other from non-defaunated sites or one from rain forest and other from semideciduous forest) and estimate the β-allelic diversity and its components among these groups (i.e. β-allelic diversity between defaunated and non-defaunated sites or between rain forest and semideciduous forest). This procedure were carried out 499 times (bootstrapping) and we obtained the mean, standard deviation and 95% confidence interval for each β diversity component. We constructed a null model with 1000 permutation and a 95% confidence interval to verify whether the results of the β diversity analysis for each hypothesis differed from random. With this analysis we could estimate the importance of each β-allelic diversity component for the overall allelic richness dissimilarity between defaunated and non-defaunated sites or between rain forest and semideciduous forest. Second, we randomly sampled all palm recruits from two sites of each group (i.e. two sites from defaunated sites) and estimate the β-allelic diversity and its components among the sites from each group. We carried out this procedure 499 times (bootstrapping) and obtained the mean, standard deviation and 95% confidence interval for each β diversity component. The same was done for all groups of all hypotheses (defaunated, non-defaunated, rain forest and semideciduous forest). With this analysis we could compare and contrast the β-allelic diversity results between defaunated and non-defaunated sites or between rain forest and semideciduous forest. For example, we expected that β diversity might be lower among individuals in defaunated sites compared to non-defaunated sites.
Spatial component of genetic differentiation
The spatial distribution of individuals and populations are also important factors that affect genetic differentiation patterns. For example, some populations can follow an isolation by distance model while others may follow an island model. To deal with spatial effects on genetic differentiation, we conducted a Redundancy Analysis (RDA). RDA is a multivariate method that assesses the influence of a matrix of independent variables (e.g. geographic coordinates) on a matrix of dependent variables (e.g. alleles frequency)61. RDA also allows the use of environmental data as independent variables61. We performed a RDA using an allele frequency matrix as the response variable and the spatial variable, defaunation and biogeographical regions as the explanatory variables. For RDA model selection, we performed RDA using the forward selection method with double stop criteria of Blanchet62. To reduce the incorporation of too many variables into the model and to not inflate type I error, Blanchet et al.62 proposed a forward selection based on two criterion: the forward selection is stop if either the significance level alpha is reached or the global adjusted coefficient of multiple determination (R2adj) is exceeded. We used the adegenet R package53 to calculate allele frequencies for each site and packfor63 and vegan64 R package to carry out RDA and forward selection.
How to cite this article: Carvalho, C. S. et al. Defaunation leads to microevolutionary changes in a tropical palm. Sci. Rep. 6, 31957; doi: 10.1038/srep31957 (2016).
Dirzo, R. et al. Defaunation in the Anthropocene. Science 345, 401–406 (2014).
Galetti, M. & Dirzo, R. Ecological and evolutionary consequences of living in a defaunated world. Biol. Conserv. 163, 1–6 (2013).
McConkey, K. R. et al. Seed dispersal in changing landscapes. Biol. Conserv. 146, 1–13 (2012).
Anderson, S. H., Kelly, D., Ladley, J. J., Molloy, S. & Terry, J. Cascading effects of bird functional extinction reduce pollination and plant density. Science 331, 1068–1071 (2011).
Wotton, D. & Kelly, D. Frugivore loss limits recruitment of large-seeded trees. Proc. R. Soc. B 278, 3345–3354 (2011).
Holbrook, K. M. & Loiselle, B. A. Dispersal in a Neotropical tree, Virola flexuosa (Myristicaceae): Does hunting of large vertebrates limit seed removal? Ecology 90, 1449–1455 (2009).
Pérez-Méndez, N., Jordano, P., García, C. & Valido, A. The signatures of Anthropocene defaunation: cascading effects of the seed dispersal collapse. Sci. Rep. 6, 24820 (2016).
Pérez-Méndez, N., Jordano, P. & Valido, A. Downsized mutualisms: consequences of seed dispersers’ body-size reduction for early plant recruitment. Perspect. Plant Ecol. Evol. Syst. 17, 151–159 (2015).
Galetti, M. et al. Functional extinction of birds drives rapid evolutionary changes in seed size. Science 340, 1086–1090 (2013).
Markl, J. S. et al. Meta-analysis of the effects of human disturbance on seed dispersal by animals. Conserv. Biol. 26, 1072–1081 (2012).
Carvalho, C. S., Ribeiro, M. C., Cortes, M. C., Galetti, M. & Collevatti, R. G. Contemporary and historic factors influence differently genetic differentiation and diversity in a tropical palm. Heredity 115, 216–224 (2015).
García, C. & Grivet, D. Molecular insights into seed dispersal mutualisms driving plant population recruitment. Acta Oecologica 37, 632–640 (2011).
Voigt, F. A., Arafeh, R., Farwig, N., Griebeler, E. M. & Böhning-Gaese, K. Linking seed dispersal and genetic structure of trees: a biogeographical approach. J. Biogeogr. 36, 242–254 (2009).
Epps, C. W. & Keyghobadi, N. Landscape genetics in a changing world: disentangling historical and contemporary influences and inferring change. Mol. Ecol. 24, 6021–6040 (2015).
Bacles, C. F. E. & Jump, A. S. Taking a tree’s perspective on forest fragmentation genetics. Trends Plant Sci. 16, 13–18 (2011).
Hardy, O. J. et al. Comparative phylogeography of African rain forest trees: A review of genetic signatures of vegetation history in the Guineo-Congolian region. Comptes Rendus Geosci. 345, 284–296 (2013).
Henderson, A., Galeano-Garces, G. & Bernal, R. Field guide to the palms of the Americas. (Princeton University Press, 1997).
Pizo, M. A., Von Allmen, C. & Morellato, L. P. C. Seed size variation in the palm Euterpe edulis and the effects of seed predators on germination and seedling survival. Acta Oecologica 29, 311–315 (2006).
Pacheco, L. F. & Simonetti, J. A. Genetic structure of a mimosid tree deprived of it seed dispersed, the spider monkey. Conserv. Biol. 14, 1766–1775 (2000).
Choo, J., Juenger, T. E. & Simpson, B. B. Consequences of frugivore-mediated seed dispersal for the spatial and genetic structures of a neotropical palm. Mol. Ecol. 21, 1019–1031 (2012).
Hardesty, B. D., Dick, C. W., Kremer, A., Hubbell, S. & Bermingham, E. Spatial genetic structure of Simarouba amara Aubl. (Simaroubaceae), a dioecious, animal-dispersed Neotropical tree, on Barro Colorado Island, Panama. Heredity 95, 290–297 (2005).
Karubian, J., Sork, V. L., Roorda, T., Durães, R. & Smith, T. B. Destination-based seed dispersal homogenizes genetic structure of a tropical palm. Mol. Ecol. 19, 1745–1753 (2010).
Thibert-Plante, X. & Hendry, A. P. When can ecological speciation be detected with neutral loci? Mol. Ecol. 19, 2301–2314 (2010).
Cordeiro, N. J. & Howe, H. F. Forest fragmentation severs mutualism between seed dispersers and an endemic African tree. Proc. Natl. Acad. Sci. USA 100, 14052–14056 (2003).
Laurance, S. G. W., Stouffer, P. C. & Laurance, W. F. Effects of road clearings on movement patterns of understory rainforest birds in Central Amazonia. Conserv. Biol. 18, 1099–1109 (2004).
Nogueira, E. M., Fearnside, P. M., Nelson, B. W. & França, M. B. Wood density in forests of Brazil’s ‘arc of deforestation’: Implications for biomass and flux of carbon from land-use change in Amazonia. For. Ecol. Manage. 248, 119–135 (2007).
Fahrig, L. Effects of habitat fragmentation on biodiversity. Annu. Rev. Ecol. Evol. Syst. 34, 487–515 (2003).
Young, A., Boyle, T. & Brown, T. The population genetic consequences of habitat fragmentation for plants. Trends Ecol. Evolut. 11, 413–418 (1996).
Ribeiro, M. C. et al. In Biodiversity hotspots: Distribution and Protection of Conservation Priority Areas (eds Zachos, F. E & Habel, J. C. ) 405–434 (Springer, 2011)
Da Silva, J. M. C. & Casteleti, C. H. M. In The Atlantic Forest of South America: Biodiversity Status, Threats, and Outlook (eds Galindo-Leal, C. & Câmara, I. G. ) 43–59 (CABS and Island Press, 2003).
Oliveira‐Filho, A. & Fontes, M. Patterns of floristic differentiation among Atlantic Forests in Southeastern Brazil and the influence of climate. Biotropica 32, 793–810 (2000).
Diniz-Filho, J. A. F., Collevatti, R. G., Soares, T. N. & Telles, M. P. D. C. Geographical patterns of turnover and nestedness-resultant components of allelic diversity among populations. Genetica 140, 189–95 (2012).
Cordeiro, N. J. & Howe, H. F. Low recruitment of trees dispersed by animals in African forest fragments. Conserv. Biol. 15, 1733–1741 (2001).
Caughlin, T. T. et al. Loss of animal seed dispersal increases extinction risk in a tropical tree species due to pervasive negative density dependence across life stages. Proc. R. Soc. London B Biol. Sci. 282, 1–9 (2014).
Loveless, M. D. & Hamrick, J. L. Ecological determinants of genetic structure in plant populations. Annu. Rev. Ecol. Syst. 15, 65–95 (1984).
Beckman, N. G. & Rogers, H. S. Consequences of seed dispersal for plant recruitment in tropical forests : Interactions within the seedscape. Biotropica 45, 666–681 (2013).
Hendry, A. P. & Kinnison, M. T. An introduction to microevolution: rate, pattern, process. Genetica 112, 1–8 (2001).
Brancalion, P. H. S., Novembre, A. D. L. C. & Rodrigues, R. R. Seed development, yield and quality of two palm species growing in different tropical forest types in SE Brazil: implications for ecological restoration. Seed Sci. & Technol. 39, 412–424 (2011).
Brancalion, P. H. S. & Rodrigues, R. R. Seed size-number trade-off in Euterpe edulis . Sci. Agric. 71, 226–231 (2014).
Heuertz, M., Duminil, J., Dauby, G., Savolainen, V. & Hardy, O. J. Comparative phylogeography in rainforest trees from lower Guinea, Africa. PLoS One 9, e84307 (2014).
Moran, C., Catterall, C. P. & Kanowski, J. Reduced dispersal of native plant species as a consequence of the reduced abundance of frugivore species in fragmented rainforest. Biol. Conserv. 142, 541–552 (2009).
Wang, B. C., Sork, V. L., Leong, M. T. & Smith, T. B. Hunting of mammals reduces seed removal and dispersal of the Afrotropical tree Antrocaryon klaineanum (Anacardiaceae). Biotropica 39, 340–347 (2007).
Kramer, A. T., Ison, J. L., Ashley, M. V. & Howe, H. F. The paradox of forest fragmentation genetics. Conserv. Biol. 22, 878–885 (2008).
Jordano, P. In Seeds: the ecology of regeneration in natural plant communities (ed. Gallagher, R. S. ) 18–61 (Commonwealth Agricultural Bureau International, 2013).
Joly, C. A., Metzger, J. P. & Tabarelli, M. Experiences from the Brazilian Atlantic Forest: ecological findings and conservation initiatives. New Phytol. 204, 459–473 (2014).
Ribeiro, M. C., Metzger, J. P., Martensen, A. C., Ponzoni, F. J. & Hirota, M. M. The Brazilian Atlantic Forest: How much is left, and how is the remaining forest distributed? Implications for conservation. Biol. Conserv. 142, 1141–1153 (2009).
Morellato, L. P. C. & Haddad, C. F. B. Introduction: The Brazilian Atlantic Forest. Biotropica 32, 786–792 (2000).
Galetti, M. & Fernandez, J. C. Palm heart harvesting in the Brazilian Atlantic forest: changes in industry structure and the illegal trade. J. Appl. Ecol. 35, 294–301 (1998).
Gaiotto, F. A., Grattapaglia, D & Vencovsky, R. Genetic structure, sating system, and long-distance gene flow in heart of palm (Euterpe edulis Mart.). J. Hered. 94, 399–406 (2003).
Rother, D. C., Pizo, M. A. & Jordano, P. Variation in seed dispersal effectiveness: the redundancy of consequences in diversified tropical frugivore assemblages. Oikos 125, 336–342 (2016).
Gaiotto, F. A., Brondani, R. P. V & Grattapaglia, D. Microsatellite markers for heart of palm –Euterpe edulis and E. oleracea Mart. (Arecaceae). Mol. Ecol. Notes 1, 86–88 (2001).
Jombart, T., Devillard, S. & Balloux, F. Discriminant analysis of principal components: a new method for the analysis of genetically structured populations. BMC Genet. 11, 1–15 (2010).
Jombart, T. et al. Package ‘adegenet’. Bioinforma. Appl. Note 24, 1403–1405 (2008).
Titus, K., Mosher, J. A. & Williams, B. K. Chance-corrected classification for use in discriminant analysis: ecological applications. Am. Midl. Nat. 111, 1–7 (1984).
Nei, M. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89, 583–590 (1978).
El Mousadik, A. & Petit, R. J. High level of genetic differentiation for allelic richness among populations of the argan tree [Argania spinosa (L.) Skeels] endemic to Morocco. Theor. Appl. Genet. 92, 832–839 (1996).
Weir, B. S. & Cockerham, C. C. Estimating F-statistics for the analysis of population structure. Evolution. 38, 1358–1370 (1984).
Goudet, J. FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.2). (2002). at http://www2.unil.ch/popgen/ (Accessed: 31 May 2016).
Do, C. et al. NeEstimator v2: re-implementation of software for the estimation of contemporary effective population size (Ne) from genetic data. Mol. Ecol. Resour. 14, 209–214 (2014).
Baselga, A. & Orme, C. D. L. betapart: an R package for the study of beta diversity. Methods Ecol. Evol. 3, 808–812 (2012).
Meirmans, P. G. Seven common mistakes in population genetics and how to avoid them. Mol. Ecol. 24, 3223–3231 (2015).
Blanchet, F. G., Legendre, P. & Borcard, D. Forward selection of explanatory variables. Ecology 89, 2623–2632 (2008).
Dray, S., Legendre, P. & Blanchet, G. Packfor: forward selection with permutation. R package. Version 0.0-8/r109. (2013). Available at https://R-Forge.R-project.org/projects/sedar/ (Accessed: 31May 2016).
Dixon, P. & Palmer, M. W. Vegan, a package of R functions for community ecology. J. Veg. Sci. 14, 927–930 (2003).
This work was supported by the competitive grants CNPq 477843/2011-5, CNPq 401258/2012-2 and FAPESP 2014/01986-0 that we gratefully acknowledge. CSC received a CNPq (project no. 401258/2012-2) and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, no. 2014/01029-5) scholarship. RGC and MG have been continuously supported by grants and scholarships from CNPq. PJ thanks to Conselho Nacional de Desenvolvimento Científico (CNPq, Prog. Ciências sem Fronteiras) and a Junta de Andalucía Excellence Grant (RNM-5731), as well as a Severo Ochoa Excellence Award from the Ministerio de Economía y Competitividad (SEV-2012-0262), Spain. We thank C Brocardo, S Nazareth and A Cintra for field assistance, Fundação Florestal and IBAMA for permission to sample in State Reserves and Federal Reserves and many farm owners for permission to sample in private properties.
The authors declare no competing financial interests.
About this article
Cite this article
Carvalho, C., Galetti, M., Colevatti, R. et al. Defaunation leads to microevolutionary changes in a tropical palm. Sci Rep 6, 31957 (2016). https://doi.org/10.1038/srep31957
This article is cited by
Potential cultivation areas of Euterpe edulis (Martius) for rainforest recovery, repopulation and açai production in Santa Catarina, Brazil
Scientific Reports (2023)
Biodiversity and Conservation (2023)
Anthropogenic threats and habitat degradation challenge the conservation of palm genetic resources—an appraisal of current status, threats and look-ahead strategies
Biodiversity and Conservation (2023)
Patterns of genetic diversity and structure of a threatened palm species (Euterpe edulis Arecaceae) from the Brazilian Atlantic Forest