Acta Oecologica 99 (2019) 103447

Contents lists available at ScienceDirect

Acta Oecologica

journal homepage: www.elsevier.com/locate/actoec

Distinct seeds in contrasting habitats: Morphological and reproductive responses in to new environmental conditions T

Rafael Candido-Ribeiroa,b,*, Miguel Busarello Lauterjunga, Tiago Montagnaa, Alison Paulo Bernardia, Newton Clóvis Freitas da Costaa, Marcia Patricia Hoeltgebauma, Maurício Sedrez dos Reisa a Núcleo de Pesquisas em Florestas Tropicais, Pós-Graduação em Recursos Genéticos Vegetais, Centro de Ciências Agrárias, Universidade Federal de , Rodovia Admar Gonzaga, 1346, Florianópolis, Santa Catarina, b Department of Forest and Conservation Sciences, University of British Columbia, 2424 Main Mall, Vancouver, BC, Canada

ARTICLE INFO ABSTRACT

Keywords: Distinct environments drive particular responses in the reproduction pattern of and may have ecological Phenotypic variation and evolutionary consequences for local populations. Currently, populations of the palm Butia eriospatha are Fruit morphology found in forests and open grasslands in southern Brazil, providing opportunities to investigate how this species Habitat transformation responds to recent habitat transformations. We assessed phenology and fruit morphology in two geographically Reproductive status close populations of B. eriospatha under contrasting environmental conditions – forest and open grassland – to Seed dispersal address the following questions: 1) Are there phenotypic differences between populations of B. eriospatha oc- curring in contrasting habitats? 2) What ecological and evolutionary consequences could those differences impose on local populations? Greater variation was observed within the forest population. The grassland po- pulation showed lighter endocarps on average, reflecting a greater proportion of pulp per fruit, but smaller seeds, which may suggest plasticity, local adaptation, or both after the habitat transformation. Reproductive status is dependent on individuals' sizes in the forest environment but not for the open grassland population. Additionally, the average production of infructescences per individual is lower in the forest environment. Our findings indicate that the transformation of B. eriospatha's habitat has promoted important phenotypic changes. We emphasize the importance of forest environments in promoting dispersal selectivity, which may increase population fitness through time.

1. Introduction in each particular environment (Fragoso et al., 2003; Beckman and Muller-Landau, 2011). Local environmental conditions are critical in determining how Natural ecosystems have been drastically transformed since the rise populations perform and adapt through time (Anderson et al., of modern agriculture around the world (Foley et al., 2005; Laurance 2012). Distinct habitats with contrasting environmental conditions may et al., 2014; Newbold et al., 2015). Over the past century, humans have be associated with different population patterns in phenology driven landscape transformation across many Central and South (Duchoslav, 2009), reproduction (Seguí et al., 2018), and fruit mor- American ecosystems, generally from forested to non-forested systems phology (Munthali et al., 2012; Mkwezalamba et al., 2015), among (Pacheco et al., 2011; Ribeiro et al., 2009), yet some biomes in this other traits (Guerrero et al., 2018). Such changes in population per- region are naturally open and constrained by abiotic factors (e.g. Cer- formance arise from natural selection, where phenotypes with the rado and Pampa). During the transformation process, some tree and highest fitness prevail, from individual plasticity (Jacobs and palm species have been retained or even planted in these new, open Lesmeister, 2012; Michel et al., 2014), or from both. As a result, distinct environments. The reasons for this are not well documented for all the ecological responses, and therefore evolutionary consequences, may be ecosystems where it has happened. Nevertheless, pieces of evidence expected (Gram and Sork, 2001). In addition, responses in other or- suggest that cultural importance and human use, both economic and ganisms interacting with the plant, e.g., seed dispersers or predators, are non-economic, have generally played roles in this process (Bondar, also expected to vary depending on the number of interactions involved 1964; Harvey and Haber, 1999; Pulido and Caballero, 2006; Harvey

* Corresponding author. Department of Forest and Conservation Sciences, University of British Columbia, 2424 Main Mall, Vancouver, BC, Canada. E-mail address: [email protected] (R. Candido-Ribeiro). https://doi.org/10.1016/j.actao.2019.103447 Received 14 December 2018; Received in revised form 5 May 2019; Accepted 1 July 2019 1146-609X/ © 2019 Elsevier Masson SAS. All rights reserved. R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447

Fig. 1. Study site where the two populations of Butia eriospatha were assessed and sampled. The pictures illustrate the environment of each population (FOR – Forest, GRA – open grassland); the established plots are represented by the rectangles near each picture. All the reproductive and non-reproductive individuals are represented by plus signs (+) and black dots (•), respectively, inside each plot. et al., 2011; Santos and Mitja, 2011; Silva et al., 2012). Anecdotal eriospatha is considered critically endangered by the national environ- evidence, such as chainsaws and axes being damaged by the fibrous mental ministry in Brazil (MMA, 2014) and is listed as vulnerable on palm stems, also help to explain the maintenance of this group of plants the IUCN Red List (Noblick, 1998). It is at high risk of local extirpation in open landscapes. Despite the large-scale transformation of the South due to demographic restrictions caused by cattle predation and illegal American landscape, some fragmented forests have been left behind or poaching of adult individuals for sale as ornamental plants (Nazareno were less exploited (Ribeiro et al., 2009; Vibrans et al., 2012). These and Reis, 2013a, b), leading to a call for more studies regarding the remaining fragments vary in size, form, connectivity, and in terms of species' ecology. Despite previous efforts in characterizing the species’ current disturbances, but still are valuable sites for comparison with morphology, ecology and neutral genetic diversity, no study has con- more intensively modified environments such as open grasslands when sidered populations of B. eriospatha in natural forest environments. the same species is present under both conditions. Therefore, understanding the differences between forest and open Several palms are currently found in contrasting environments in grassland populations of B. eriospatha is critically important to clarify Brazilian biomes as a consequence of anthropogenic interventions, e.g., the possible impacts of habitat transformation on the ecology and Mauritia flexuosa (Rull and Montoya, 2014); Attalea speciosa (Barot evolution of this species. et al., 2005; Silva et al., 2012; Tucker Lima et al., 2018); A. dubia Phenotypic differences observed in natural populations can be (Meiga and Christianini, 2015); A. phalerata (Sopchaki, 2014; Tucker characterized by fruit morphological descriptors (Franco and Hidalgo, Lima et al., 2018). This represents an opportunity for the investigation 2003). Fruit and seed size variability, within and among populations of responses, risks and adaptive potential that emerge from habitat are an important component in the adaptation and evolution of trees in differentiation (Barot et al., 2005; Meiga and Christianini, 2015), tropical and sub-tropical ecosystems. This is especially true for species especially for species with urgent needs for conservation and manage- dependent on zoochoric dispersal (Howe, 1986; Wheelwright, 1993), ment strategies. This is the case for Butia eriospatha (Martius ex. Drude) including most palm species (Zona and Henderson, 1989). Dispersal is Beccari, a remarkable palm species that occurs in the Southern Plateau one of the most important factors in determining spatial distribution in of Brazil. This species is associated with the Araucaria Forest, a type of palms (Eiserhardt et al., 2011) and may be intimately related to the forest formation in the biome that has experienced process of natural selection in this taxon (e.g., Galetti et al., 2013). Butia massive transformation in the past century (Ribeiro et al., 2009). eriospatha is an important food source for animals that eat its fruits Butia eriospatha is a monoecious palm with seasonal reproduction (Nazareno and Reis, 2012), suggesting dependence on vertebrates for and predominantly allogamous with potential for self-fertilization when dispersal. Traits related to phenology and reproduction in palms are solitary despite protandrous inflorescences (Nazareno and Reis, 2012). greatly affected by the environment (Amadeu et al., 2016; Tucker Lima In general, adult individuals grow to 6 m and typically occur in clusters et al., 2018), and divergent conditions may lead to differentiation (Reitz, 1974). B. eriospatha is currently found in two contrasting habi- among populations. tats, Araucaria Forest fragments and open grasslands. Its ripe fruits are In spite of its use potential and ecological importance, little is globose and yellowish with a sweet and fleshy mesocarp – containing known about existing variation in fruit morphology, reproductive status one to three seeds inside the endocarp (Lorenzi et al., 2010) – highly and infructescence production among and within wild B. eriospatha appreciated by local communities for food (Bourscheid, 2011). Butia populations, particularly with regard to those in the Araucaria Forest.

2 R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447

Therefore, the aim of this study is primarily to describe how B. erios- 2,448 fruits from 62 different individuals. patha occurring in two contrasting environments (Araucaria Forest and Fruits were assessed for the following morphological variables: i) open grassland) vary in terms of fruit morphology, reproductive pat- fresh fruit mass (g); ii) pyrene (endocarp + seeds) mass (g); iii) max- terns and infructescence production. Secondly, we test the hypothesis of imum fruit diameter (cm); and iv) fruit length (cm). Fruit and pyrene phenotypic differences between grassland and forest environments. mass (obtained immediately after pulp removal) were assessed with a Finally, we discuss possible ecological and evolutionary consequences semi-analytical weighing scale. The maximum diameter and the length of phenotypical differences in B. eriospatha found between contrasting of fruits were measured with a digital caliper. We also estimated the environments. pulp mass (g) and the proportion of pulp per fruit using the following Pij equations: Pij=−FS ij ij, and Prij = , where Pij is the pulp mass for 2. Material and methods ( Fij ) the jth fruit from the ith individual, F is the fresh fruit mass, S is the pyrene mass and Pr is the proportion of pulp per fruit. 2.1. Study site and populations After averaging the measurements of fruits for each sampled in- dividual, we used descriptive statistics to describe variation within each Two wild populations of B. eriospatha, located on a private property population. We also visualized the data sets for each morphological (−27°12′1″; −50°36′31″) in Southern Brazil, were selected for this variable in box plots for a better interpretation of how each trait varied study (Fig. 1). This region has a maritime climate (Cfb) according to within and among individuals within populations (Supplementary Köppen's classification, with annual average temperatures between material S1). 16 °C and 17 °C and precipitation between 1,600 mm and 1,700 mm In order to test the differences in fruit morphology between the two (Alvares et al., 2014). This area is characterized by plateau areas under environments we fitted a mixed effect model for each descriptor (de- the influence of the Araucaria Forest, where B. eriospatha shows a broad pendent variable) using the lme function implemented in the package but discontinuous distribution, generally found in altitudes between nlme (Pinheiro et al., 2018) in R software – v. 3.3.0 (R Core Team, 800 m and 900 m (Bourscheid, 2011). According to the property owner, 2016). For each model, habitat (forest and grassland) was considered a the adult individuals of B. eriospatha are more than 100 years old and fixed effect and the fruits were nested within each reproductive in- were used for decades for crina vegetal production (palm fibers pro- dividual to be counted as a random effect with a nested structure duced from leaves for multiple purposes, such as mattress and cushion (random intercept). This approach was used to control for non-in- manufacturing). dependence among the measured fruits within each individual One of the selected populations (GRA) occurs in a landscape of open (Harrison et al., 2018). Likelihood ratio tests were performed between grassland where all trees except B. eriospatha were removed more than full and reduced models to test the significance of the estimated coef- 60 years ago for timber production and livestock has been raised in this ficients for the fixed effect and consequently the significance of the area for more than 100 years (property owner, personal communica- difference between FOR and GRA. tion). The maintenance of B. eriospatha is likely associated with the We used non-parametric approach to compare DBH and height be- crina vegetal production, the appreciation for its fruits, and the use of tween the two habitats using 95% confidence intervals constructed with leaves and fruits to feed the livestock. In this population, all B. erios- bootstraped samples (1,000 samples for 95%). The bootstraps were patha individuals are exposed to direct sunlight. Management practices constructed using standard functions (resampling) in R software – v. including frequent mowing and occasional burning has been applied to 3.3.0 (R Core Team, 2016). If the confidence intervals from the two this area, which maintains the open character of the landscape (Fig. 1). populations overlapped, the means were not considered significantly The other population (FOR) is situated in a fragment of Araucaria different. Forest, where B. eriospatha individuals grow beneath the canopies of Relationships among all morphological descriptors of fruits within larger trees, such as and Podocarpus lambertii. populations were analyzed using Spearman correlation implemented in Systematic, low-impact timber harvesting has been practiced in this the package “Hmisc” (Harrell, 2019) in R software – v. 3.3.0 (R Core forest fragment within the past century but left little impact on the Team, 2016). The relationships were depicted in scatterplots with forested landscape of the fragment (Fig. 1). This area does not receive trendlines and visually compared to assess possible differences between any other management interventions besides the constant presence of the two populations. cattle. We established one permanent plot in each environment for ob- servations (220 m × 220 m in FOR; and 220 m × 260 m in GRA). All B. 2.3. Reproductive status eriospatha individuals (sample units) within each plot were measured for diameter at breast height (DBH) and total height, and were classi- In order to identify and quantify reproductive individuals, all B. fied as adults, according to Nazareno and Reis (2013a). The two plots eriospatha individuals (sample units) within each plot were monitored are approximately 2 km apart and, since there appears to be low genetic over one reproductive season (October 2015 to March 2016). An in- differentiation between the two populations (based on neutral nuclear dividual was considered reproductive if any reproductive structure – markers; R. Candido-Ribeiro, unpublished data), we assume they male/female inflorescence or infructescence – was observed during the formed a single population in the forest environment prior to human season. A generalized linear model using logistic regression and fol- landscape transformation. Additionally, no major differences are ex- lowing a binomial distribution was used to test the hypothesis of po- pected in climatic conditions between them. sitive association between individual size and reproductive status in each population, and to verify if the populations diverge in terms of 2.2. Morphological descriptors of fruits size-related reproductive patterns. We used the model equation: μ ⎞ gμ()=+ β βX, wheregμ( ) is the logit function (ln ⎟, μ is the 01 ()1 − μ We sampled fruit from 32 to 22 randomly selected reproductive ⎠ individuals (sample units) within the GRA and FOR plots, respectively. predicted reproductive status given the growth variable X (either

Due to the low number of individuals bearing mature fruit in the FOR height or DBH) and β0 and β1are the estimated intercept of the model plot, we also sampled eight individuals just outside the plot but still and the estimated coefficient of the variable X , respectively. within the same forest fragment, totaling 30 sampled individuals for The growth variables used to fit the models (height and DBH) were this population. We collected 40 mature fruits from each sampled in- fitted separately to verify which of the variables best predicts re- dividual (single infructescences) in the GRA population and 31–40 productive status of individuals. For this, we used the “nlme” package mature fruits per individual in the FOR population, producing a total of (Pinheiro et al., 2018) in R software – v. 3.3.0 (R Core Team, 2016). The

3 R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447 models were compared based on the Akaike Information Criterion (AIC) Table 2 (Burnham and Anderson, 2002). A log-likelihood ratio test was per- Differences between grassland (GRA) and forest (FOR) B. erispatha populations formed to test the fitted model against a reduced model (without the for the measured fruit morphological descriptors. p-values were extracted from explanatory variable) and to obtain the significance of the coefficients. the likelihood ratio tests between full and reduced models. Populations 2.4. Number of infructescences through time FOR GRA ff The di erence between the two environments concerning timing NM 30 32 and number of infructescences per individual was also tested. We ex- NF 1,168 1,280 – amined randomly selected reproductive individuals (sample units) five NF/M 31 40 40 H‾ (m) ns 5.49 6.25 times from October 2015 to March 2016 in the FOR (NP = 38) and in ns fi DBH (cm) 32.14 34.7 the GRA (NP = 34) plots, respectively. The rst and the second ob- Fruit descriptors servation were performed approximately three months apart from each Fresh fruit mass (g) ns Mean ± SD 6.55 ± 1.92 6.94 ± 1.78 other and the three last observations were performed approximately (p = 0.308) Range (min. - 3.158–13.50 3.54–11.71 every two weeks. The total number of infructescences on assessed in- max.) CV (%) 29% 26% dividuals (Ni) and the average number of infructescences per individual (∑ Ni ) in each population were obtained for each observation time. We Pyrene mass (g) * Mean ± SD 1.88 ± 0.45 1.54 ± 0.44 NP (p = 0.015) Range (min. - 1.257–3.11 0.70–2.68 modeled the number of infructescences per individual in relation to max.) time by fitting a Generalized Linear Mixed Model, using a Poisson re- CV (%) 24% 29% gression (Supplementary material S2 for detailed information). Max. diameter of fruits (cm) Mean ± SD 2.25 ± 0.24 2.32 ± 0.21 ns 3. Results (p = 0.124) Range (min. - 1.68–3.05 1.851–2.82 max.) CV (%) 11% 9% 3.1. Population and fruit descriptors Fruit length (cm) ns Mean ± SD 2.05 ± 0.19 1.97 ± 0.19 – – −1 (p = 0.187) Range (min. - 1.69 2.38 1.55 2.33 We observed 109 adult individuals in the GRA plot (19.1 ind ha ), max.) −1 101 of which were reproductive (17.7 ind ha ), and 209 in the FOR CV (%) 9% 10% −1 −1 plot (43.2 ind ha ), 58 of which were reproductive (12.0 ind ha ). Pulp mass (g) ns Mean ± SD 4.66 ± 1.67 5.40 ± 1.44 The maximum number of infructescences observed (including both ripe (p = 0.065) Range (min. - 1.90–11.32 2.68–9.72 and unripe fruits) during the assessments was 108 in GRA (across 34 max.) individuals), ranging from one to six infructescences per individual, and CV (%) 36% 27% with a maximum average of 3.18. This value was greater (p < 0.05) Proportion of pulp per fruit Mean ± SD 0.70 ± 0.06 0.78 ± 0.03 than the maximum average number of infructescences per individual *** – – observed in the FOR population (1.97; in 38 individuals), which ranged (p < 0.001) Range (min. - 0.59 0.84 0.70 0.83 max.) from one to four and totaled a maximum of 75 infructescences during CV (%) 9% 4% the assessments (Table 1). When all the individuals within the plots were considered, no significant difference was observed for height be- NM: number of individuals sampled for fruit morphology. NF: number of sam- tween the two populations (p > 0.05). However, for the mean DBH, pled fruits. NF/M: number of sampled fruits per individual. H‾ : average height FOR presented a significantly higher value (p < 0.05) (Table 1). (m) of sampled individuals. DBH: average DBH (cm) of sampled individuals. SD: Standard deviation. ***: p ≤ 0.001. *: p ≤ 0.05. ns: p > 0.05. CV: coef- We observed significant differences (p < 0.05) between the GRA ficient of variation. and FOR populations for two of the six studied morphological fruit descriptors. The GRA population showed the greatest mean value for greater in the FOR population (1.88 g, p = 0.015) (Table 2). The FOR the proportion of pulp per fruit (0.78, p < 0.001) (Table 2); and the population also showed the widest range for most fruit descriptors. mean value of pyrene mass (endocarp + seeds) was significantly However, the coefficient of variation (CV) was similar between the populations for most of the descriptors, with exception for the pulp Table 1 mass and the proportion of pulp per fruit (Table 2). The average size Population descriptors for the sampled B. eriospatha grassland (GRA) and forest (FOR) populations. (height and DBH) of sampled individuals for fruits collection was not significantly different between the two populations (p > 0.05). Var- Population descriptors Populations iance in the evaluated descriptors was high both within and among

FOR GRA reproductive individuals of each population (Supplementary material S1). − N (ind ha 1) 43.2 19.1 H‾ (m) ns 5.78 5.65 DBH(cm)* 33.37 31.92 3.2. Correlation between morphological descriptors −1 NREP (ind ha ) 12.0 17.7 N /N 0.28 0.92 REP With the exception of correlations of proportion of pulp per fruit NP 38 34

Ni 75 108 with fresh fruit mass, fruit maximum diameter, and fruit length in both * Ni/NP 1.97 3.18 populations, and the correlation between proportion of pulp per fruit

− and pyrene mass in the FOR population, and between proportion of N: number of individuals per hectare (ind ha 1). H:average height (m) of in- ‾ pulp per fruit and pulp mass in the GRA population (p > 0.05), all dividuals within the plot. DBH: average DBH (cm) of individuals within the −1 other correlations were significant in both populations (Fig. 2). While plot. NREP: number of reproductive individuals per hectare (ind ha ). NREP/N: proportion of reproductive individuals within the plot. N : number of assessed the two populations presented similar correlations between most mor- P ff individuals for phenology. Ni: maximum number of infructescences on the as- phological traits, some di erences were detected. sessed individuals. Ni/NP: average number of infructescences per assessed in- The correlation between proportion of pulp per fruit and pyrene dividual. *: p ≤ 0.05. ns: p > 0.05. mass was negative and significant for the GRA population (rs = −0.50;

4 R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447

Fig. 2. Scatterplots with the Spearman correlations (rs) among all the measured fruit morphological descriptors in B. eriospatha. Trendlines were added to facilitate the visualization of each association. p-values for the correlations are depicted within each graph. FOR: forest population. GRA: grassland population. p = 0.004), contrasting with a non-significant correlation in the FOR 3.3. Reproductive status population (rs = −0.31; p = 0.094; Fig. 2). The correlation of the proportion of pulp per fruit with the pulp mass was positive and sig- A positive and significant association between size (both height and nificant in FOR (rs = 0.50; p = 0.005), while non-significant in GRA DBH) and reproductive status of individuals was observed in the FOR (rs = 0.16; p < 0.368). The correlation between fresh fruit mass and population (p < 0.001 and p < 0.05, respectively). However, height pulp mass for the GRA population (rs = 0.97; p < 0.001) and for the (ΔAIC = 0.0) was found to be a better predictor of reproductive status FOR population (rs = 0.96; p < 0.001), in addition to the correlation over DBH (ΔAIC = 15.5) in this population (Fig. 3A and B). No sig- between fresh fruit mass and maximum fruit diameter for GRA popu- nificant association between individual size (height or DBH) and re- lation (rs = 0.93; p < 0.001) and FOR population (rs = 0.92; productive status was observed in the GRA population (p = 0.519 and p < 0.001) were the strongest positive correlations observed (Fig. 2). p = 0.506, respectively), indicating that all classes of height or DBH present a similar proportion of reproductive individuals when B. eriospatha is in an open grassland environment (Fig. 3C and D).

5 R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447

Fig. 3. Logistic regressions representing the association between reproductive status – the probability of individuals of B. eriospatha showing any reproductive structure – and size. A and C - association between height (m) and reproductive status in the FOR and GRA populations, respectively. B and D - association between DBH (cm) and reproductive status in the FOR and GRA populations, respectively. Each point in each graph represents an assessed individual palm. Confidence intervals (95%) are represented by dashed lines. p values smaller than 0.05 represent a significant association.

3.4. Number of infructescences through time 4.1. Population and fruit descriptors

A significant difference between the two populations was observed The two studied populations presented a remarkable difference in for the mean number of infructescences (p < 0.05) at most observation census size. The FOR population has more than twice the total number times (Fig. 1 – Supplementary material S2). The largest difference was of individuals per hectare when compared to the GRA population (43.2 − − observed when the populations reached their maximum production of ind ha 1 and 19.1 ind ha 1, respectively) and is comparable to what infructescences per individual i.e., between 89 and 108 days of as- was found by Nazareno and Reis (2013a) in three other grassland po- − sessment ≈ January 2016 (Fig. 1, Supplementary material S2). The pulations studied in the same region (39, 58 and 40 ind ha 1). The difference between FOR and GRA for the number of infructescences same authors also studied another grassland population with a similar − produced through time was confirmed to be highly significant number of individuals to our grassland population (16 ind ha 1). Since (p < 0.001) after the log-likelihood ratio test between the full and the the FOR and GRA populations are close to each other, it is assumed they reduced models (Supplementary material S2). once formed a single population, before the transformation of the landscape which occurred at least 60 years ago. Therefore, the observed difference in census size between the two populations, together with 4. Discussion the fact that fallen and dying palms were seen in the grassland area (personal observation), suggests that the landscape change may also This is the first attempt to report and compare the variation in fruit have affected the demographic dynamic of the species in addition to morphology, reproductive status and production of infructescences in fruit morphology and reproductive patterns. Nevertheless, long term two populations of B. eriospatha that currently occur in different en- demographic studies with more populations are necessary to confirm vironments, forest and open grassland. Despite the short distance be- this pattern. Another study of four open grassland populations of B. tween the two populations (≈2 km), we detected significant differences eriospatha estimated a reduction up to 50% in effective population size across two important fruit morphological descriptors, reproductive over the next 40 years (Nazareno and Reis, 2013a), owing to mortality patterns, production of fruits, and demographic parameters. These and low seedling recruitment, which aligns with the observed differ- differences have likely been triggered by differences in local environ- ence between FOR and GRA in our study. mental conditions resulting from the transformation of the landscape. The FOR and GRA populations differed significantly in two of the Local adaptation or extirpation due to evolutionary forces are expected evaluated fruit traits. While GRA presented a higher proportion of pulp through time (disregarding the effect of cattle predation). Here we per fruit, FOR showed fruits with heavier pyrenes (endocarp + seeds), discuss how differences between the habitats may be playing a role in which reflected a lower proportion of pulp. FOR presented the broadest local adaptation of B. eriospatha populations. range of values for most of the studied traits. This suggests that, with the transformation of the landscape, GRA was submitted to distinct selective pressures (e.g., light and water availability), pushing the

6 R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447 population to express different phenotypes (with smaller seeds and 1998), and Acromia aculeata (Scariot et al., 1995). In our study, this more pulp per fruit), but using the preexisting genetic variation from pattern was not observed in open grassland environment, where almost the original environment (forest). all individuals were reproductive. In the GRA population, the re- A possible explanation for the observed differences in fruits and productive status of individuals had no relation to their sizes. No cor- pyrenes between FOR and GRA is the cross-pollination rate, as sig- relation was found between DBH and reproductive status of B. erios- nificant increases in pulp yield per fruit is characteristic of self-polli- patha individuals in another open grassland population studied by nated individuals of B. odorata (Eloy, 2013). This suggests that when Nazareno and Reis (2012). This reinforces the claim that light may be a geitonogamy takes place – a form of self-pollination observed in open determining factor in the reproduction of B. eriospatha, and that the grassland populations of B. eriospatha (Nazareno and Reis, 2012) – transformation of the landscape, producing open habitats, may have pyrenes, and consequently the seeds, exhibit a reduced size, which in changed the mechanisms involved in the selection of which individuals turn, is associated with higher proportion of pulp per fruit. Similarly, will reproduce in the population. Complementary studies in B. erios- heavier pyrenes in the FOR population, and thus larger seeds, suggest patha involving multiple populations in both environments are neces- more outcrossing in this environment. Larger seeds of Mauritia flexuosa sary to validate if this pattern persists throughout the species natural – a palm species native from northern South America – were also ob- range after the habitat transformation. served in forested environments when compared to disturbed savannas Another possible factor acting upon the flowering of B. eriospatha in northern Brazil (Rosa et al., 2014). individuals under forest cover may be a reproductive fluctuation me- Differences between populations have been observed in fruit de- chanism from one season to the next. Rosa et al. (1998) observed a scriptors of other species of the Butia (B. odorata, Rivas and fluctuation in the reproductive and non-reproductive state of in- Barilani, 2004; B. capitata, Silva, 2008) but some differences may re- dividuals in a population of B. catarinensis on the southern coast of present variation between reproductive seasons (Silva, 2008). The Brazil, suggesting the possibility of individuals alternating their re- possible explanations for these differences were not thoroughly dis- productive state through time. A large variation in number of re- cussed by the authors; nevertheless, there is an argument for environ- productive individuals among reproductive events was also observed in mental and genetic causes. Euterpe edulis (Silva, 2011) and A. mexicanum (Piñero and Sarukhán, 1982), both understory palm species. Despite the evidence, long-term 4.2. Reproductive patterns studies following the reproductive phenology of wild populations of B. eriospatha are necessary to confirm the occurrence of reproductive state The proportion of reproductive individuals was nearly threefold fluctuation in forest environments and whether this mechanism occurs higher in GRA than FOR (0.92 and 0.28, respectively). This proportion in open grassland populations. in GRA was similar to that observed in previously-studied open grass- land populations (Nazareno and Reis, 2013a). The average number of 4.3. Possible evolutionary consequences infructescences per reproductive individual was higher in GRA (3.18) than in the FOR population (1.97) and similar to the number of in- Larger seeds may represent both a selective advantage and dis- florescences observed in another grassland population of B. eriospatha advantage for palm trees and could be understood as a trade-off be- studied in the same region (3.15 in 2010 and 2.93 in 2011, Nazareno tween survival and dispersal (Howe, 1986). Larger seeds of the palm and Reis, 2012). However, the values presented in this study are the Euterpe edulis, for instance, show a higher potential of germination, and number of inflorescences and therefore, might overestimate the number seedlings generated from larger seeds show higher total biomass values, of infructescences per individual. In other Butia species, average indicating a greater potential of survival (Pizo et al., 2006). A reduction number of infructescences varies substantially (Rosa et al., 1998; in seed size of the palm E. edulis may even decrease the average fitness Soares, 2013; Amadeu et al., 2016). of a population (Galetti et al., 2013). Furthermore, while studying seed The high environmental heterogeneity common in forested en- size in different groups of tropical tree species, Foster (1986) suggested vironments (e.g., light, water, wind, moisture and soil nutrient avail- that larger seeds could either increase the ability of seedlings to persist ability), alongside higher inter- and intraspecific competition, are pos- in an environment with low light intensity beneath a canopy cover, or sible explanations for the FOR population's low proportion of even accelerate the initial vegetative growth of seedlings to reach strata reproductive individuals showing fewer infructescences. Palms occur- with available light. On the other hand, an increase in seed size may ring in the understory of forest environments may show accelerated imply a reduction in the number of seeds produced by the population as growth and higher reproductive rates in light gaps produced by falls or a whole, since more energy is allocated to producing larger seeds the death of larger trees (Ash, 1988). For populations of Astrocaryum (Foster, 1986). This pattern is observed in the FOR population where mexicanum, a palm that occurs under the canopy of tropical rainforests the number of reproductive individuals and the number of in- in Mexico and Central America, the influx of light into the forest pro- fructescences per plant were significantly lower than in the GRA po- duced by gaps likely plays a fundamental role in the reproductive status pulation, but the pyrenes (and, by inference, seeds) were larger. This of adult individuals (Piñero and Sarukhán, 1982). After studying the reduction in reproductive potential in the FOR population may re- relationship of growth rates and reproductive status in Geonoma mac- present a selective pressure, resulting in fewer fruits and seeds pro- rostachys – an understory palm – with the presence of canopy gaps, duced by the population, but with seedlings generated by more capable Svenning (2002) posited that fine-scale spatial and temporal variation and well-adapted individuals, which tends to increase the adaptive in the opening of canopies of tropical trees plays a key role in the value of the population as a whole (Moore, 2001). A reduction in the ecology of shade-tolerant understory plant populations. Since B. erios- total number of fruits produced by a population may also increase the patha individuals occurring in forest environments do not occupy the selectivity of dispersers in finding the most rewarding individuals upper canopy and are therefore under the influence of gap dynamics, (Howe, 1986). light availability is possibly one of the main determining factors in the From another adaptive perspective, fruit characteristics in some reproductive potential of individuals. groups of plants have been found to be associated with variation in seed In the forest environment, taller and thicker B. eriospatha in- dispersal rates (Herrera, 1988). Brewer (2001) observed that for the dividuals are more likely to have a reproductive structure than shorter palm A. mexicanum, smaller seeds might be less predated and dispersed and thinner ones, and height is a better predictor of reproduction than at longer distances, which contrasts with the adaptive advantages DBH. Significant associations between size and reproductive status conferred by larger seeds described above. The same trade-off was were also observed in other palms in their natural habitats, e.g., A. observed for other tropical trees in previous studies (Howe, 1986). mexicanum (Piñero and Sarukhán, 1982), B. catarinensis (Rosa et al., Therefore, the larger seeds of B. eriospatha observed in the forest

7 R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447 environment may increase the species’ chance of persisting in this type Acknowledgement of environment, since it potentially confers better germination and greater seedling survival in the population. In contrast, a higher number This work was supported by the National Council of Technological of smaller seeds, observed in the open grassland environment, possibly and Scientific Development (CNPq), process no. FAPESC/2780/2012-4 represents a greater chance for dispersing across longer distances with a (PRONEX), for the Master’s scholarship to R.C.R (130894/2015-0), for lower probability of predation. Hence, there seems to be a compensa- the doctoral scholarship to M.P.H. (140386/2013-0), and for the award tory adaptive mechanism in B. eriospatha populations once the forest to M.S.R. (309128/2014-5), and the Coordination for the Improvement habitat is converted to open grasslands habitats, where germination and of Higher Education Personnel (CAPES) for the Master's scholarship to survival after germination may be reduced, but production of seeds and M.B.L., doctoral scholarships to A.P.B., T.M., and N.C.F.C. We would their dispersal could be more effective. While seeds are more exposed to like to thank Jonathan Degner, and the two anonymous reviewers for predation in open environments, colonization of new areas and avoid- the thoughtful comments and English review, the Federal University of ance of density-dependent mortality near parents could be considered Santa Catarina (UFSC-Curitibanos) and the Ecology Lab for the logis- plausible advantages of a longer dispersal (Howe and Miriti, 2000). tical support, and the Chico Mendes Biodiversity Institute (ICMBio) and Considering that fruit productivity in palms can be highly influ- the landowner of the property who authorized and facilitated the enced by the environment (Freitas et al., 2016), the observed smaller sample collections. variations in fruits of the GRA population may be related to the homogeneity of conditions in the open grassland habitat e.g., in light Appendix A. Supplementary data availability. Differences in fruit productivity due to differences in light availability have been observed in other understory palms (Martínez- Supplementary data to this article can be found online at https:// Ramos et al., 2016; Amadeu et al., 2016). Moreover, other environ- doi.org/10.1016/j.actao.2019.103447. mental factors may influence patterns in fruit production of palms (e.g., Rosa et al., 2014; Schlindwein et al., 2017; Tucker Lima et al., 2018), References suggesting that habitat transformation may lead to important changes in the reproductive mechanisms and fruit traits in this group of plants. Alvares, C.A., Stape, L., Sentelhas, P.C., et al., 2014. Köppen’s climate classification map If a different selective pressure has taken place during the past years for Brazil. Meteorol. Z. 22, 711–728. https://doi.org/10.1127/0941-2948/2013/ 0507. since the transformation of the landscape and at least one new gen- Amadeu, L.S.N., Sampaio, M.B., Santos, F.A.M., 2016. Influence of light and plant size on eration was formed in the GRA population, an adaptive genetic com- the reproduction and growth of small palm tree species: comparing two methods for ponent could be involved in the phenotypic variation observed between measuring canopy openness. 103, 1–9. https://doi.org/10.3732/ajb.1600178. Anderson, J.T., Willis, J.H., Mitchell-olds, T., 2012. Evolutionary genetics of plant B. eriospatha populations. High heritability values for traits related to adaptation. Trends Genet. 27, 258–266. https://doi.org/10.1016/j.tig.2011.04.001. fruit production have been found in other palms (Passos et al., 2007; (Evolutionary). Okoye et al., 2009; Neto et al. 2013) and are expected in B. eriospatha. Ash, J., 1988. Demography and production of Balaka microcarpa burret (), a – However, future experiments are necessary to verify the genetic control tropical understorey palm in Fiji. Aust. J. Bot. 36, 67 80. https://doi.org/10.1071/ BT9880067. of each studied trait and determine which portion of the variation could Barot, S., Mitja, D., Miranda, I., et al., 2005. Reproductive plasticity in an Amazonian be attributed to the adaptive genetic component. palm. Evol. Ecol. Res. 7, 1051–1065. Beckman, N.G., Muller-Landau, H.C., 2011. Linking fruit traits to variation in predispersal vertebrate seed predation, insect seed predation, and pathogen attack. Ecology 92 (11), 2131–2140. https://doi.org/10.1890/10-2378.1. 5. Conclusions Bondar, G., 1964. Palmeiras Do Brasil. Palmae** Arecaceae**. Instituto de Botânica. São Paulo. Our study empirically demonstrated that the transformation of B. Bourscheid, K., 2011. Butia eriospatha – butiá-da-serra. In: Coradin, L., Siminski, A., Reis, A. (Eds.), Espécies nativas da flora brasileira de valor econômico atual ou potencial: eriospatha habitat in one population from Araucaria Forest to open plantas para o futuro – Região Sul. MMA, Brasília, pp. 156–158. grassland more than 60 years ago might have caused shifts in local Brewer, S.W., 2001. Predation and dispersal of large and small seeds of a tropical palm. phenotypic variation of traits related to its reproduction and fruit Oikos 92, 245–255. morphology. These findings suggest not only possible changes in pat- Burnham, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. Springer, New York. terns of fruit production, but also of seedling survival and seed dis- Duchoslav, M., 2009. Effects of contrasting habitats on the phenology, seasonal growth, persal, possibly affecting ecological interactions with other organisms. and dry-mass allocation pattern of two bulbous geophytes (Alliaceae) with partly ff – If this is true for other populations, over generations, differences in di erent geographic ranges. Pol. J. Ecol. 57, 15 32. Eiserhardt, W.L., Svenning, J.C., Kissling, W.D., Balslev, H., 2011. Geographical ecology conditions between remaining environments where of B. eriospatha of the palms (Arecaceae): determinants of diversity and distributions across spatial occurs (e.g., forest and open areas) may act as selective drivers of local scales. Ann. Bot. 108, 1391–1416. https://doi.org/10.1093/aob/mcr146. adaptation and subsequent population differentiation, speciation, or Eloy, J., 2013. Polinização, produção e qualidade de butiá (Butia odorata Barb. Rodr.) Noblick & Lorenzi. Dissertation. Universidade Federal de Pelotas. extirpation. Given the uncertain evolutionary trajectory of this critically Foley, J.A., Defries, R., Asner, G.P., et al., 2005. R EVIEW global consequences of land threatened species, further studies aiming to describe and confirm the use. 8, 570–574. https://doi.org/10.1126/science.1111772. evolutionary consequences of these phenotypic changes in local popu- Foster, S., 1986. On the adaptive value of large seeds for tropical moist forest trees: a – fi review and synthesis. Bot. Rev. 52, 260 299. https://doi.org/10.1007/BF02860997. lations of B. eriospatha associated with speci c habitats are crucial. Fragoso, J.M.V., Silvius, K.M., Correa, J.A., 2003. Long-distance seed dispersal by tapirs increases seed survival and aggregates tropical trees. Ecology 84, 1998–2006. https://doi.org/10.1890/01-0621. Author contributions Franco, T.L., Hidalgo, R., 2003. Análisis Estadístico de Datos de Caracterización Morfológica de Recursos Fitogenéticos - Boletin Técnico Nro 8. Instituto Internacional de Recursos Fitogenéticos IPGRI. RCR and MSR conceived the study and designed the sampling Freitas, C., Costa, F.R.C., Barbosa, C.E., Cintra, R., 2016. Restriction limits and main method. RCR, MBL, TM, APB, NCFC and MPH collected the data. RCR drivers of fruit production in palm in central Amazonia. Acta Oecol. 77, 75–84. analyzed the data. RCR wrote the manuscript. MBL, TM, APB, NCFC, https://doi.org/10.1016/j.actao.2016.09.003. fi Galetti, M., Guevara, R., Cortes, M.C., et al., 2013. Functional extinction of birds drives MPH and MSR edited the text. All the authors have approved the nal rapid evolutionary changes in seed size. Science 80 (340), 1086–1090. https://doi. version of the manuscript. org/10.1126/science.1233774. Gram, W.K., Sork, V.L., 2001. Association between environmental and genetic hetero- geneity in forest tree populations. Ecology 82, 2012–2021. https://doi.org/10.1890/ Declarations of interest 0012-9658(2001)082[2012:ABEAGH]2.0.CO;2. Guerrero, J., Andrello, M., Burgarella, C., Manel, S., 2018. Soil environment is a key driver of adaptation in Medicago truncatula: new insights from landscape genomics. None. New Phytol. 219, 378–390. https://doi.org/10.1111/nph.15171.

8 R. Candido-Ribeiro, et al. Acta Oecologica 99 (2019) 103447

Harrell Jr., F.E., 2019. Hmisc: Harrell Miscellaneous. R Package Version 4.2-0. 0700-0461. Harrison, X.A., Donaldson, L., Correa-Cano, M.E., et al., 2018. A brief introduction to Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., R Core Team, 2018. Nlme: Linear and mixed effects modelling and multi-model inference in ecology. PeerJ 6, e4794. Nonlinear Mixed Effects Models. R Package Version 3. pp. 1–137. https://doi.org/10.7717/peerj.4794. Pizo, M.A., Allmen, C Von, Morellato, L.P.C., 2006. Seed size variation in the palm Euterpe Harvey, C.A., Haber, W.A., 1999. Remnant trees and the conservation of biodiversity in edulis and the effects of seed predators on germination and seedling survival. Acta Costa Rican pastures. Agrofor. Syst. 44 (1), 37–68. https://doi.org/10.1023/ Oecol. https://doi.org/10.1016/j.actao.2005.11.011. A:1006122211692. Pulido, M.T., Caballero, J., 2006. The impact of shifting agriculture on the availability of Harvey, C.A., Villanueva, C., Esquivel, H., et al., 2011. Conservation value of dispersed non-timber forest products: the example of Sabal yapa in the Maya lowlands of tree cover threatened by pasture management. For. Ecol. Manag. 261, 1664–1674. Mexico. For. Ecol. Manag. 222 (1–3), 399–409. https://doi.org/10.1016/j.foreco. https://doi.org/10.1016/j.foreco.2010.11.004. 2005.10.043. Herrera, C.M., 1988. The fruiting ecology of Osyris quadripartita : individual variation and R Development Core Team, 2016. R: A Language and Environment for Statistical evolutionary potential. Ecology 69, 233–249. Computing. R Foundation for Statistical Computing, Vienna, Austria. http://www.R- Howe, H.F., 1986. Seed dispersal by fruit-eating birds and mammals. In: Murray, D.R. project.org. (Ed.), Seed Dispersal. Academic Press, Sydney, Australia, pp. 123–190. Reitz, R., 1974. Flora Ilustrada Catarinense - Palms. Herbario Barbosa Rodrigues. Itajaí. Howe, H.F., Miriti, M.N., 2000. No question: seed dispersal matters. Trends Ecol. Evol. 15 Ribeiro, M.C., Metzger, J.P., Martensen, A.C., et al., 2009. The Brazilian Atlantic Forest: (11), 434–436. https://doi.org/10.1016/S0169-5347(00)01965-0. how much is left, and how is the remaining forest distributed? Implications for Jacobs, B.S., Lesmeister, S.A., 2012. Maternal environmental effects on fitness, fruit conservation. Biol. Conserv. 142, 1141–1153. https://doi.org/10.1016/j.biocon. morphology and ballistic seed dispersal distance in an annual forb. Funct. Ecol. 26, 2009.02.021. 588–597. https://doi.org/10.1111/j.1365-2435.2012.01964.x. Rivas, M., Barilani, A., 2004. Diversidad, potencial productivo y reproductivo de Los Laurance, W.F., Sayer, J., Cassman, K.G., 2014. Agricultural expansion and its impacts on Palmares de (Mart.) becc. de . Agrociencia 8, 11–20. tropical nature. Trends Ecol. Evol. 29, 107–116. https://doi.org/10.1016/j.tree. Rosa, L., Castellani, T.T., Reis, A., 1998. Biologia reprodutiva de Butia capitata (Martius) 2013.12.001. Beccari var. odorata (Palmae) na restinga do município de Laguna, SC. Braz. J. Bot. Lorenzi, H., Noblick, L.R., Kahn, F., Ferreira, E., 2010. Flora Brasileira Lorenzi: Arecaceae 21, 3. (Palmeiras). Instituto Plantarum, Nova Odessa. Rosa, R.K., Barbosa, R.I., Koptur, S., 2014. Which factors explain reproductive output of Martínez-Ramos, M., Ortiz-Rodríguez, I.A., Piñero, D., et al., 2016. Anthropogenic dis- Mauritia flexuosa (Arecaceae) in forest and savanna habitats of northern Amazonia? turbances jeopardize biodiversity conservation within tropical rainforest reserves. Int. J. Plant Sci. 175, 307–318. https://doi.org/10.1086/674446. Proc. Natl. Acad. Sci. Unit. States Am. 1–6. https://doi.org/10.1073/pnas. Rull, V., Montoya, E., 2014. Mauritia flexuosa palm swamp communities: natural or 1602893113. human-made? A palynological study of the Gran Sabana region (northern South Meiga, A.Y.Y., Christianini, A.V., 2015. Changes in seed predation of Attalea dubia in a America) within a neotropical context. Quat. Sci. Rev. 99, 17–33. https://doi.org/10. gradient of Atlantic forest disturbance in Brazil. Palms 59, 135–144. 1016/j.quascirev.2014.06.007. Michel, M.J., Chevin, L.M., Knouft, J.H., 2014. Evolution of phenotype-environment as- Santos, AM dos, Mitja, D., 2011. Pastagens arborizadas no projeto de assentamento sociations by genetic responses to selection and phenotypic plasticity in a temporally benfica, município de Itupiranga, Pará, Brasil. Rev. Árvore 35, 919–930. https://doi. autocorrelated environment. Evolution 68, 1374–1384. https://doi.org/10.1111/ org/10.1590/S0100-67622011000500017. evo.12371. Scariot, A., Lleras, E., Hay, J.D., 1995. Flowering and fruiting phenologies of the palm Mkwezalamba, I., Munthali, C.R.Y., Missanjo, E., 2015. Phenotypic variation in fruit Acrocomia aculeata: patterns and consequences. Biotropica 27, 168–173. https://doi. morphology among provenances of Sclerocarya birrea (A . Rich.) hochst. Int. J. For. org/10.2307/2388992. Res. 1–8. https://doi.org/10.1155/2015/735418. 2015. Schlindwein, G., Tonietto, A., Azambuja, AC De, et al., 2017. Pindo Palm fruit yield and MMA, 2014. Ministério do Meio Ambiente. Portaria nº 443 de 17 de dezembro de 2014. its relationship with edaphic factors in natural populations in . Moore, D., 2001. Insects of palm flowers and fruits. In: Howard, F.W., Moore, D., Giblin- Ciência Rural. 47, 1–7. https://doi.org/10.1590/0103-8478cr20151371. Davis, R.M., Abad, R.G. (Eds.), Insects on Palms. CABI Publishing, Wallingford, UK, Seguí, J., Lázaro, A., Traveset, A., et al., 2018. Phenotypic and reproductive responses of pp. 233–266. an Andean violet to environmental variation across an elevational gradient. Alpine Munthali, C.R.Y., Chirwa, P.W., Akinnifesi, F.K., 2012. Phenotypic variation in fruit and Bot. 128, 59–69. https://doi.org/10.1007/s00035-017-0195-9. seed morphology of Adansonia digitata L. (baobab) in five selected wild populations Silva, J.Z., 2011. Fundamentos da produção e consumo de frutos em populações naturais in Malawi. Agrofor. Syst. 85, 279–290. https://doi.org/10.1007/s10457-012-9500-1. de Euterpe edulis Martius. Dissertation. Universidade Federal de Santa Catarina. Nazareno, A.G., Reis, MS Dos, 2012. Linking phenology to mating system: exploring the Silva, MR da, Júnior, OA. de C., Martins É de, S., et al., 2012. Mulivariate factor analysis reproductive biology of the threatened palm species Butia eriospatha. J. Hered. 103, applied to the characterization of Babassu (Attalea Speciosa Mart. ex Spreng) occur- 842–852. https://doi.org/10.1093/jhered/ess070. rence area in Cocal river basin. Soc. Nat. 24, 267–281. Nazareno, A.G., Reis, M.S., 2013a. At risk of population decline? An ecological and ge- Silva, P.A.D., 2008. Ecologia populacional e botânica econômica de Butia capitata (Mart.) netic approach to the threatened palm species Butia eriospatha (Arecaceae) of Beccari no no norte de . Dissertation. Universidade de Brasília. southern Brazil. J. Hered. 105, 120–129. https://doi.org/10.1093/jhered/est065. Soares, K.P., 2013. O gênero Butia (Becc.) Becc. (Arecaceae) no Rio Grande do Sul com Nazareno, A.G., dos Reis, M.S., 2013b. Where did they come from? Genetic diversity and ênfase nos aspectos ecológicos e silviculturais de (Mart.) Becc. E Butia forensic investigation of the threatened palm species Butia eriospatha. Conserv. Genet. witeckii K. Soares & S. Longhi. Dissertation. Universidade Federal de Santa Maria. 15, 441–452. https://doi.org/10.1007/s10592-013-0552-1. Sopchaki, D.P.D.S., 2014. Estratégia de colonização de Uricuri (Attalea phalerata mart. ex Neto, JT. de F., Clement, C.R., Resende, MDV de, 2013. Estimativas de parâmetros spreng. Arecaceae) em áreas antropizadas por atividade pecuária no leste do Acre. genéticos e ganho de seleção para produção de frutos em progênies de polinização Dissertation. Instituto Nacional de Pesquisas da Amazônia. aberta de pupunheira no Estado do Pará, Brasil. Bragantia, Campinas 72, 122–126. Svenning, J.-C., 2002. Crown illumination limits the population growth rate of a neo- Newbold, T., Hudson, L.N., Hill, S.L.L., et al., 2015. Global effects of land use on local tropical understorey palm (Geonoma macrostachys, Arecaceae). Plant Ecol. 159, terrestrial biodiversity. Nature 520, 45–50. https://doi.org/10.1038/nature14324. 185–199. Noblick, L., 1998. Butia eriospatha. The IUCN Red List of Threatened Species 1998. Tucker Lima, J.M.T., Caruso, N.M., Clugston, J., Kainer, K.A., 2018. Landscape change https://doi.org/10.2305/IUCN.UK.1998.RLTS.T38462A10114794.en, Accessed date: alters reproductive phenology and sex expression in Attalea palms (Arecaceae) of 9 July 2018. southwestern Amazonia. Plant Ecol. 219 (10), 1225–1245. https://doi.org/10.1007/ Okoye, M.N., Okwuagwu, C.O., Uguru, M.I., 2009. Population improvement for fresh fruit s11258-018-0874-7. bunch yield and yield components in oil palm (Elaeis guineensis Jacq .). Am. J. Sci. Vibrans, A.C., McRoberts, R.E., Lingner, D.V., Nicoletti, A.L., Moser, P., 2012. Extensão Res. 4, 59–63. original e atual da cobertura florestal de Santa Catarina. In: Vibrans, A.C. (Ed.), Pacheco, P., Aguilar-Støen, M., Börner, J., et al., 2011. Landscape transformation in Inventário florístico florestal de Santa Catarina – Diversidade e Conservação dos tropical Latin America: assessing trends and policy implications for REDD+. Forests Remanescentes Florestais. Editora da FURB. Blumenau, pp. 65–76. 2, 1–29. https://doi.org/10.3390/f2010001. Wheelwright, N.T., 1993. Fruit size in a tropical tree species: variation, preference by Passos, C.D., Aragão, W.M., Passos, E.E.M., 2007. Herdabilidade de Caracteres re- birds, and heritability. Vegetatio 107–108, 163–174. https://doi.org/10.1007/ produtivos de Cultivares de Coqueiro Anão. Rev. Brasileira Biociências 5, 336–338. BF00052219. Piñero, D., Sarukhán, J., 1982. Reproductive behaviour and its individual variability in a Zona, S., Henderson, A., 1989. A review of animal-mediated seed dispersal of palms. tropical palm, Astrocaryum mexicanum. J. Ecol. 70, 461–472 doi: 0022-0477/82/ Selbyana 11, 6–21.

9