Introduction

The tropics hold by far the largest biodiversity worldwide, especially in tropical forests. For instance, the Amazonia harbor more than 16,000 tree species (ter Steege et al. 2013), reaching in a single hectare 200 species or even more (Gentry 1992). However, it is also possible to find in the tropics monodominant forests, rare phytophysiognomies with more than 60% of the canopy dominated by a single species (Hart et al. 1989; Peh et al. 2011). Few occurrences of this unusual situation have been recorded on a pan-tropical scale (Peh et al. 2011; Brookshire and Thomas 2013; Nik Norafida et al. 2018), with just two cases described for Amazonia: Peltogyne gracilipes Ducke (Fabaceae) in north (Nascimento et al. 1997; Nascimento and Proctor 1997), and Brosimum rubescens Taub. (Moraceae) in the southern region (Marimon et al. 2001a, b).

The causes of the tropical monodominance are not yet fully understood, but surely these are not related to just a special feature, but to a set of ecological and evolutionary mechanisms (Peh et al. 2011; Brookshire and Thomas 2013; Nascimento et al. 2017; Marimon-Junior et al. 2019). Some ecological mechanisms proposed are related to the lack of disturbance (Connell and Lowman 1989; Hart et al. 1989), shade tolerance and survival of seedlings under closed cover (Hart 1995; Torti et al. 2001), and soil-nutrient factors, as Mg/Ca ratio and N dynamics along the ecological succession (Brookshire and Thomas 2013; Nascimento et al. 2017; Elias et al. 2018). The proposed evolutionary mechanisms involve functional and structural traits of monodominant species as large seeds to overcome deep litter (Torti et al. 2001; Peh et al. 2011), ectomycorrhizal association (Connell and Lowman 1989; McGuire et al. 2008) and low efficiency in seed dispersal (Hart 1985).

Current studies have integrated the knowledge of interspecific variation of functional traits to explain the coexistence of tropical tree species and the relationship between functional diversity and dominance in plant assemblies (Cornwell and Ackerly 2010; Aiba et al. 2020), including trade-offs between functional traits, growth, and mortality (Wright et al. 2010). Functional traits such as wood density, seed volume, and total height have already been identified as good predictors of tree species competition and survival rates in Neotropical forests (Poorter et al. 2008), all of them possibly important for species dominance. It has also been shown that dominance can be positively correlated with leaf mass by area, and negatively with leaf size (Aiba et al. 2020). Recent studies have looking for more consistent explanations about monodominance at the functional level, for example, in Africa Gilbertiodendron dewevrei monodominance is associated with low local functional diversity (Kearsley et al. 2017), as well as functional acquisition strategies (Hall et al. 2020).

Many species have different survival strategies that allow them to thrive in various ecological conditions (Araújo et al. 2021a, b). These strategies can manifest in the population structural parameters of a species. Parameters such as growth and recruitment rates vary between populations due to different environmental conditions as a survival strategy (Marimon et al. 2020). Also, these population structural parameters may indicate underlying conditions, e.g., functional traits, which keep an ecosystem in its climax state or alternative successional state such as monodominance (Kearsley et al. 2017). Associating population structural parameters and leaf functional traits seems a logical step in elucidating many of the dynamics in forest ecosystems.

Understanding how functional traits shape the species monodominance is a major challenge in ecology. For example, the combination of leaf functional traits can represent distinct ecological strategies (Araújo et al. 2021b), such as drought tolerance and competition ability, influencing the survival, growth, and reproduction of the organisms (Ackerly 2003; Violle et al. 2007), and consequently regulating its abundance (Aiba et al. 2020). Some species can become such efficient competitors that they end up causing local competitive exclusion (Gause 1932), increasing their abundance (Grubb 1982; Miller and Werner 1987), which can favor monodominance.

As for the monodominant forests of B. rubescens, early studies suggested that the monodominance of this species is episodic and depends on small disturbances, such as tree fall gaps (Marimon et al. 2001a, b). The most recent study evaluated whether lower soil water retention could explain the monodominance (Marimon-Junior et al. 2019). The authors rejected this hypothesis and argue that only the integration of several studies involving different mechanisms will be able to reveal the causes and consequences of tropical monodominance.

This study with B. rubescens in Southern Amazonia is the first evidence of integration of leaf functional traits explaining tropical monodominance. This monodominant forest occurs adjacent to a mixed forest, with much greater species diversity, but both under the same climatic condition (Marimon and Felfili 2006; Marimon et al. 2014; Morandi et al. 2016). We evaluated and compared the leaf functional traits of B. rubescens in both monodominant and mixed forests to answer two questions: (i) Do the leaf functional traits differ between the populations of B. rubescens occurring in the monodominant and mixed forests? (ii) Can the relationship between leaf functional traits and structural parameters explain the monodominance of this species?

Materials and methods

Study area and species description

We carried out the study in a B. rubescens monodominant forest and a mixed forest in southern Amazonia (Marques et al. 2020) (Fig. 1) 800 m apart, located at the Legal Reserve of Vera Cruz Farm (14° 50′ 47″ S e 52° 08′ 37″ W), in Nova Xavantina municipality, Mato Grosso state. The climate is Aw type according to Köppen's classification, characterized by two well-defined seasons, the rainy, from March to October, and the dry, from April to September (Alvares et al. 2013). The annual averages of precipitation and temperature in the study area are 1600 mm and 25 °C, respectively (Marimon et al. 2002, 2010). The forests grew on dystrophic Ferrasols (FAO/UNESCO 1992), well-drained, acidic and with high Mg/Ca ratio (Marimon et al. 2001b, 2014).

Fig. 1
figure 1

Tropical forests (Monodominant-B. rubescens and mixed) in Southern Amazonia, Brazil, South America. Here we use the official IBGE map; however, for ecological purposes, we are considering the study area as Southern Amazonia, based on Marques et al. (2020)

Full size image

B. rubescens has a wide distribution in South America, having been recorded in the Amazonia and Atlantic Forest (Lima et al. 2017; ter Steege et al. 2019), but being monodominant only in stretches of forest in southern Amazonia, transition with the Cerrado biome (Marimon et al. 2014; Marques et al. 2020). The trees can reach 45 m in height and 90 cm in diameter, presenting a long-life cycle that can reach up to 700 years (Laurance et al. 2004). Wood is widely exploited both for structural and supporting buildings purposes and used by indigenous communities for making utensils (Marimon and Felfili 2001). Its fruits are also used as food by those communities and by wild fauna (Marimon et al. 2008).

Structural parameters and leaf morphological and anatomical traits estimates for B. rubescens

Structural parameters

Each forest (monodominant and mixed) is represented by an area of 0.6 ha, subdivided into 10 × 10 m (100 m2) sub-plots. In each forest, we randomly selected 15 subplots with adult individuals of B. rubescens (> 10 cm DBH). In each subplot, we collected a set of leaves from only one representative of the species, being the leaves mature, open, and exposed to the sun following a standardized protocol. Within the same subplots, we used the density, height, diameter, and above-ground biomass of the B. rubescens trees, provided by Marimon et al. (2014).

Leaf morphological and anatomical traits estimates

We selected eight leaves for each individual in each forest, five for morphological characterization and three leaves for anatomical determinations. Complete list of traits measured and their description can be found in Table S1. As a standardization criterion, we collected fully expanded leaves, exposed to full sunlight, and free of pathogens (i.e., leaf standardization protocol). We kept the plastic bags with the samples inside coolers during transport to the laboratory, which is very close to the studied areas, and on the same day, we processed the samples morphological. Morphologically, we measured leaf thickness (mm) using an electronic digital micrometer (± 0.001 mm) and using the caliper (± 0.001 mm) the petiole length (mm). With the precision balance (± 0.001 g), fresh and dry mass and we calculate the water mass content in the leaves (mg g−1); with LI-COR model LI-3100C we measure the leaf area, and calculate the specific leaf area (cm2 g−1) by dividing the leaf area by the dry mass. To obtain the dry mass, we packed the leaves in paper bags, and we put them in an oven with forced air circulation at 65 °C until constant weight (Fidalgo and Bononi 1984; Pérez-Harguindeguy et al. 2016).

For the anatomical traits, we selected three leaves for each individual that followed the same collection and storage protocol, performed the procedure for reversing herborization (Smith and Smith 1942) and store the samples in 70% alcohol (Johansen 1940). We made freehand cross cuts with a steel blade's aid, we used the method of clarifying the cuts with 2% sodium hypochlorite and stained the material with Astra blue and basic fuchsin. (Roeser 1962; Kraus et al. 1998). For the epidermal analysis, we performed the Franklin method (1945), the leaf portions were submitted to an aqueous solution (hydrogen peroxide 30 volumes and glacial acetic acid in a 1:1 ratio) and kept in an oven at 65 °C for 24 h. After this period, we wash the samples in distilled water and separate the epidermal surfaces with a brush and stain them with basic fuchsin (Roeser 1962). With the colored sections, we set up semi-permanent slides and recorded photomicrographs with the LAZ EZ 1.7.0 software from a microscope. (Leica® ICC50) attached to a computer.

We calculated stomatal density, for each individual, being the average of the number of stomata counted in the same fields of view registered previously. Then we estimated the average stomatal density, length, and width, measuring 25 stomatal complexes per individual. We measure the length of the guard cells (L, in µm), the width of the guard cell pair (W, in µm), the size of the stomata (S, estimated as S = L*W, according Franks et al. 2009, 2012), and the maximum area of the stomatal pores (amax, in µm2). The maximum area of the stomatal pore was calculated as amax = α*S, being α = 0,12 (Franks et al. 2009). We measure leaf traits with the aid of the program Anati Quant 2® UFV (Aguiar et al. 2007) and software ImageJ (Schneider et al. 2012; http://rsb.info.nih.gov/ij) (Table S1).

We applied an index based on the equation of the maximum and minimum medians to determine the phenotypic plasticity of the leaf traits (Valladares et al. 2006).

Statistical analysis

We compared the structural parameters and leaf functional traits between forests (monodominant vs. mixed) using the permutation t-test with RVAideMemoire package (Hervé 2021). Also, we generated a Pearson’s correlation matrix to evaluate the relationship between the leaf functional traits. We produced linear models to assess the relationship between leaf functional traits and variation between forests, and developed a principal component analysis (correlation PCA), to investigate how leaf functional traits are distributed among forests, using vegan (Oksanen et al. 2020) and psych (Revelle and Revelle 2020).

We also created a variance partition to understand which factors best explain the variation of each functional traits using different groupings of generalized linear mixed models (Zuur et al. 2009) and adjusting the separate models for each functional trait (Rosas et al. 2019). Based on the methodology used by these authors, we introduced the vegetation type and individuals as nested random factors and leaf functional traits as response variables for each model. We emphasize that the term individuals is independent of the forest type.

To understand whether leaf functional traits can predict the monodominance of B. rubescens, we used generalized linear models (Zuur et al. 2009). We used as variable responses in the models the density of individuals, height, diameter, and above-ground biomass. To choose the predictor variables, we based on other studies about functional traits in plants (Rossatto and Kolb 2010, 2012; Kearsley et al. 2017; Hall et al. 2020), but we also used the statistical methods. Based on a correlation matrix and in the literature, we selected the predictors and ran the first models. However, due to the autocorrelation between the variables, we chose to apply an automatic method the vifcor function of usdm package (Naimi and Araújo 2016). Therefore, to select the predictive traits and at the same time eliminate those that were multicollinear, we used the variance inflation factor (VIF) method, considering a VIF < 2 (Quinn and Keough 2002). Thus, of a total of 13 variables, only seven were maintained in the models. After running GLM, we selected the best models based on the value of AICc (Borcard et al. 2011) and calculated the average model with the R model.avg function using the dredge function, both from the MuMIn package (Barton and Barton 2015). We considered the best models those with ΔAICc < 2 (Burnham and Anderson 2002). All analyzes were performed using the R program, version 3.6.0 (R Development Core Team 2019), considering 5% of significance level.

Results

For all leaf functional traits here considered, only adaxial epidermis thickness (AET) and leaf dry matter content (LDM) did not present different values between the two forests (monodominant and mixed). The specific leaf area (SLA), maximum opening of the stomatal pore (MOS), petiole length (PEL), and stomatal size (STS) of B. rubescens were higher in the monodominant forest (Fig. 2). The stomatal density (STD) and thickness of the cuticle (ACT), palisade (PPT), spongy parenchyma (SPT), and thickness leaf (LET) were higher in the mixed forest. In addition, we note that all structural parameters (i.e., density, biomass, height, and DBH of trees) differ between forests and are greater in the monodominant forest (Fig. S4).

Fig. 2
figure 2

Leaf functional traits of B. rubescens in monodominant and mixed forests in Southern Amazonia. Boxplots represent medians and confidence intervals, and different lowercase letters indicate significant differences (Permutation t test, P < 0.05). SLA specific leaf area, LET leaf thickness, MOS maximum opening of the stomatal pore, PEL petiole length, ACT adaxial cuticle thickness, AET adaxial epidermis thickness, SPT spongy parenchyma thickness, PPT palisade parenchyma thickness, STD stomata density, STS stomata size, LDM leaf dry matter content

Full size image

Our linear models showed strong relationship between several leaf functional traits (Table S2), with some trade-offs, as the increase in specific leaf area (SLA) predicting reduction in leaf thickness (LET), (Fig. 3A) and adaxial epidermis (AET), (Fig. 3B), as well as in the palisade and spongy parenchyma (PPT and SPT), (Fig. 3C and D) and leaf dry matter content (LDM), (Fig. 3E). On the other hand, the larger the specific leaf area (SLA), the larger the stomata size (STS), (Fig. 3F). On the other hand, thicker leaves (LET) or thicker adaxial epidermis (AET) presented thicker palisade parenchyma (PPT), (Fig. 3G and J). Furthermore, the greater the leaves thickness, the smaller the stomata size (STS) and the maximum opening of the stomatal pore (MOS), (Fig. 3H and I). Also, the greater the adaxial epidermis thickness (AET), the greater the spongy parenchyma thickness (SPT), (Fig. 3K). Also, the increase in stomata density (STD) predicts decrease in stomata size (STS) and in maximum opening of the stomatal pore (MOS), (Fig. 3L and M).

Fig. 3
figure 3

Linear models between leaf functional traits of B. rubescens in monodominant and mixed forests in Southern Amazonia. MOS maximum opening of the stomatal pore, STS stomata size, STD stomata density, SPT spongy parenchyma thickness, PPT palisade parenchyma thickness, AET adaxial epidermis thickness, ACT adaxial cuticle thickness, PEL petiole length, LET leaf thickness, LDM leaf dry matter content, SLA specific leaf area

Full size image

The differences in leaf functional traits were strong enough to segregate individuals from both forests (Table S3). Such a condition can be seen in PC1 (Fig. S3), mainly determined by higher values of specific leaf area (SLA), petiole length (PEL), stomatal size (STS), and maximum opening of the stomatal pore (MOS) registered for individuals in the monodominant forest (Fig. S3). The PCA explained 64% of the total data variation in the first two axes.

The highest values of phenotypic plasticity were recorded for leaf functional traits linked to water saving, such as adaxial cuticle thickness (ACT) and stomatal pore, density (STD), and size (STS). The same was observed for leaf functional traits related to the acquisition of resources, such as specific leaf area (SLA) and petiole length (PEL). On the other hand, the lowest values were observed for leaf dry matter content (LDM) and leaf thickness (LET) (Fig. S2).

For most of the leaf functional traits here evaluated, the most significant variations were traits to differences between individuals (Fig. 4). The forest type explained 47–84% of the change in stomatal dimensions (MOS, STD, and STS) and petiole length (PEL), and the individual level explained 53–89% of the variation in the thickness of the spongy and palisade parenchymas (PPT and SPT), as well as in the cuticle and epidermis thickness (AET and ACT) (Fig. 4). The percentage of explanation at the individual level draws attention to the leaf dry matter content (LDM), which was higher than 80% (Fig. 4).

Fig. 4
figure 4

Partitioning of variance of the nested linear models of the leaf functional traits of B. rubescens in monodominant and mixed forests in Southern Amazonia. MOS maximum opening of the stomatal pore, STS stomata size, STD stomata density, SPT spongy parenchyma thickness, PPT palisade parenchyma thickness, AET adaxial epidermis thickness, ACT adaxial cuticle thickness, LDM leaf dry matter content, SLA specific leaf area, PEL petiole length, LET leaf thickness. Within means the residual error, all data were transformed (log-10) before analysis

Full size image

We observed that the petiole length (PEL) is an important functional trait because it can positively predict the density of individuals and the height of B. rubescens trees (Fig. 5 and Table S4, S5). On the other hand, adaxial cuticle thickness (ACT) influences negatively both the height and diameter of B. rubescens trees (Fig. 5 and Table S6). No functional leaf traits were effective on the variation in biomass (Fig. 5 and Table S7). The complete models explained between 41 and 53% of the change in monodominance and population structure of B. rubescens (adjusted R2 values).

Fig. 5
figure 5

Correlates of structural parameters of B. rubescens populations in monodominant and mixed forests in Southern Amazonia. Variables that were not captured as good predictors in the models were adjusted toward the zero-point line. DBH diameter at breast height, LET leaf thickness, PEL petiole length, SPT spongy parenchyma thickness, STD stomata density, ACT adaxial cuticle thickness, LDM leaf dry matter content

Full size image

Discussion

In this first study evaluating the relationship between leaf functional traits and the monodominance of B. rubescens, we found strong differences between the populations of a monodominant and a mixed forest. The set of functional traits of individuals of B. rubescens in the mixed forest reveals that such population invests in conservative strategies, in contrast with acquisitive strategies of the monodominant population. Such a condition indicates that individuals from both forests may have suffered different selective pressures resulting in bidirectional segregation in the leaf economics spectrum. Differences between individuals and forests types were the factors that best explained the variation of leaf functional traits. Our results also reveal that the petiole length and the adaxial cuticle thickness could be associated with B. rubescens monodominance. Explanations for these results are detailed below.

Trade-offs in leaf functional traits and implications for B. rubescens monodominance

The variation in leaf traits showed a functional divergence in the ecological strategies of B. rubescens. Individuals in the monodominant forest presented strategies linked to “acquisition and use of resources”. In contrast, individuals in the mixed forest showed the opposite strategies linked to “conservation of resources”, constituting extreme patterns of the leaf economics spectrum (Wright et al. 2004).

B. rubescens individuals in the monodominant forest showed higher values of specific leaf area, stomata size, maximum opening of the stomatal pore and petiole length, which are linked to a higher photosynthetic rate, primary productivity and, consequently, greater growth of individuals (Poorter and Bongers 2006; Ogburn and Edwards 2010). These functional traits allow maximize light capture (Takenaka 1994), being advantageous during a canopy opening (gaps) or an acquisition strategy (e.g., higher values of specific leaf area and length of the petiole). This can increase the efficiency in capturing the resource and, consequently, allow B. rubescens to overlap its density above other species, occupying a large part of the space (~ 90%) and maintaining its monodominant pattern (Marimon et al. 2014).

In contrast, individuals of B. rubescens in the mixed forest showed higher values of the adaxial cuticle thickness, palisade and spongy parenchymas thickness and also leaf thickness, normally linked to protection and support mechanisms (conservative strategy), and can also facilitate the uptake of water and their maintenance in the tissues (Fahn and Cutler 1992; Gratani et al. 2006). This condition increases the efficiency of these individuals in the conservation of nutrients and water use (Pallardy 1981; De Micco and Aronne 2012), indicating higher resistance to drought (Franco 2002; Goldstein et al. 2008) and better competition for water. Also, it reduces leaf damage caused by herbivores, excessive sunlight and high temperatures (Turner 1994; Rozendaal et al. 2006; Rossatto and Kolb 2010; Araújo et al. 2021b), ensuring higher integrity of the leaf mesophyll. On the other hand, these traits represent lower productivity rate and, in turn, slower growth (Reich 2014), which may be associated with the lower population density and lower dominance of B. rubescens in the mixed forest.

This network of trade-offs between leaf functional traits can be useful to understand the differences in the population density of B. rubescens between monodominant and mixed forests. We argue that the highest values of specific leaf area, petiole length, stomata size, and maximum opening of the stomatal pore are important to ensure competitive advantage in the search for light, water, and space (Aiba et al. 2020; Araújo et al. 2021b). Resource acquisition strategies have also been suggested as the main factor contributing to Gilbertiodendron dewevrei monodominance in Africa (Hall et al. 2020). Therefore, leaf functional traits linked to acquisitive strategies (Donovan et al. 2011) can be a proxy to help explain the monodominance of B. rubescens in Southern Amazonia. Thus, we believe that new studies should evaluate whether monodominant species present convergent ecological strategies, as well as the efficiency and safety in physiological mechanisms to predict the vulnerability and resistance of monodominant species in a hotter and drier climate.

Our findings confirm that populations of B. rubescens in adjacent monodominant and forests, present ample phenotypic plasticity in leaves functional traits. This ecological condition can provide higher adaptive capacity to different environmental and climatic patterns (Silveira et al. 2013; Lima et al. 2017), conditioning the species persistence (Chevin and Lande 2010; Franks et al. 2014; Araújo et al. 2021b) and possible increases in their populations and geographic distribution over time and space.

Extreme climatic events, such as increased temperature and drought in Southern Amazonia (Meehl and Tebaldi 2004; Collins et al. 2013; Rifai et al. 2018), can cause severe changes in species abundance, composition, and distribution (Walther et al. 2002; Menzel et al. 2006), consequently increasing the tree mortality (Phillips et al. 2010). This is particularly important for B. rubescens in the monodominant forest, as the species have leaf functional traits that are less tolerant to increases in intensity and frequency of drought events and increase in temperature, and therefore may be negatively affected if future climate changes (Marimon et al. 2020). However, theoretically, phenotypic plasticity can ensure its persistence as a monodominant species (Vitasse et al. 2010), through potential changes in the leaf economics spectrum. In this way, revealing the individual responses of this species will make it possible to launch a perspective of future risk management for B. rubescens populations in both forests, especially if we consider the economic importance of the wood of this species and also its use by local indigenous communities (Marimon and Felfili 2001).

Possibly the difference in community structure of both forests and subtle variations in soil water availability (Marimon et al. 2014; Elias et al. 2018; Marimon-Junior et al. 2019) may have generated alterations in leaf tissues, where individuals of B. rubescens have a wide range variation at the individual level. This reinforces the species' ability to adapt to different ecosystems, such as the Amazonia and the Atlantic Forest (Silveira et al. 2013; Lima et al. 2017), as well as being a strong competitor in Southern Amazonia, to the point of becoming monodominant in several patches.

Leaf functional traits can predict B. rubescens monodominance

We found that the petiole length is an essential predictor of the increase in the density of individuals and the height of B. rubescens trees, these results corroborate our findings in Fig. S4. This trait maximizes the absorption of photosynthetically active radiation (Poorter and Bongers 2006), reducing leaf clumping and overlap (Takenaka 1994), which allows the plant to be more efficient in capturing light in shaded environments (Weijschedé et al. 2008). Longer petiole length improves leaf distribution and orientation (King and Maindonald 1999) and can promote higher growth rates and density of individuals, as in the case of B. rubescens. However, we cannot rule out the importance of the specific leaf area and the size of the stomata. The literature suggests that longer petioles support larger leaves (Reich 2014) that enable an increase in photosynthetic rate and primary productivity, reflecting rapid growth (Westoby 1998; Wright et al. 2004). In this way, these morphological traits acting together can further increase the efficiency in the absorption of light (King and Maindonald 1999). Also, a larger specific leaf area with larger stomatal sizes can facilitate CO2 absorption (Beaulieu et al. 2008; Rossatto et al. 2009) and, consequently, assist in higher investments in the height and density of individuals, as we can see in Fig. S4.

The negative relationship between the thickness of adaxial cuticle with the height and diameter of the individuals was not expected in this study. This trait increases the capacity of plants to retain water (Rossatto and Kolb 2010; Araújo et al. 2021b), preventing the rapid dehydration of tissues and the loss of excessive water to the atmosphere, which can allow B. rubescens to increase the water use efficiency. It is also commonly observed that plants under conditions of water deficit and intense solar radiation presented higher cuticle thickness (Shepherd and Griffiths 2006; Araújo et al. 2021b), which may be the case of B. rubescens individuals in the mixed forest, which has a more open canopy (Marimon et al. 2008). In contrast, the adaxial cuticle is thin in the forest where the species is monodominant, which allows higher photosynthetic efficiency and tree growth (Westoby 1998; Koch et al. 2009).

The cuticle has a fundamental role in maintaining the water status of the plant (Larcher 1995), especially for B. rubescens, which can reach up to 45 m in height. As the height of the trees increases the hydraulic vulnerability also increases and can lead trees to death (Klein et al. 2018). In this case, conserving more water in the leaves may be essential for B. rubescens to maintain the growth and survival (Ambrose et al. 2009). Thus, the higher cuticle thickness of individuals in the mixed forest may be a strategy of the species to better compete with the others for water. Perhaps, this contributed to the species becoming monodominant in the past; however, reducing the thickness of the cuticle according to the increase in dominance. If climate change does not occur B. rubescens can increase the number of individuals and become denser in forests where it is still present with few individuals, mainly driven by its phenotypic plasticity in leaf functional traits, which can be an essential component for the persistence of the species over time.

Our findings demonstrate that individuals of B. rubescens have different functional strategies to establish and maintain their population in monodominant and mixed forests, proving to be a strong competitor for resources, such as water and light, through the variation in their leaf functional traits. Also, we revealed that the integration of some leaf functional traits, such as the length of the petiole and the thickness of the adaxial cuticle, could be an important part of the set of conditions driving the monodominance of B. rubescens in Southern Amazonia.