Sympatric and phenotypic variation between two representative species of the Atlantic Forest: an analysis of the hydraulic system

,


Introduction
Atlantic Forest has one of the most significant species diversities related to distinct heterogeneous landscapes throughout its extensive east coast distribution of the South American continent (Oliveira-Filho & Fontes 2000;Scarano 2002).This biome is formed by a complex of ecosystems that offer goods and services to the largest cities in Brazil (Rezende et al. 2018;SOSMA & INPE 2019).However, the remnants that still resist must be better understood for their conservation and adoption of measures to restore degraded areas due to reducing their original area.The Atlantic Forest in Bahia and Rio de Janeiro has a high degree of richness and endemism of plant species.Furthermore, they have macroclimatic and rainfall differences imposed by spatial separation.
Several researchers seek to understand the different mechanisms by which plants adjust the anatomy of their xylem tissues to meet various macro and microclimatic characteristics (Hacke et al. 2006;Zanne et al. 2010;Trueba et al. 2017;Rosas et al. 2019).Tropical forests are the most diverse on the planet, and a wide range of tree strategies have evolved to promote the survival and distribution of these species.Understanding which characteristics contribute to the evolutionary success of plants in such diverse forests is still a challenge (O'Brien 2017).Traits such as vessel length and diameter can clarify the relationship of efficiency and safety in water use (Tyree & Zimmermann, 2002).Thus, the comparative wood anatomy allows a better understanding of the xylem structure-function relationships and hydraulic strategies against environmental variations.
Wood anatomy gathers data of proven relevance for recognizing taxonomic groups and contributes to solving problems about the systematic position of several taxa (ex.: Carlquist & Zona 1988;Baas et al. 2004;Lens et al. 2007).Thus, the wood anatomy has been used successfully in the botanical identification of species of wood-producing trees and, in particular, commercial material devoid of the vegetative part (Coradin 2002).In addition, ecological wood anatomy includes studying environmental influences on structural variations of species, showing that environmental factors directly influence the anatomical and morphological structure (Baas 1973).The wood anatomy also acts on phenotypic variation and plasticity and, in the long run, performs in the natural selection of species.Wherefore, the relationships between the structure of the wood and the environmental characteristics have been widely studied (Zuidervaart & Anderson 2016).Thus, studies of ecological anatomy that present the hydraulic system's adaptive strategies and the plants responses to the water stress open a new field in the ecological wood anatomy, mainly in the tropical forests (Costa et al. 2020;Simioni et al. 2020).
The ability of a tree or even a population to undergo morphological or physiological variations in the face of changes in environmental conditions is called phenotypic plasticity (Valladares et al. 2006;Gianoli & Valladares, 2012).The phenotypic plasticity is the fastest mechanism for responding to environmental change and plays a key role in the persistence of trees in the face of climate change (Aubin et al. 2016).Simultaneously, the gradient of morphological or physiological variation between related individuals resulting from an adaptive process is called phenotypic variability (Gratani, 2014).
Studies in forest environments about adaptation conditions by trees concerning morphological and functional characteristics enable a better understanding of ecosystem organization (e.g., Missio et al. 2017).In this sense, the hydraulic system may reflect structural adjustments as means of adaptation by native species to various phytophysiognomies in the Atlantic Forest (e.g., Marques et al. 2012;Campbell et al. 2016;2018;Macedo et al. 2019).Then, the structure and function of wood, mainly focused on tree hydraulics, is an essential field in biology because it allows a better comprehension of forest ecology and evolutionary analysis of plants (Baldin et al. 2016;Hacke 2015).
Bathysa C. Presl is widely distributed in the Northeast and Southeast regions of Brazil.This genus is represented by about 15 species, trees that reach 15 m, or shrubs.The species are exclusively Neotropical, occurring in hillside forest formations at 600-800 (-1,100) meters.In Brazil, seven species occur, all exclusive to the Atlantic Forest (Germano-Filho 1999).Some species of this genus are considered relevant by the Importance Value Index (IVI) for studied remnants indicating great representativeness of these species (Guedes-Bruni 1998).This genus has species known as "falsas quinas" that encompass all plants that contain barks with a bitter taste and a red, yellow, or gray color.According to Correa (1984), the bark provides tonics, which are often used in treating anemias, cachexia, fevers, hookworm, convalescent states, and others, in addition, are used as dyes.Although the genus has a high representation in Atlantic Forest remnants, little is known about the ecology of its species.
Considering the environmental variables that occur in a given location throughout the year, as well as the latitude and differences between locations, the present work investigates the hypothesis that the intrinsic environmental variables of different environments, for example, latitude and rainfall, can result in a structural adjustment in the secondary xylem of Bathysa cuspidata and Bathysa nicholsonii in two different Atlantic Forest remnants.One of the studied remnants is in the Corredor Central of the Atlantic Forest in the Northeast Region of Brazil.The other is in Corredor da Serra do Mar of the Atlantic Forest in the Southeast Region.Therefore, the present work aimed to answer the following questions: (i) Are there anatomical traits that highlight adaptive strategies of the studied species of Bathysa in the different studied forest remnants?And, if so, (i) What can be inferred from the hydraulic system of the four analyzed tree groups?(Increment Borer, SUUNTO, USA) at approximately 1.30 m above the ground (DBH).For arboreous individuals, samples were taken using a Pressler probe at approximately 1.30 m above the ground, whereas for shrubby individuals, samples were taken as stem disks using a machete.Part of the wood samples and permanent slides were deposited and registered in the wood collection of the Universidade Estadual do Norte Fluminense (HUENFw), whereas the remainder was used for anatomical identification and exsiccates were deposited at the Herbarium of Universidade Estadual do Norte Fluminense (HUENF).Wood samples were obtained from the sapwood, according to Coradin & Muniz (1992).The samples were sectioned with a sliding microtome (SM2010 R, Leica, Germany) in transverse and longitudinal (radial and tangential) planes, at an average thickness of 20 µm.Sectioned material was dehydrated in ethanol and stained with Astra Blue and hydro-alcoholic Safranin (Burger & Richter 1991).Wood fragments were subjected to dissociation by Franklin solution (1945), washed and stained with 1 % aqueous safranin.Anatomical measurements were performed using a light microscope (Axioplan Zeiss, Oberkochen, Germany); measurement techniques and terminology followed IAWA guidelines (IAWA committee 1989).To establish the relationship between the environment and the wood characters, vulnerability (VI = tangential diameter of vessels/vessel frequency) and mesomorphy (MI = VI x length of vessels) indices were calculated, following Carlquist (2001).Samples for scanning electron microscopy were subjected to critical point drying with CO 2 (CPD 030, Baltec, Heerbrugg, Switzerland), adhered to stubs using carbon adhesive tape and covered with a layer of approximately 20 nm of gold (SCD 050, Baltec, Heerbrugg, Switzerland).Images were obtained using a ZEISS-DSEM 962 scanning electron microscope (Oberkochen, Germany) at a voltage of 25 kV.

Materials and methods
Quantitative data were subjected to the Kolmogorov-Smirnov normality test (Hollander & Wolfe 1973).Nonparametric tests were used to verify the significance of the differences between species in the two forest remnants in Bahia and Rio de Janeiro, respectively, through multiple comparisons using the Kruskal-Wallis test (Kruskal & Wallis 1952).Analyses were performed with Statistica 7 software.The principal component analysis (PCA) was performed with the PC-ORD version 6.0 software, using a matrix with anatomical characteristics in columns and sampling units (individuals / species) in rows, to verify a possible pattern within plant traits describing plant strategies in the different environments.

Results
Anatomical traits of both Bathysa cuspidata and B. nicholsonii are shown in Table 2 according to sampling site, REBIO-RJ, and RPPNSB-BA.In general, the qualitative wood anatomy traits proved to be similar in both species regardless of the studied sites, showing: diffuse porosity; solitary vessels or in multiple radial arrangements of 2 to 4 elements; vessels with simple perforation plates located on inclined end walls and, sometimes, on side walls, and with alternate and ornamented intervessel pits; septate fibretracheids; axial parenchyma absent or extremely rare; and heterogeneous uniseriate and/or multiseriate rays formed by procumbent, upright and square cells (Fig. 1).The distinct traits for B. cuspidata (Figs.S1, S3) are indistinct or absent growth rings; lower vessel frequency; fibre-tracheids with thick walls; and multiseriate rays of 3 to 4 cells wide.The distinct traits for B. nicholsonii (Figs.S2, S4) are presence of poorly distinct growth rings, marked by the greater thickening and radial flattening of the fibres of the latewood; higher vessel frequency; uniseriate, biseriate and triseriate rays, sometimes fused; and the presence of tyloses.
Quantitative comparative analysis of wood anatomy in relation to the study areas (Tab.2) revealed that vessel frequency and vessel wall thickness are greater in RPPNSB-BA for both species, however, tangential diameter of vessels and ray width did not differ between sampling areas for both species.In addition, both species had lower values for the vulnerability index, longer vessels and fibres, and larger intervessel pits at REBIO-RJ.The principal component analysis of the quantitative data explained 80% of the total variation in wood anatomical traits (Fig. 2) and ordered the four analyzed species separately.Axis 1 explained 56 % of  the total variation and ordered the RPPNSB-BA samples on the positive side of the axis and the REBIO-RJ samples on the negative side, regardless of species.The traits that most influenced this segregation were vessel frequency and diameter.Axis 2 explained 24 % of the variation and ordered the two species separately, with B. cuspidata negatively positioned on this axis and B. nicholsonii (RJ) positively positioned.The traits that most influenced this separation were fibre diameter and vessel length.

Discussion
Bathysa cuspidata and B. nicholsonii wood was described and compared structurally for two different areas of the Atlantic Forest (Rio de Janeiro and Bahia), with traits being identified with value for the systematics of the genus and for species ecology in the forest environment.Our results suggest that some traits that can be considered taxonomically useful at the level of the genus for Bathysa, such as diffuse porosity or pores in multiple radials, simple perforation plate, absent axial parenchyma and septate fibre-tracheids with simple pits, which agree with the wood characteristics of Group II, according to Jansen et al. (2002).
Wood anatomical traits are essential in studies of ecological structural variation due to microclimatic changes (e.g., Carlquist 1977;Campbell et al. 2018).In the present study, the wood anatomical traits showed different strategies of the species in the different remnants.At REBIO-RJ species showed lower frequency, higher vessel length, and vessel diameter; longer fiber and ticker fibre-tracheids walls; and higher levels of vulnerability.In contrast, species from RPPNSB-BA showed higher frequency, and lower length, and diameter of vessels; lower length and thickness of walls of fibre-tracheids; and lower levels of vulnerability.Such traits may be associated with microclimate differences between the areas, resulting in adaptations to the hydraulic system of both species.These findings corroborate other studies of intraspecific variation in wood involving morphological, anatomical, and physiological traits.For example, Lens et al. (2004) analyzed ecological trends in tropical forests and described the importance of the amount of precipitation in influencing the continuous traits of wood.Melo Júnior et al. (2018) analyzed species under different climate and soil types and found differences in xylem structure favoring safety and water conduction efficiency, showing how hydraulic processes can influence plants.
Qualitatively, the studied species were very similar.Quantitatively, however, the species showed differences related to the distinct environmental pressures of the two analyzed sites, although not characterized as a set of convergent responses.Both species had larger vessel diameters and lower vessel frequencies in REBIO-RJ compared to RPPNSB-BA.These traits reflect a tradeoff between efficiency and safety in water transport, corroborating to Carlquist (2001);Melo Junior et al. (2018) and Simioni et al. (2020).
The environmental variables occur at a given location throughout the year, latitude, and between sites.For example, annual rainfall is greater in REBIO RJ.Individuals in the REBIO-RJ have greater vessel length and diameter, slightly larger intervascular pits than individuals in RPPNSB- BA, suggesting that the vessels have smaller diameters and lengths in drier environments (Luchi 2004;Lima et al. 2009).Furthermore, longer vessels offer less resistance in the cell wall, favoring higher resistance and greater conductive efficiency (Sperry et al. 2005).At RPPNSB-BA, both species had a majority of relatively long vessels, ranging from 1335.9 µm to 846.71 µm.They both also had higher vessel frequencies and smaller vessel diameters, shorter vessel lengths, lower vulnerability indices, and greater thickness of fibre-tracheids, which together represent a strategy that provides greater safety in water conduction.These differences in anatomical wood traits can be attributed to the structural adjustments of individuals in the different areas (Lens et al. 2004;Melo Júnior et al. 2018;Costa et al. 2020).
Although fibres or fibre-tracheids vary in length and width in the wood, the sampling of the present study was standardized for the same collection time, tree diameter, and the distance from bark (sapwood), allowing comparisons to be made of the length of these cells between the four groups.In this sense, fibre-tracheids were shown to be longer and wider at REBIO-RJ, corroborating previous studies that related greater fibre length to environments with greater precipitation (Carlquist 1975;Barajas-Morales 1985;Noshiro & Baas 2000).Mesomorphy indices above 200 indicate species from mesic environments, while values below 200 indicate plants from xeric environments (Carlquist 1977).In the present study, both species showed mesomorphy indices well above 200, corroborating with other results obtained for Rubiaceae in mesic environments of the Atlantic Forest (e.g., Callado et al. 2003;Marques et al. 2015;Campbell et al. 2016;Melo Júnior et al. 2017).
Based on the ecological anatomy of Carlquist, Olson (2020) established the vulnerability index by relating vessel frequency to vessel diameter because they are traits that demonstrate the risk of embolism.That is, the lower the index, the safer the conductive element and, consequently, the lower its hydraulic conductivity.In the present study, the vulnerability indices were near 1, and only lower for B. nicholsonii at RPPNSB-BA, which may reflect an adjustment to the local climatic conditions with less annual rainfall.This adjustment characterizes greater safety in the hydraulic system by reducing the risk of embolism and cavitation and corroborates other studies with species from different environments (Longui et al. 2009;Campbell et al. 2018, Melo Junior et al. 2018).
The differences in microclimatic conditions between the two areas favored structural adjustments to hydraulic traits, such as vessel frequency, length, and diameter.These traits were also shown to be greater in REBIO-RJ, representing a strategy that favors transport efficiency.This improved efficiency is due to the lumen having a greater volume and capacity for water transport, which contrasts with the conservative strategy of the species at RPPNSB-BA, with higher density, smaller diameter, and shorter length for the vessels.These are traits that provide safe transport and less risk of cavitation and have been shown to be wood anatomical traits that are influenced by the environment (Baas et al. 2004;Poorter et al. 2010;Marques et al. 2015;Campbell et al. 2016Campbell et al. , 2018;;Melo Junior et al. 2018;Costa et al. 2020;Vasconcelos & Callado 2020).
Both studied areas varied in precipitation, but the REBIO-RJ shows higher rainfall, although geographically distant.As a result, higher frequency, shorter length, and smaller vessel diameter providing both species greater safety in transport at RPPNSB-BA.In contrast, at REBIO-RJ, both species have vessels with larger diameters and longer lengths, which is more efficient for transport.Thus, the traits found confer different strategies for the same species in different environments, suggesting responses to different macro-and micro-environmental pressures.
Although each species hydraulic system shows acclimatization variations to different sites of occurrence, the structural differences in wood ensured the structural adjustments between the species studied in each of these sites.These results corroborate the relevance of the wood anatomy studies as a subsidy for a better definition of morphologically similar taxa and show structural aspects aimed at survival in the Atlantic Forest.The combination of anatomical and environmental traits is often neglected.Here we showed that this combination could be an essential tool to understanding tropical biodiversity.

Four
individuals from Bathysa cuspidata (A.St. Hil.) Hook.F. and Bathysa nicholsonii K. Schum of approximately the same height and diameter breast height were (DBH) selected from interior areas the Dense Ombrophilous Rainforest in two forest remnants (Tab. 1) Reserva Particular do Patrimônio Natural Serra Bonita, Camacan, Bahia (RPPNSB-BA) and Reserva Biológica do Tinguá, Nova Iguaçu, Rio de Janeiro (REBIO -RJ).Wood samples were obtained non-destructively using a Pressler probe

Figure 1 .
Figure 1.Comparison of wood anatomy of co-occurring species at RPPNSB-BA and REBIO-RJ.A, E, I: sections in three planes of Bathysa cuspidata RPPNSB-BA, note smaller diameter and greater frequency of vessels (arrow, triangle).B, F, J: sections in three planes of Bathysa cuspidata at REBIO-RJ, note greater diameter and lesser density of vessels (arrow, star).C, G, L: sections in three planes of Bathysa nicholsonii at RPPNSB-BA, note greater vessel density and smaller width of rays (arrow, star).D, H, M: sections in three planes of Bathysa nicholsonii at REBIO-RJ, note lower vessel density and greater width of rays (arrow, star).Scale A-M = 100 µm.

Table 1 .
Coordinate, climate and precipitation of study sites.

Table 2 .
Average and standard deviation of the quantitative parameters of the wood of the four Bathysa species studied.