Xylem Functional Anatomy of Pure-Species and Interspecific Hybrid Clones of Eucalyptus Differing in Drought Resistance

Abstract

Climate extremes threaten the resilience of Eucalyptus plantations, yet hybridization with drought-tolerant species may enhance stress tolerance. This study analyzed xylem anatomical and functional drought responses in commercial Eucalyptus grandis (GG) clones and hybrids: E. grandis × camaldulensis (GC), E. grandis × tereticornis (GT), and E. grandis × urophylla (GU1, GU2). We evaluated vessel traits (water transport), fibers (mechanical support), and wood density (D) in stems and branches. Theoretical stem hydraulic conductivity (k_Stheo), vessel lumen fraction (F), vessel composition (S), and associations with previous hydraulic and growth data were assessed. While general drought responses occurred, GC had the most distinct xylem profile. This may explain it having the highest performance in different irrigation conditions. Red gum hybrids (GC, GT) maintained k_Stheo under drought, with stable F and a narrower vessel size, especially in branches. Conversely, GG and GU2 reduced F and S; and stem k_Stheo declined for a similar F in these clones, indicating vascular reconfiguration aligning the stem with the branch xylem. Almost all clones increased D under drought in any organ, with the highest increase in red gum hybrids. These results reveal diverse anatomical adjustments to drought among clones, partially explaining their growth responses.

1. Introduction

Eucalyptus species cover over 20 million hectares worldwide due to their high productivity and adaptability [1]. In the subtropical and warm-temperate regions of South America, Eucalyptus grandis is widely cultivated for its superior pulp and timber yields. However, an increasing occurrence of droughts, frosts, and heatwaves [2,3] threatens plantation resilience, particularly during early growth stages. Consequently, breeding programs prioritize genotypes exhibiting traits linked to abiotic stress tolerance, such as efficient water use, sustained hydraulic function, and stable growth under adverse conditions. Within this context, hybridization with drought-tolerant red gums (E. camaldulensis, E. tereticornis) and E. urophylla is commonly used to enhance adaptability and productivity.

Balancing xylem hydraulic efficiency and safety is a fundamental physiological strategy that allows plants to maintain function, survive, and remain productive under drought conditions [4,5,6]. During water stress, increased xylem tension frequently induces cavitation and embolism, which reduce hydraulic conductivity and limit gas exchange and carbon assimilation. The potentially severe consequences of this process, including hydraulic failure and plant mortality, underscore the importance of understanding the trade-off between efficiency and safety in xylem function [4,7]. In angiosperms, susceptibility to hydraulic dysfunction is closely linked to anatomical features such as vessel diameter, vessel grouping, and pit membrane structure, which collectively influence both conductivity and resistance to cavitation [8,9]. Vessel wall reinforcement also plays a critical role in preserving conduit integrity under negative pressure [8], while fibers primarily provide mechanical support, although they may also contribute to sustaining hydraulic function under stress [10,11,12].

Overall, vessel wall reinforcement and thicker fiber walls contribute to increase wood density, a trait generally associated with enhanced xylem safety under negative pressure conditions [8]. In response to drought, many woody species exhibit anatomical adjustments that favor a safer xylem structure—such as a reduced vessel diameter or increased cell wall thickness—often at the cost of lower transport efficiency, typically quantified as specific hydraulic conductivity (k_S) [13,14,15,16]. This trade-off between hydraulic safety and efficiency is a central principle in plant hydraulic architecture [17], and elucidating its underlying mechanisms is essential for understanding species-specific responses to water stress [18,19,20]. In this context, structural traits such as the lumen fraction (F) and vessel composition (S) are particularly informative, as they influence both the water transport capacity and vulnerability to embolism. These integrative traits offer valuable insights into how plants coordinate hydraulic function in response to environmental variability [21].

The Eucalyptus genus, comprising over 700 species, exhibits a wide range of hydraulic strategies to cope with drought. In a comparative study of 31 species along an aridity gradient, Pfautsch et al. [22] reported that species native to drier environments tend to develop narrower vessels and denser wood. These traits are typically associated with lower theoretical specific hydraulic conductivity (i.e., maximum k_S based on vessel dimensions), reflecting a trade-off that favors hydraulic safety over transport efficiency. Such characteristics are largely genetically determined and exhibit limited plasticity. For example, Eucalyptus grandis displays high stem ks values due to its large vessel diameters and high vessel area, features generally linked to a lower wood density and greater conductive capacity. However, under water-limited conditions, E. grandis adopts a water-conservative strategy, primarily through tight stomatal regulation and the maintenance of stable water potentials [23,24,25]. In contrast, red gums such as E. camaldulensis and E. tereticornis follow a water-spending strategy, maintaining higher rates of gas exchange and growth even under drought stress.

The hydraulic strategy employed by red gum species involves a greater degree of hydraulic risk. While relatively resistant to cavitation, these species often experience significant reductions in specific conductivity due to limited stomatal regulation and anisohydric behavior [26]. In branch wood, E. camaldulensis develops wider vessels than E. grandis [27], a trait generally linked to lower cavitation resistance via increased pit area, a pattern observed across several Eucalyptus species [28]. However, fast-growing commercial genotypes may deviate from this expected safety–efficiency trade-off. For instance, although it possesses wider vessels, E. camaldulensis has been shown to be less vulnerable to cavitation than E. grandis, suggesting that organ-specific deviations from the expected trade-off may occur. Similarly, E. tereticornis combines wide vessels with high cavitation resistance [28]. In contrast, E. urophylla, a species native to humid tropical regions, exhibits steep conductivity losses under drought, triggering strong stomatal closure and reduced gas exchange [29]. This response reflects xylem vulnerability, comparable to that observed in the fast-growing E. grandis.

Interspecific hybrids can exhibit anatomical and physiological traits that are intermediate or even superior to those of their parental species [30]. In E. grandis and its hybrids, hydraulic performance shows considerable genotypic variation, which influences their capacity to tolerate drought. In South Africa, hybrid clones differed from E. grandis in growth efficiency and canopy dieback under severe drought, despite displaying only moderate differences in vulnerability to cavitation [31]. Similarly, a previous study conducted in Uruguay reported high in situ levels of cavitation in Eucalyptus hybrids with red gum parentage, with percent loss of conductivity (PLC) reaching up to 85%, despite their presumed lower vulnerability. This unexpected behavior was attributed to weak stomatal control, combined with osmotic and elastic adjustments that allowed continued growth under both well-watered and drought conditions [32]. Hybrids such as E. grandis × E. camaldulensis and E. grandis × E. urophylla may thus adopt high-risk hydraulic strategies under water deficit [33]. However, it has been shown that among hybrids sharing E. urophylla parentage, drought vulnerability can vary considerably, despite similar wood anatomical features [34], thus suggesting that hybridization can lead to rather unpredictable functional combinations. Moreover, although drought-induced xylem downsizing and increased tissue density have been observed in E. grandis, the functional significance of these responses in hybrid genotypes remains unclear [35].

Most studies on xylem hydraulics have focused on a single organ—leaves, branches, stems, or roots—despite substantial anatomical and functional variation along the plant axis. In addition to vessel tapering that facilitates vertical water transport, some Eucalyptus species, including E. grandis, exhibit hydraulic decoupling between stems and branches as an adaptive response to drought. In E. grandis, stems typically contain large-diameter vessels and exhibit high specific conductivity, supporting rapid water transport and growth. In contrast, the vessel diameter in branches declines markedly along the vertical axis, from approximately 250 µm at the base to 20 µm at the distant ends [36]. While narrower vessels generally confer greater resistance to cavitation [37], E. grandis branches possess anatomical features such as larger pits [27] and thinner pit membranes [36], which enhance hydraulic conductivity but increase vulnerability to embolism. As a result, E. grandis often sheds branches during severe drought, a response interpreted as a means of protecting the stem from embolism propagation. This branch abscission strategy was not observed in E. globulus or E. camaldulensis, which exhibited greater resistance to cavitation at the branch level [27]. These organ-specific differences in vulnerability highlight the importance of assessing the whole-plant hydraulic architecture when evaluating drought responses and tolerance strategies in Eucalyptus [38].

Given its dual role in mechanical support and water transport, the wood-specific density is another key trait linked to the hydraulic function and drought response. In Eucalyptus species, wood density ranges from 400 to 1000 kg m⁻³ and varies across plant organs [21,39], potentially influencing hydraulic efficiency and vulnerability. Although differences in wood density between stems and branches are commonly reported, their direction and magnitude remain inconsistent across studies [40,41,42,43], suggesting complex underlying mechanisms that may depend on species-specific strategies, developmental stage, or environmental condition. Overall, the relationships among clonal drought resistance, xylem anatomical traits, wood density, and the coordination of hydraulic properties across different organs remain poorly understood [44,45,46].

The objective of this study was to investigate the anatomical and functional responses of xylem to drought in commercial clones of Eucalyptus grandis and interspecific hybrids of E. grandis × camaldulensis, E. grandis × tereticornis, and E. grandis × urophylla grown under controlled greenhouse conditions. Despite the ecological and commercial importance of these genotypes, there is a significant knowledge gap regarding how such anatomical and physiological traits diverge between E. grandis and its hybrids, particularly in the context of drought stress. In previous work, we characterized the different physiological strategies employed by these genotypes to cope with water deficit [32]. Here, we advance on the wood anatomical traits that may underlie or help explain the differential drought responses previously observed.

We hypothesized that clone-specific differences in the size and arrangement of conductive (vessels) and supportive (fibers) cells under drought would contribute to variations in hydraulic performance and wood density in stems and branches, thereby affecting water supply to the foliage and plant growth under varying water availability regimes. These anatomical differences may reflect distinct drought-response strategies among Eucalyptus grandis and its interspecific hybrids, with direct implications for water transport efficiency and growth under contrasting moisture regimes. Understanding hydraulic differentiation and xylem plasticity among genotypes is essential for elucidating drought adaptation and guiding the selection of resilient genotypes for water-limited environments.

2. Materials and Methods

2.1. Plant Material and Growth Conditions

This study was conducted at the Department of Plant Biology, Centro Universitario de Tacuarembó, Uruguay (31.73° S, 55.97° W), using Eucalyptus clones widely cultivated in Uruguay, South America. The plant material consisted of rooted cuttings of Eucalyptus grandis (GG), E. grandis × camaldulensis (GC), E. grandis × E. tereticornis (GT), and E. grandis × E. urophylla (GU1 and GU2), propagated in a clonal nursery operated by Lumin Forestry Company. Cuttings were initially grown in a controlled environment (20–22 °C; 90%–95% relative humidity) in 120 cm³ tubes filled with inert substrate and periodically watered with a Biorend^® nutrient solution (10 cm³ L⁻¹) (Bioagro S.A., Tierra del Fuego, Chile). After five months, plants were transplanted into 3 L pots containing Carolina Soils^® substrate (Carolina Soil do Brasil Ltda., Santa Cruz do Sul, RS, Brasil). At six months of age, the plants were transferred into 18 L plastic pots and moved to the experimental greenhouse for acclimation to its specific environmental conditions.

2.2. Experimental Design and Treatments

A split-plot experimental design was employed to evaluate drought responses across two cycles. The main plots consisted of two watering treatments—water restriction (drought) and well-watered control—both applied uniformly across the entire plot. Within each main plot, subplots consisted of five Eucalyptus clones (GG, GC, GT, GU1, GU2). The experiment comprised two drought cycles, each lasting 60 days and separated by a 15-day recovery period, conducted between 29 October 2019 (austral spring) and 22 March 2020 (early autumn). The first drought cycle (29 October 2019–2 January 2020) was followed by a recovery period, after which the second cycle was applied (18 January–22 March 2020).

Each treatment × clone combination included six individual plants, totaling 60 plants. These were randomly assigned within four rows of 15 plants each to minimize spatial variability. This design allows for the assessment of the main effect of watering treatment, the effect of clones, and their interaction.

Irrigation was supplied automatically via a drip system twice daily (6:30 and 19:30 h) for 15 min per session. The system employed two opposing emitters positioned 5 cm below the substrate surface, each with a flow rate of 8 mL min⁻¹. Plants classified as well-watered (WW) received enough irrigation to compensate for the losses caused by evapotranspiration entirely. In contrast, water-restricted (WR) plants were supplied with 30% of that amount, a previously validated level for applying drought stress in these genotypes.

Water loss was periodically measured for each pot using gravimetric assessments [32]. During the recovery period, all plants were irrigated at the WW rate. Substrate water content (θ, %, w/w) was measured biweekly during drought cycles using a time-domain reflectometry sensor (Decagon^®, Pullman, WA, USA). Air temperature (T, °C) and relative humidity (RH, %) were measured at 15 min intervals throughout the experimental period using an RHT10 sensor (Extech Instruments Corp., Nashua, NH, USA) positioned 1 m above the ground. Data were used to calculate the air vapor pressure deficit (VPD, kPa) following the method described by [47]. Within each drought cycle, plant water status and physiological parameters were measured over a one-week period at the beginning and end of each cycle. More details are provided in [32]. The air vapor pressure deficit (VPD, kPa) during the experimental period presented mid-morning and midday averages ranging from 4 to 5.5 kPa. Soil moisture levels were maintained below 10% in all water-restricted plants and above 35% under well-watered conditions.

2.3. Sample Processing

For this study, anatomical analyses were conducted on stem and branch wood tissue collected at the end of the second drought cycle. From each plant, one stem segment (basal section) and one branch segment were sampled. The branch segment was taken from the proximal portion of a subapical branch—i.e., the section closest to the stem. Both segments measured 6–8 cm in length. In branches, all growth occurred during the water stress treatment. In stems, growth started before the experiment; however, the wood formed before and after treatment was not separated. Nevertheless, over two-thirds of stem growth occurred during the treatment (Section S4). As these are diffuse-porous species, distinguishing between pre- and post-treatment growth is technically challenging, which may reduce the precision of radial growth estimates (more details are provided in Section S4).

Cross-sections of each segment, 15–18 µm thick, were obtained using a manual microtome (Euromex^® MT 5501, Euromex Microscopen BV, Duiven, The Netherlands) and stained with a 0.5% (v/v) safranin solution, followed by rinsing in distilled water. Images of the stained sections were captured to measure vessel dimensions, using an AmScope^® MU853B digital camera (United Scope LLC, Irvine, CA, USA) mounted on an Olympus^® CX21-9 light microscope (Olympus Corporation, Tokyo, Japan) (see Appendix A for examples of cross-section images).

Macerates were prepared to evaluate the anatomical traits related to fibers. Three wood samples (30 mg each) were taken longitudinally from stem and branch segments. The samples were placed in glass test tubes containing a 1:1 (v/v) mixture of acetic acid and hydrogen peroxide and incubated at 60 °C for 24 h to ensure complete fiber dissociation. After incubation, samples were rinsed three times with distilled water. A 1 mL aliquot was then taken from each tube and stained with a 0.5% (v/v) safranin solution. A drop of the stained suspension was mounted on a glass slide for light microscopy. Images of each sample were captured using the previously described digital camera mounted on the microscope (see Appendix A for examples of cross-section images).

2.4. Wood Basic Density

Basic density (D, g cm⁻³) of each stem or branch segment was calculated as the ratio of oven-dry mass (m) to saturated volume, i.e., $\mathrm{D} = \mathrm{m}/\mathrm{V}$, assuming a water density of 0.9983 g cm⁻³ at 20 °C. Samples were first carefully debarked and immersed in distilled water for 72 h to ensure saturation. Saturated volume was determined by water displacement, with samples blotted before submersion. Subsequently, samples were oven-dried at 105 ± 2 °C until a constant mass was reached and then weighed to determine their dry mass. Data were used to calculate the branch-to-stem density ratio (Dif_bs, %) as [$(\mathrm{Dbranch} - \mathrm{Dstem})/\mathrm{Dstem}] \times 100$.

2.5. Xylem Anatomical Variables

Images captured from stained cross-sections were analyzed using ImagePro^® Plus software, version 6.3 (Media Cybernetics, Rockville, MD, USA). Three images, each covering an area of 0.723 mm², were taken from each sample to quantify the mean vessel diameter (Vd, µm), maximum vessel diameter (Vd_max, µm), minimum vessel diameter (Vd_min, µm), average vessel range $(\mathrm{Vdif} = {\mathrm{Vd}}_{\max } - {\mathrm{Vd}}_{\min }, \mu \mathrm{m})$, and mean vessel area (Va, µm²). The number of vessels per image was recorded to calculate the vessel density (N, n° mm⁻²). Using these measurements, the vessel lumen fraction $(\mathrm{F} = \mathrm{Va} \times \mathrm{N}$, unitless), which reflects the relative area available for water transport, and the vessel composition $(\mathrm{S} = \mathrm{Va}/\mathrm{N}, {\mathrm{mm}}^4)$, representing the average vessel size within that area, were calculated following the approach described in [21].

Theoretical hydraulic conductivity (ks_theo) was determined based on the Hagen–Poiseuille law as ${\mathrm{k}\mathrm{s}}_{\mathrm{theo}} = (\pi \rho {\mathrm{Vd}}^4)/(\eta 128 \mathrm{N})$ (kg m⁻¹ MPa⁻¹ s⁻¹), where Vd is the vessel diameter (m), η is the dynamic viscosity of water at 20 °C (1.002 × 10⁻⁹ MPa⁻¹ × s⁻¹), ρ is the density of water at 20 °C (998.3 kg m⁻³), and N is the vessel density.

Images obtained from macerates were analyzed to assess fiber length (Fl, µm; n = 60 per sample), fiber total diameter (Ftd, µm; n = 60 per sample), and fiber lumen diameter (Fld, µm; n = 60 per sample) using AmScope ToupView 3.7 version (2022) Data were used to calculate the fiber lumen fraction$(\mathrm{Ff} = \mathrm{F}\mathrm{ld}/\mathrm{F}\mathrm{td}, \mathrm{unitless})$, and fiber wall thickness (Fwt, µm), calculated as the difference between Ftd and Fld, divided by two (i.e., single wall thickness).

2.6. Hydraulic Conductivity

Stem hydraulic conductivity (k_h, kg s⁻¹ m MPa⁻¹) was measured in 60 plants at the end of the second drought cycle. To minimize sampling-induced cavitation, plants were irrigated the evening before harvest. Subapical shoot segments were collected the following day, defoliated underwater, and trimmed 30 cm from the apex. Hydraulic conductivity (k_h) was determined using the pipette method [48], followed by maximum conductivity (k_hmax) after embolism removal via pressurized perfusion. Percent loss of conductivity (PLC) was calculated as $\mathrm{PLC} = [(1 - {\mathrm{k}}_{\mathrm{h}}/{\mathrm{k}}_{\mathrm{hmax}})] \times 100.$Specific conductivity (ks, ks_max, kg s⁻¹ m⁻¹ MPa⁻¹) was normalized by cross-sectional area. Further methodological details are available in [32].

2.7. Plant Growth

Plant height (Ht, cm) and stem diameter at ground level (Φ, mm) were measured for all 60 plants. Relative growth (RG, %) was calculated as the percentage change in Ht and Φ from the beginning of the experiment and the end of the second drought cycle as $\mathrm{RG} = \left[\right({\mathrm{G}}_{\mathrm{n}} - {\mathrm{G}}_{\mathrm{n}-1})/({\mathrm{G}}_{\mathrm{n}-1}\left)\right] \times 100$, where G represents the plant height (Ht) or diameter (Φ). For more details on the methodology, see [32].

2.8. Statistical Analyses

To test the hypothesis of differential clonal responses to drought, two-way ANOVA was performed for each variable, considering the effects of the clone, water regime, and their interaction. Assumptions of normality and homogeneity of variances were assessed using the Shapiro–Wilk and Levene’s tests, respectively. When significant main effects or interactions were detected, post hoc comparisons were conducted using Tukey’s test (p ˂ 0.05). For variables violating ANOVA assumptions, the non-parametric Kruskal–Wallis test was applied, followed by Dunn’s test for pairwise comparisons. Proportional data was arcsine square-root transformed prior to ANOVA. All analyses were conducted using InfoStat^® software version 2020 (InfoStat Group, Córdoba, Argentina) [49]. Relationships between anatomical variables and wood density were examined by multiple linear regression. The influence of the lumen fraction (F) and vessel spatial composition (S) on the theoretical specific conductivity (k_Stheo) of branches and stems across clones and treatments was explored via linear or non-linear regression analyses.

To assess the relationship between multiple xylem traits and overall plant performance under well-watered and drought conditions, multivariate analyses were conducted. Due to missing data in some variables, only individuals with complete records across all measured traits (29 variables) were selected, retaining three to four individuals per clone and treatment. A small number of remaining missing values (29 out of 957) were imputed using the mean of the other individuals from the same clone and treatment. A correlation matrix was first computed to exclude variables with high correlations (Pearson’s r ≥ 0.7), resulting in a final set of 22 variables. These were used to perform an exploratory cluster analysis based on standardized data, using the Euclidean distance and the Ward.D linkage method in Navure Professional Version 2.6.1 software [50], which was also used for the correlation analysis. Subsequently, a principal component analysis (PCA) was performed in RStudio version 2025.05.1+513 [51] using the FactoMineR package v2.11 [52], along with Factoextra v1.0.7, Factoinvestigate v1.9, and additional packages (ggplot2, v3.5., dplyr v1.1.4, tibble v3.2.1) for data handling and graphical representation.

3. Results

3.1. Wood Density

Wood density varied among genotypes and between plant organs. Red gum hybrids (GC and GT) exhibited a stem wood density approximately 20% higher than Eucalyptus grandis clones (GG) (p < 0.0001), with mean values of 0.500 and 0.401 g cm⁻³, respectively. The GC clone exhibited the highest wood density in both stem and branches. GU hybrids presented intermediate values, approximately 10%–12% lower than those of GC and GT. On average, branch wood was 25% denser than stem wood across genotypes. The largest proportional difference between stem and branch density was observed in the E. grandis clone (34%), calculated between stem wood density, which was not affected by the treatment, and branch wood density under drought conditions. In contrast, the smallest difference occurred in GC under the same treatment (13%) (

Table 1. Mean stem and branch basic wood density (D, g cm⁻³) in E. grandis (GG), E. grandis × camaldulensis (GC), E. grandis × tereticornis (GT), and E. grandis × urophylla (GU1 and GU2) clones under well-watered (WW) and water-restricted (WR) conditions. The proportional difference between organs (Dif_bs, %) was calculated for each clone and treatment as $\left[\right(\mathrm{Dbranch} - \mathrm{Dstem})/\mathrm{Dstem}] \times 100$. Data are presented as mean ± standard error (SE). Different lowercase letters indicate statistically significant differences within each clone for a given treatment (ANOVA, p < 0.05).

Clone	Water Regime	Stem Wood Density	Branch Wood Density	Difbs (%)
GG	WW	0.391 ± 0.013 a	0.499 ± 0.044 b	27.62
	WR	0.412 ± 0.013 a	0.603 ± 0.040 a	46.36
GC	WW	0.465 ± 0.028 b	0.620 ± 0.025 a	33.33
	WR	0.516 ± 0.032 a	0.624 ± 0.043 a	21.32
GT	WW	0.481 ± 0.020 b	0.543 ± 0.029 b	12.89
	WR	0.525 ± 0.017 a	0.627 ± 0.024 a	19.43
GU1	WW	0.443 ± 0.009 a	0.519 ± 0.035 b	17.16
	WR	0.459 ± 0.010 a	0.600 ± 0.036 a	30.72
GU2	WW	0.416 ± 0.014 a	0.524 ± 0.026 a	25.96
	WR	0.440 ± 0.013 a	0.528 ± 0.027 a	20.00

).

Drought increased wood density—considering stems and branches combined—across all genotypes, with the greatest relative increase (30%) observed in the GG clone. When analyzing organs separately, branches showed a stronger response to water deficit, exhibiting an average wood density increase of 11% across the clones, compared to a 5% increase in stems. Only the red gum hybrids showed a significant stem wood density increase of approximately 10% under drought conditions. In contrast, branch wood density increased in all clones except for GC (Table 1).

3.2. Xylem Anatomical Traits

3.2.1. Variation in Vessel Characteristics

At the stem level, vessel density (N) increased under drought conditions (p < 0.0001) by an average of 60%, ranging from 50 to 64 vessels mm⁻² in water-stressed plants. The greatest increase was observed in the red gum hybrids (reaching up to 75%), followed by the E. grandis clone (

Table 2. Mean stem vessel density (N, n° mm⁻²), vessel diameter (Vd, µm), difference between maximum and minimum vessel diameter (Vdif, µm), vessel lumen area (Va, µm²), lumen fraction (F), vessel size distribution (S, mm⁴), fiber length (µm), fiber diameter (Ftd, µm), fiber lumen diameter (Fld, µm), fiber lumen fraction (Ff), and fiber wall thickness (Fwt, µm) in E. grandis (GG), E. grandis × camaldulensis (GC), E. grandis × tereticornis (GT), and E. grandis × urophylla (GU1 and GU2) clones under well-watered (WW) and water-restricted (WR) conditions. Data are presented as mean ± standard error (SE) or median with interquartile range (IQR), according to data distribution. Lowercase letters denote significant differences between treatments within each clone (p < 0.05), determined by Tukey’s HSD test for parametric variables (N, Vd, Vdif, F, Ftd, Fld, Ff) and Dunn’s post hoc test for non-parametric variables (Va, S, Fl, Fwt).

Clone	Conductive Elements							Biomechanical Support
	Water Regime	N (n° mm−2)	Vd (µm)	Vdif (µm)	Va (µm2)	F (Unitless)	S (mm4)	Fl (µm)	Ftd (µm)	Fld (µm)	Ff (Unitless)	Fwt (µm)
GG	WW	35.03 b	141.40 a	133.82 a	19,099 a	0.19 a	4.5 × 10−4 a	728.27 a	22.14 a	14.05 a	0.63 a	4.05 b
	WR	61.37 a	116.36 b	123.37 a	14,498 b	0.15 b	1.9 × 10−4 b	647.39 b	16.32 b	5.80 b	0.36 b	5.26 a
GC	WW	36.17 b	140.32 a	154.24 a	18,281 a	0.18 a	3.8 × 10−4 a	797.31 a	19.43 a	9.28 a	0.48 a	5.08 a
	WR	63.75 a	115.47 b	127.24 b	12,361 b	0.15 a	1.5 × 10−4 b	804.62 a	14.88 b	4.64 b	0.32 b	5.12 a
GT	WW	34.25 b	137.18 a	136.29 a	21,635 a	0.21 a	4.8 × 10−4 a	796.97 a	15.83 a	7.43 a	0.47 a	4.23 a
	WR	53.60 a	117.56 b	135.75 a	17,650 a	0.18 a	2.9 × 10−4 b	754.49 a	12.38 b	5.53 b	0.45 a	4.29 a
GU1	WW	45.81 b	113.99 a	162.30 a	16,297 a	0.23 a	2.9 × 10−4 a	721.65 a	17.16 a	10.98 a	0.64 a	3.14 a
	WR	62.73 a	115.98 a	151.81 a	16,181 a	0.16 b	2.1 × 10−4 a	669.08 b	14.70 b	6.99 b	0.58 a	3.86 a
GU2	WW	34.97 b	130.94 a	152.82 a	17,705 a	0.22 a	4.2 × 10−4 a	705.04 a	17.48 a	10.35 a	0.60 a	3.47 a
	WR	55.60 a	113.39 a	147.44 a	15,602 a	0.18 b	2.3 × 10−4 b	593.60 b	14.24 b	6.98 b	0.49 a	3.63 a
SE		4.28	4.70	13.29	1409	0.015	0.50 × 10−4	25.12	0.43	0.67	0.03	0.40

). Vessel diameter (Vd) was reduced (p = 0.0007) under water restriction in red gum hybrids (17%) and GG clones (12%), while the GU hybrids showed no response. As observed in Vd, vessel lumen area (Va) significantly decreased (p = 0.0222) in drought-stressed plants of GG and GC, with the most pronounced decrease (35%) found in GC. Additionally, the difference between maximum and minimum vessel diameter (Vdif), an indicator of lumen size uniformity, was reduced by 21% on average (p = 0.0134). The vessel composition within the sapwood (S) declined under drought (H = 25.86, p = 0.0021), except in the GU1 clone, where no change was observed. Notably, the drought-induced reduction in S was twice as large in red gum hybrids and GU1 compared to the E. grandis clone. The vessel lumen fraction (F) also declined under drought, particularly in GG (p = 0.0088) and GU clones (p = 0.0248), with an average decrease of 25%. F remained unchanged in the red gum hybrids (Table 2).

Drought affected branches differently than stems. Vessel density (N) increased under water restriction, but this response was limited to the GG clone (p = 0.010). Conversely, vessel diameter (Vd) was significantly reduced in the GG (p = 0.0114) and GU2 (p = 0.0027) clones. Unlike stems, Vd in the GT clone’s branches remained unaffected by drought, and no significant changes in the vessel lumen area (Va) were detected in any clone. Vessel diameter variation (Vdif) also exhibited organ-specific patterns, decreasing in branches of GC (34%) and GT (30%) hybrids, whereas in stems, this reduction was restricted to GC only. Vessel composition (S) followed a similar organ-specific trend, showing an average 36% reduction in branches of both GU hybrids under drought (p = 0.0409;

Table 3. Mean branch vessel density (N, n° mm⁻²), vessel diameter (Vd, µm), difference between maximum and minimum vessel diameter (Vdif, µm), vessel lumen area (Va, µm²), lumen fraction (F), vessel size distribution (S, mm⁴), fiber length (µm), fiber diameter (Ftd, µm), fiber lumen diameter (Fld, µm), fiber lumen fraction (Ff), and fiber wall thickness (Fwt, µm) in E. grandis (GG), E. grandis × camaldulensis, E. grandis × tereticornis, and E. grandis × urophylla clones under well-watered (WW) and water-restricted (WR) conditions. Data are presented as mean values ± standard error (SE) or median with interquartile range (IQR), according to data distribution. Lowercase letters denote significant differences between treatments within each clone (p < 0.05), determined by Tukey’s HSD test for parametric variables (N, Vd, Vdif, F, Ftd, Fld, Ff) and Dunn’s post hoc test for non-parametric variables (Va, S, Fl, Fwt).

Clone	Conductive Elements							Biomechanical Support
	Water Regime	N (n° m−2)	Vd (µm)	Vdif (µm)	Va (µm2)	F (Unitless)	S (mm4)	Fl (µm)	Ftd (µm)	Fld (µm)	Ff (Unitless)	Fwt (µm)
GG	WW	80.36 b	101.36 a	92.03 a	9661 a	0.25 a	9.2 × 10−5 a	593.07 a	12.04 a	6.33 a	0.46 a	3.74 a
	WR	109.56 a	77.06 b	97.85 a	9335 a	0.15 a	9.6 × 10−5 a	477.07 b	12.52 a	5.50 a	0.46 a	3.25 a
GC	WW	89.93 a	87.40 a	83.78 a	7080 a	0.15 a	6.2 × 10−5 a	636.08 a	13.27 a	6.22 a	0.47 a	3.52 b
	WR	82.17 a	86.14 a	55.16 b	7526 a	0.15 a	6.5 × 10−5 a	590.76 a	13.28 a	5.36 a	0.41 b	3.96 a
GT	WW	83.17 a	98.21 a	99.99 a	9362 a	0.13 a	9.8 × 10−5 a	648.94 a	14.99 a	5.95 a	0.40 a	4.04 b
	WR	82.13 a	84.44 a	70.47 b	6698 b	0.17 a	6.3 × 10−5 a	616.59 a	12.66 b	4.59 b	0.36 b	4.52 a
GU1	WW	74.70 a	97.47 a	105.62 a	8876 a	0.17 a	9.3 × 10−5 a	597.67 a	12.04 a	4.78 a	0.40 a	3.63 a
	WR	85.93 a	90.15 a	96.21 a	8240 b	0.14 b	7.1 × 10−5 b	472.76 b	12.52 a	5.07 a	0.40 a	3.73 a
GU2	WW	81.17 a	97.70 a	80.12 a	9109 a	0.17 a	8.7 × 10−5 a	563.74 a	12.38 a	4.22 a	0.30 a	3.71 a
	WR	87.33 a	77.43 b	63.23 a	5843 b	0.12 b	5.7 × 10−5 b	434.04 b	13.45 a	4.50 a	0.34 a	4.04 a
SE		9.30	3.90	11.11	637	0.02	4.76 × 10−6	16.79	0.99	0.70	0.03	0.39

). In contrast, the reduction was only observed in stems for GU2 (Table 2).

3.2.2. Anatomical Characteristics of Xylem Fibers

In stems, fiber length (Fl) varied significantly among clones (p < 0.0001) and between water regimes (p = 0.024). However, red gum hybrids, which exhibited the highest Fl, were not affected by the drought treatment in this variable (Table 2). Fiber diameter (Ftd) decreased markedly under drought conditions (p < 0.0001), with the most significant reductions observed in E. grandis (28%) and the GC clone (25%). The GC clone also had the highest Ftd (19.23 ± 0.82 µm), whereas the lowest was recorded in GT (14.11 ± 0.82 µm). A similar but more pronounced pattern was observed for fiber lumen diameter (Fld), with the GG clone showing a 50% reduction from 14.05 µm in well-watered plants to 5.80 ± 0.70 µm under drought. The clones exhibited a reduction in the fiber lumen fraction (Ff) of varying degrees, with significant differences in GC and the most pronounced in the GG clone (Table 2). Additionally, GG was the only clone to show an increase in fiber cell wall thickness (Fwt) under drought, with a 30% rise (Table 2).

In branches, fiber length (Fl) was also highest in the red gum hybrids, averaging 15% greater length than in the other clones, and—as observed in stems—remained unaffected by water drought. In contrast, the remaining clones exhibited a 21% reduction in Fl (p < 0.0001). The GT hybrid was the only clone to exhibit significant reductions in fiber diameter (Ftd), fiber lumen diameter (Fld), and fiber lumen fraction (Ff), accompanied by a 12% increase in fiber wall thickness (Fwt) in response to drought. A similar increase in Fwt was also observed in the GC clone (Table 3).

Overall, drought caused greater anatomical variation in stems than in branches, particularly in terms of fiber traits. In stems, drought led to reductions in fiber diameter and fiber lumen size across clones, while an increase in cell wall thickness was exclusive to GG. In branches, fiber length remained stable in red gum hybrids, which also exhibited an increase in Fwt. In contrast, reductions in Ftd and Fld under drought were detected only in the GT hybrid.

3.3. Relationship Between Anatomical Traits and Wood Density

The models developed to predict wood basic density using stem and branch data identified stem fiber length (Fl) and total fiber diameter (Ftd) as significant predictors of stem wood density in the studied clones. These variables had the highest associated t-values, indicating a stronger contribution to the model. Both showed a positive linear relationship with wood density (

Table 4. Coefficients and associated statistics from multiple linear regression models predicting stem wood density based on anatomical variables in stems of E. grandis (GG), E. grandis × camaldulensis, E. grandis × tereticornis, and E. grandis × urophylla clones under well-watered (WW) and water-restricted (WR) conditions.

Predictor Variable	Coef	SE	LI (95%)	LS (95%)	t	p-Value	R2
const	422.11	6.55	287.60	556.61	6.44	<0.0001	0.58
Fwt	9.53	5.61	−1.98	21.05	1.70	0.1009
Fl	0.17	0.07	0.02	0.32	2.37	0.0254
Ftd	−11.15	2.21	−15.67	−6.62	−5.06	<0.0001
Vd	0.37	0.33	−0.31	1.05	1.12	0.2733

). In contrast, the model for branch wood density did not yield statistically significant coefficients (p > 0.05).

3.4. Theoretical Hydraulic Conductivity

Under drought conditions, the theoretical specific hydraulic conductivity (ks_theo) decreased significantly in both stems (H = 29.45, p = 0.00054) and branches (H = 24.89, p = 0.0031). However, the magnitude and significance of this reduction varied between plant organs and among clones. In stems, significant decreases were observed in the clones GG, GC, and GU2, with GU2 showing the most pronounced decline (up to 64%). In branches, significant reductions in ks_theo occurred in water-stressed plants of the GG, GU1, and GU2 clones. It is worth noting that the E. grandis clone (GG) exhibited the most severe decline overall, with a 67% reduction in branches (

Table 5. Mean theoretical hydraulic specific conductivity (ks_theo, kg s⁻¹ m⁻¹ MPa⁻¹) in stems and branches of E. grandis (GG), E. grandis × camaldulensis (GC), E. grandis × tereticornis (GT), and E. grandis × urophylla (GU1 and GU2) clones under well-watered (WW) and water-restricted (WR) conditions. Values are presented as mean ± standard error (SE). Different lowercase letters indicate significant differences between treatments within each clone (ANOVA, p < 0.05).

Clone	Water Regime	Stem Theoretical ks (kg s−1 m−1 MPa−1)	Branch Theoretical ks (kg s−1 m−1 MPa−1)
GG	WW	7.42 ± 0.84 a	2.14 ± 0.39 a
	WR	3.46 ± 0.92 b	0.71 ± 0.16 b
GC	WW	7.20 ± 1.03 a	1.07 ± 0.11 a
	WR	3.32 ± 0.51 b	1.10 ± 0.11 a
GT	WW	6.54 ± 0.93 a	1.06 ± 0.48 a
	WR	4.59 ± 0.58 a	1.80 ± 0.21 a
GU1	WW	6.00 ± 1.13 a	1.68 ± 0.26 a
	WR	3.60 ± 0.82 a	1.33 ± 0.07 b
GU2	WW	8.66 ± 1.28 a	1.66 ± 0.11 a
	WR	3.13 ± 0.64 b	0.75 ± 0.15 b

).

In the same plants, the maximum specific hydraulic conductivity (ks_max) was measured in apical stem segments after embolism removal, as previously reported by [32]. A significant correlation (Pearson’s r) was found between measured stem ks_max and theoretical ks across all clones and treatments (r = 0.60; p = 6.8 × 10⁻⁵). Clone-specific analyses revealed the strongest correlation in GU1 (r = 0.86; p = 0.010) and GU2 (r = 0.76; p = 0.015). In contrast, the GC clone showed the weakest but still highly significant correlation (r = 0.51; p = 0.024), suggesting a weaker relationship between hydraulic conductivity and the studied vessel traits. Intermediate correlation values were observed in clones GG (r = 0.63; p = 0.009) and GT (r = 0.67; p = 0.007). Similar correlations were observed between ks_max and ks_theo within each water regime, with r = 0.70 (p < 0.0001) under drought conditions and r = 0.65 (p < 0.0001) in well-watered plants.

Additionally, the regression slopes of these parameters versus the vessel-related parameter F, which quantifies the lumen fraction at the tissue level, were consistently lower in branches than in stems across all clones (Figure 1A). Although F values were similar between organs within each clone (x-axis of Figure 1; mean values in Table 2 and Table 3), the ks_theo was higher in stems than in branches for a given F (Figure 1) across WW and WR plants. However, in clones GG and GU2, the relationship between ks_theo and F differed between stems developed under drought and control conditions. For a given lumen fraction and comparable variability, drought-stressed stems exhibited lower k_Stheo values, as indicated by reduced regression slopes (red lines in Figure 1). Despite F remaining relatively stable or even increasing under drought, k_Stheo decreased substantially, leading to functional convergence between stressed stems and branches in these clones (Figure 1).

In contrast to F, which presents similar ranges in stem and branches, S, which quantifies the size distribution or composition of vessels, was consistently lower in branches than stems within each clone. Theoretical k_S increased with the logarithm of S across both organs and irrigation treatments in all clones (Figure 1B). As a whole, these results indicate that the total amount of lumen space is not the most important determinant of conductivity but rather how that space is distributed across different vessel sizes and numbers.

Each row corresponds to a different clone; Figure 1A and 1B within each row show the respective relationships for that genotype. Regressions lines include both well-watered and water-restricted treatments, due to consistent patterns, except for stems of GG and GU2, which are shown separately. In these cases, light blue symbols represent stems from well-watered plants and red symbols represent stems from water-restricted plants. Note that S is plotted on a log₁₀ scale to account for magnitude differences between stems and branches.

3.5. Multivariate Analyses

Exploratory cluster analysis distinguished two main groups, consisting of WW and WR individuals of all clones. Within each group, the GC clone appeared as the most differentiated, while the remaining clones clustered differently depending on the irrigation treatment (Section S1). The PCA explained a relatively low percentage of variability, with the first three principal components (PC1, PC2, and PC3) accounting for 48.1% of the total variance (Figure 2). When irrigation treatments were not considered, the GC clone was separated from the GG clone, being more closely associated with a higher k_Smax, height growth, stem and branch wood density, and fiber wall thickness. The confidence intervals of both GU clones overlapped with those of the GG clone, and GT was not differentiated from the GC clone (Section S2). When irrigation treatments were taken into account, as in the cluster analysis, two distinct groups were separated along the x-axis: all the clones grown under well-watered conditions clustered on the center-right, while those subjected to water-restricted conditions were grouped on the left (Figure 2). PC1 (Dim1 in right panel of Figure 2 and Section S2), which explained the highest percentage of the variance and most effectively separated the two groups, was primarily associated with vessel number (VN), vessel diameter (Vd), vessel composition (S), and fiber diameter (Ftd) in one or both organs (stems and/or branches). Vessel number was higher in drought-stressed plants, whereas fiber diameter was greater in well-watered individuals.

Within these general groupings, however, some clones were further differentiated, mainly along the y-axis of the first biplot (PC1 and PC2). Among WR plants, the GC clone (upper left) was clearly separated from the others (lower left). In addition, within this group of four clones, the GG clone (orange in Figure 2) was distinguished from GT (green). The WR-GC clone was associated with a higher maximum hydraulic conductivity (ks_max), stem wood density (WD_st), and fiber wall thickness (Fwt_st), traits that reflect a combination of high hydraulic efficiency and safety. In contrast, the WR-GG and WR-GU clones were associated with a higher vessel number (VN) and lumen fraction (VF) in stem tissues. Notably, in these WR clones, higher lumen fractions did not correspond to higher theoretical ks values (Figure 1, red dots and lines).

On the other hand, among WW plants, the red gum clones (GC and GT) occupied a similar region in the biplot and were associated with a greater diameter growth (RGd) and longer fiber length (Fl) in both stems and branches. In contrast, the GG and both GU clones grouped together and were associated with a broader range of vessel diameter (Vdrange) (Figure 2). The second biplot (Dim1 and Dim3) did not distinguish clones within each irrigation treatment; however, the two main groups remain separated along the x-axis.

4. Discussion

In this study, we investigated the effects of drought on anatomical and functional variation in xylem traits across commercial E. grandis (GG) clones and their hybrids E. grandis × camaldulensis (GC), E. grandis × tereticornis (GT), and E. grandis × urophylla (GU). We assessed changes in conductive and supportive xylem elements of stems and branches to gain a better picture of whole-plant hydraulic responses to drought. By quantifying vessel and fiber traits, theoretical hydraulic conductivity, and wood density, we identified structural adjustments with potential hydraulic consequences, highlighting xylem plasticity under drought stress. These results were then analyzed in light of other functional and performance traits of the exact clones previously reported by [32].

In that earlier work, red gum hybrids (GC and GT) exhibited higher transpiration rates, osmotic or elastic adjustments in foliar tissues, and a greater percentage loss of conductivity (PLC), likely due to their water-spending strategy. Despite their common physiological patterns, superior stem growth under drought conditions was only observed in the GC clone. In contrast, the GG clone exhibited a more drought-avoidant strategy, characterized by tighter stomatal regulation, resulting in low in situ cavitation, as indicated by the lowest PLC, and lower growth. GU hybrids displayed conservative water-use traits similar to those of the GG clone. However, they achieved greater stem growth under drought, specially GU1, suggesting a more favorable trade-off between hydraulic safety and growth [32]. In the present study, we observed that, although all the clones showed relatively common general responses to drought at the xylem level, differing in the magnitude rather than the direction of the changes, xylem traits of the GC clone were the most differentiated, with values suggesting both high efficiency (high theoretical ks) and safety (high wood density and fiber wall reinforcement). In the following sections, we discuss the observed trends of the studied traits in the different clones and their potential implications to explain their performance under well-watered and drought conditions.

4.1. Response of Xylem Traits to Drought

As revealed by the individual traits integrated through the multivariate analyses, the main variation observed in the xylem traits across clones and water availability treatments was primarily associated with the latter: all clones clustered according to the irrigation treatment. However, we identified some differences within them, which are discussed in Section 4.1.1. Overall, well-watered conditions were associated with larger vessels, longer and wider fibers, a broader range of vessel sizes (Vdif), and a higher vessel composition (S). In contrast, drought conditions resulted in a higher wood density, thickness of fiber walls, vessel frequency, and vessel lumen fraction (F). This general pattern was expected, as traits associated with hydraulic efficiency were primarily expressed in clones under well-watered conditions.

Conversely, traits associated with xylem safety were linked to clones grown under drought conditions. This suggests that plasticity at the xylem level (both the size and distribution of vessels and fibers and wood density) tends to maximize safety under drought, probably resulting in a general trade-off between efficiency and safety when considering all clones together. The significant drought-induced plasticity observed in several traits, at least within some clones, contrasts with the lack of wood anatomical variation between sites with contrasting aridity reported by [22], who compared seven eucalypt species growing both in wet and dry native forest sites. However, our findings are consistent with studies on Eucalyptus commercial genotypes, which have revealed relatively high xylem sensitivity to water availability, particularly in vessel traits, fiber characteristics, and wood density [53,54].

4.1.1. Differential Responses at Clone Level

Anatomical differences among clones help explain their divergent hydraulic strategies. In GC and GT hybrids, stem wood density was the highest, approximately 25% greater than that of the GG clone, which exhibited the lowest values. Moreover, stem wood density increased in response to drought in these two clones, whereas it remained unchanged in the other three (Table 1). This increase in density was likely associated with thicker cell walls and smaller lumen areas in conductive and/or supportive xylem elements. Such anatomical changes likely improved mechanical strength and enhanced resistance to cavitation under low water potential [8,10]. Red gum hybrids, differentiated from other clones, also maintain a relatively constant vessel lumen fraction (F) under drought, even when presenting a reduced vessel diameter (Vd), as seen in the GG clone (Table 2). Besides the mean vessel diameter, in GT, this adjustment occurred without changing the vessel diameter range (Vdif), which, in turn, was reduced in the GC clone, as well as in all the other clones. In eucalypts, it has been shown that a larger amplitude in vessel sizes is correlated with both higher ks and resistance to cavitation [27,55]. The implications of the observed maintenance in stem lumen fraction on theoretical ks are discussed below.

In contrast, the GG clone exhibited substantial changes in fiber traits in response to drought, including a 65% reduction in fiber lumen diameter (Fld) and a 30% increase in fiber wall thickness (Fwt), as well as adjustments in vessel lumen mean size and distribution. Besides the functional implications of those changes, the marked sensitivity of fiber characteristics—although not reflected in whole-stem wood density—may have important implications for the wood industrial properties of E. grandis, as suggested for E. globulus [54].

GU hybrids exhibited an intermediate behavior between red gum hybrids and the GG clone (Table 1 and Table 2), with no changes in stem wood density and a decrease in vessel lumen fraction—similar to the GG clone—; an increase in vessel frequency, as observed in all the clones; and a uniquely stable mean and range of vessel sizes (diameter and area) and composition (S). In terms of stem fiber plasticity, some traits varied similarly to the GG clone (i.e., reduced fiber length) and others similar to the red gums (no change in fiber fraction and wall thickness). It is important to note, however, that although the GU clones behaved somewhat similarly (i.e., they were grouped in the PCA), their xylem (this study) and growth [32] responses to drought were not identical (i.e., in addition to some differences at the univariate level, the cluster analyses grouped them in the control conditions but not under drought). This finding is in agreement with previous studies on hybrids of E. grandis with E. urophylla in South Africa, which showed a large variation in drought responses among hybrids related in a variable way to the xylem anatomy of the clones [34].

In summary, some clones exhibited significant variation in wood density in response to drought (red gums), while others varied in their fiber sizes, lumen diameter, or wall thickness. Overall, these shifts resulted in a greater proportion of cell wall material per unit volume, probably increasing the mechanical strength of the wood but also allowing higher internal tensions due to drought in all the clones. It is probable that these changes are accompanied by other compensatory anatomical strategies focused on vessel size modulation and pit architecture to minimize the cavitation risk, as has been described for other taxa [56,57]. However, the implications of the observed changes in functional terms (xylem safety) require measurements of vulnerability to cavitation that are still lacking in the studied clones.

4.1.2. Differential Responses at Organ Level

Most of the previous analysis was focused on the observed changes at the stem level. However, our results reveal organ-specific adjustments in xylem anatomy under drought conditions, adding an extra layer of complexity to understanding the plastic changes that a plant can display in response to water availability. These changes varied among clones, reflecting genotype-specific strategies that influenced wood anatomy and basic density in both stems and branches, ultimately leading to differences in hydraulic function.

Wood density was, on average, 25% higher in branches than in stems. This higher density in distal organs likely reflects an adaptive strategy to enhance hydraulic safety where the risk is greater due to higher tension during water flux [58]. However, the degree of differentiation between organs varied among the clones and the water availability treatments, with the highest differentiation between stem and branch in the GC clone under well-water conditions and in the GG clone under drought stress. A large difference in wood density between the stem and branches has been described for adult trees in E. grandis, larger than in E. viminalis and E. globulus, suggesting a high degree of hydraulic segmentation in this species [12].

As discussed in the previous section, the observed changes in xylem traits indicate a greater investment in mechanical reinforcement under drought [59]. However, they may also imply a reduced capacity for hydraulic compensation [60]. Although we did not directly compare xylem traits between organs, our findings suggest potential functional differentiation: stems appear to prioritize hydraulic conductivity, while branches may emphasize hydraulic safety. In this regard, under drought, wood density increased in both organs, but with a more pronounced rise in branches (11%), doubling the response observed in stems across clones. To validate this hypothesis, further studies focusing on detailed anatomical traits, especially vessel pit structure and distribution, are needed. These traits are crucial in mediating the trade-off between hydraulic efficiency and safety under drought conditions [27,61].

One limitation of this study is the exclusion of root development parameters. This decision was based on the inherent constraints of pot cultivation under greenhouse conditions, which often induce atypical root architectures, such as root circling and limited lateral expansion, which do not accurately reflect the conditions experienced by field-grown trees. These artificial constraints may mask genuine genotypic differences in root traits between Eucalyptus hybrids and E. grandis [62,63]. Furthermore, pot size can significantly influence water and nutrient availability, introducing additional variability that complicates the interpretation of belowground responses [64]. Since the main objective of this study was to evaluate aboveground growth and physiological performance, root traits were intentionally excluded to avoid confounding effects associated with artificial root restriction. Future studies conducted under field conditions or using root observation systems may help clarify potential genotype-specific differences in root development.

4.2. Vessel Architecture and Hydraulic Efficiency of the Studied Clones

The response of theoretical specific conductivity (k_Stheo) to vessel composition (S) was consistent across clones and organs, indicating a stable functional relationship. However, S values were ten times higher in stems than in branches, reflecting significant anatomical differences despite this conserved pattern. As noted by [21], the size distribution of vessels influences the trade-off between hydraulic efficiency and safety by affecting both redundancy and the risk of embolism spread. Higher S values typically indicate fewer, wider, and more interconnected vessels, which enhances conductivity but increases vulnerability to cavitation. In contrast, lower S values reflect narrower, more numerous, and less connected vessels, which favor safety by limiting the propagation of embolism.

Under drought, S declined by an average of 35% in GU hybrid branches and in stems of all clones except GU1, with a notable 60% reduction in GC. These changes suggest a shift toward greater safety at the cost of efficiency. While these patterns are broadly applicable to angiosperms, interpretation in eucalypts must consider their predominance of solitary vessels. Eucalyptus xylem anatomy is complex [65], with cell types such as vasicentric tracheids and fiber-tracheids playing functional roles [27,55]. Further research is needed on these components.

Regarding the relationship between k_Stheo and the lumen fraction (F), in contrast to common patterns across stems and branches about S, k_Stheo was consistently higher in stems than in branches for a given F value in all clones. This suggests that the distribution of vessel sizes (quantified in S) rather than the total amount of lumen space may be crucial to understanding the variation in xylem efficiency across clones and organs. Therefore, a plastic change in S could imply an important change in hydraulic conductivity, even if a similar lumen fraction is maintained. In this regard, in a study of stems of adult trees of three Eucalypus species, Barotto et al. [12] found a higher relative impact of F than S on their ks_theo. However, the degree of variation among and within species was much larger in S than in F, suggesting that the last is a more conservative feature within eucalypts xylem. In this study, GG and GU2 clones modified both F and S under drought. Red gum hybrids changed only S, while GU1 adjusted only F. These contrasting patterns highlight the diversity of anatomical responses among clones.

On the other hand, clones GG and GU2 exhibited an additional behavior in response to drought compared to the other three clones. In these clones, for a given stem F value, k_Stheo was significantly lower in drought-stressed than in well-irrigated plants, as shown by reduced regression slopes. This suggests high plasticity in stem hydraulic function, with greater efficiency loss than in other clones. Notably, the lumen fraction remained stable or increased under drought, yet k_Stheo declined, indicating a decoupling between anatomical investment and hydraulic performance. This response suggests that stems of these two clones under drought adopt xylem traits more similar to those of branches, indicating an adaptive vascular reconfiguration. Comparable patterns of organ-specific plasticity have been reported in Cunninghamia sp., where xylem shifts mediated hydraulic variation across organs under water stress [16].

In stems, k_Stheo showed a moderate-to-strong correlation (r = 0.58) with measured ks_max from our previous study. In xylem research, this magnitude is considered a strong correlation [66], with vessel lumen size alone explaining ~60% of the ks_max variation. The remaining ~40% may be attributed to traits such as pit type and density, perforation plates, and the presence of non-functional vessels, which affect overall conductivity beyond vessel dimensions. In addition to these vessel characteristics, it is important to consider that other cell types participate in water (and air) movement in eucalypts. As previously noted, vasicentric tracheids may enhance both hydraulic efficiency and safety in eucalypts, acting as conductive bridges for water and barriers to the spread of embolism [55].

4.3. Relationship Between Xylem Traits and Clones’ Performance

Clone performance, measured by height and basal diameter growth, varied among genotypes under both control and drought conditions [32]. All hybrids outperformed the GG clone in height growth across both irrigation treatments. Under well-watered conditions, stem diameter growth was relatively uniform among clones; however, under drought, the GC clone outperformed the others. GU1 also performed relatively well under drought, though still below GC. Several morpho-physiological traits related to this performance were discussed in [32]. In the present complementary study, we explored associations with functional wood anatomy and density, demonstrating that xylem traits also play a critical role in explaining clone responses to drought.

Multivariate analyses identified the GC clone as the most differentiated, consistent with its superior growth under both water regimes. This strong performance challenges the commonly reported trade-off between high growth potential and stress resistance [64]. In this study, GC exhibited xylem characteristics and plasticity indicative of both high hydraulic efficiency and safety, likely supporting sustained water flux to leaves under tension, despite high conductivity losses, as previously shown for this clone and its parental species E. camaldulensis [61]. Notably, PCA grouped both red gum hybrids under well-watered conditions but not under drought conditions, aligning with the poor drought performance of the GT clone despite its physiological similarities with GC [32]

The shared anisohydric behavior of red gums hybrids supports a fast-growth strategy under water-limited conditions [4]. However, GT’s stem diameter growth under drought did not match that of GC, likely reflecting differences in biomass allocation patterns. In terms of xylem traits, GT exhibited lower vessel plasticity compared to GC, as indicated by a narrower range of vessel diameters and areas, which allowed it to maintain theoretical stem-level hydraulic conductivity (ks_theo) under drought. Although elevated ks_theo may favor efficient water transport, it may not offset the risk-prone stomatal behavior characteristic of red gum hybrids. Further investigation into the plasticity of vulnerability to cavitation is needed to determine whether GT’s limited vessel adjustment also reflects constrained plasticity in hydraulic safety. While restricted cavitation plasticity has been documented in Eucalyptus branches [31] and E. obliqua leaves [6], few studies have addressed this trait across a range of pure and hybrid Eucalyptus clones. Additional evidence is needed to clarify whether vulnerability to cavitation represents a conservative trait across these genotypes.

The GG clone exhibited several anatomical adjustments indicative of marked plasticity in response to drought. Previous studies have reported high anatomical variability in adult plants of other E. grandis clones, suggesting considerable xylem plasticity under environmental stress [12]. However, in GG, these adjustments, while likely enhance xylem safety, did not prevent the strong stomatal control observed, in contrast to the looser regulation in red gum hybrids [32]. Water-restricted GG plants exhibited the lowest percent loss of conductivity (PLC) among all clones, even lower than their well-irrigated counterparts [32], suggesting that xylem modifications facilitated efficient dehydration avoidance. Nonetheless, this conservative strategy came at the cost of the lowest growth.

Although GU1 and GU2 shared several xylem traits and clustered together in the PCA, the clone exhibiting superior diameter growth under drought (GU1) showed a distinct pattern of stem-level vessel plasticity, as evidenced by the relationship between the theoretical ks and vessel lumen fraction (Figure 1). Moreover, GU1 increased branch wood density in response to drought, whereas GU2 did not exhibit any change. While these traits alone may not fully account for GU1’s enhanced performance under water-limited conditions, they clearly distinguish it from GU2. Comparable variation among GU hybrids has previously been reported [34], highlighting the importance of evaluating each new commercial clone independently to assess its adaptive potential under contrasting water availability.

In conclusion, this study reveals a range of xylem strategies exhibited by the studied clones that partially resemble their growth responses to drought. Inherent values and differential plasticity of xylem traits in branches and stems contribute to understanding the complex mechanisms involved in the drought performance of pure vs. hybrid clones.

5. Conclusions

This study demonstrates that drought elicits anatomically and functionally distinct responses among five commercially relevant Eucalyptus clones, reflecting divergent hydraulic strategies under water-limited conditions. In particular, red gum hybrids (E. grandis × camaldulensis and E. grandis × tereticornis) sustained the theoretical specific hydraulic conductivity during drought by maintaining the vessel lumen fraction while reducing the vessel diameter, especially in branch xylem. This combination, along with increased wood density, indicates a conservative hydraulic strategy that balances conductivity and safety. Among the tested genotypes, the E. grandis × camaldulensis clone exhibited the most distinct xylem profile, characterized by a high theoretical conductivity co-occurring with an elevated wood density, suggesting a potential for both efficient water transport and resistance to hydraulic failure. These characteristics are in agreement with its better growth performance under different water availabilities, as previously reported.

In contrast, the pure species (E. grandis) and certain hybrids (e.g., E. grandis × urophylla) exhibited signs of vascular adjustment under drought, marked by reductions in both vessel lumen fraction (F) and composition (S), despite a stable or even increased individual vessel area. These anatomical shifts were associated with significant declines in theoretical hydraulic conductivity. Such patterns underscore the role of hybridization with drought-adapted species in enhancing xylem resilience and maintaining the water transport capacity under stress. By integrating organ-level anatomical and functional analyses, this study advances the understanding of wood formation dynamics in response to drought and offers valuable criteria for the selection of drought-adaptive genotypes suited to future climate scenarios.

Future research should further explore the contrasting patterns of hydraulic regulation between stems and branches to determine inherent differences in hydraulic architecture among Eucalyptus clones. A deeper understanding of this axial variation, together with the genotype-specific plasticity observed in xylem traits, will be essential for elucidating the mechanisms underlying drought adaptation and for refining selection criteria in breeding programs targeting water-limited environments.