Physiological traits underlying sodicity tolerance in Jamun (Syzygium cumini L. Skeels) cultivars

Study site, experimental material and growth conditions

The experiment was carried out between 2018 and 2020 at Indian Council of Agricultural Research-Central Soil Salinity Research Institute located in Karnal, Haryana, India. The study area has a semi-arid climate with ~700 mm of annual rainfall. The experimental farm soils still suffer from moderate to high sodicity in the sub-surface soil. One-year old plants of jamun cultivars CISH J-37 (J-37), CISH J-42 (J-42), Goma Priyanka (GP), and Konkan Bahadoli (KB), obtained from ICAR-CHES, Godhra, Gujarat, were transferred to the experimental field in September, 2018 following acclimation. Soil analysis revealed considerable spatial and vertical differences in soil sodicity in terms of soil pH_s and exchangeable sodium percentage (ESP) (Hassani, Azapagic & Shokri, 2020). Soil pH_s was 7.81, 8.14, and 8.35 (average 8.09) in control and 8.45, 8.82, and 9.11 (8.80) in sodic treatment at 0–30, 30–60, and 60–100 cm depths, respectively. The corresponding ESP was 8.35%, 10.82%, and 16.58%, respectively, in control and 18.34%, 29.79%, and 39.60%, respectively, in sodic treatment. The scion shoots of all the cultivars were grafted on 1-year-old uniform rootstock plants raised by sowing the seeds from a single mother tree (~30 y) in polybags (25 cm × 10 cm) filled with soil and farm yard manure (3:1). Rootstock seedlings uniform in growth and having pencil thick stems were used in grafting using ~20 cm long healthy scion shoots (Singh et al., 2011). The plants were transplanted into auger-holes (diameter: 30 cm, depth: 120 cm) filled with the original soil without the addition of any soil amendment. A square system of planting was used, keeping between- and within-row distances of 6 m each. The young plants were trained properly to create a strong canopy framework. This was achieved by retaining on 3–5 well-spaced scaffold branches in different directions of the main stem above 60 cm from the ground surface. These branches served as the main framework for canopy development. Subsequent to planting, each tree was fertilized using 125 g N, 50 g P₂O₅ and 50 g K₂O in two equal splits. While half the dose was applied immediately after transplanting, the remaining half was applied in the ensuing June month (Singh et al., 2010). The irrigation water was applied into 1 m wide irrigation channels at fortnightly intervals during autumn-winters (October-February) and at weekly intervals during summers (March-June). The irrigation water had an electrical conductivity of 0.68 dS/m, pH of 8.04, Na⁺ of 2.43, K⁺ of 0.16, Ca²⁺ + Mg²⁺ of 4.14, Cl⁻ of 0.89 and HCO⁻₃ of 4.22 (me/l). The irrigation water was safe for irrigation, and unlikely to induce soil sodicity even after long-term use (Choudhary, 2017).

Soil properties

After about a year of planting, there were notable differences in tree growth, likely due to innate variations in soil properties (Marino et al., 2019; Peterson, Helgason & Ireson, 2019). Subsequently, twelve trees of each jamun cultivar were randomly selected for recording the spatial variations in soil pH, electrical conductivity, ESP, cation exchange capacity (CEC), and organic carbon (OC). Soil samples were collected using an auger from 0–30, 30–60, and 60–100 cm depths, approximately 50 cm from 2–3 directions of the trunk. The soils from the same depths were mixed to prepare the composite representative samples. Then, soil samples were air dried and sieved (2.0 mm sieve). Soil pH (pH_s) and saturation extract electrical conductivity (EC_e) were determined using digital pH and conductivity metres (Eutech, Singapore). The CEC and ESP were measured using the procedures described in Bhargava (2003). For CEC determination, the soils samples (5 g) were treated with 33 ml of 1N sodium acetate (pH 8.2) for replacing the exchangeable cations by Na⁺ ions. Ethanol was used to remove any excess sodium acetate. The adsorbed sodium was then replaced from the sample by extraction with three portions of 1N ammonium acetate. Following dilution to 100 ml, Na concentration was measured using a flame photometer (Systronics, India). The CEC was calculated using the formula:

$$\mathrm{C}\mathrm{E}\mathrm{C}\left(\mathrm{m}\mathrm{e}/100\mathrm{g}\right)={\displaystyle \frac{\mathrm{N}\mathrm{a}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\left(\mathrm{m}\mathrm{e}/\mathrm{l}\right)\mathrm{x}10}{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\left(\mathrm{g}\right)}}$$where CEC is the cation exchange capacity (milliequivalent per 100 g of soil), Na concentration refers to the Na⁺ concentration in milliequivalent per litre (me/l) from the standard curve, and soil sample (g) refers to the weight of soil sample (5 g) used.

For ESP measurement, 33 ml of 60% ethanol was poured gradually and allowed to leach. This procedure was repeated until the leachate’s electrical conductivity dropped below 40 micromhos/cm. The leachate was discarded. Following addition of three portions of 1N ammonium acetate (33 ml each) in a 100 ml volumetric flask, Na concentration was determined. The exchangeable sodium (me/100 g of soil) was calculated by the formula:

$$\mathrm{E}\mathrm{S}\left(\mathrm{m}\mathrm{e}/100\mathrm{g}\right)={\displaystyle \frac{\mathrm{N}\mathrm{a}\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{c}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\left(\mathrm{m}\mathrm{e}/\mathrm{l}\right)\mathrm{x}10}{\mathrm{s}\mathrm{o}\mathrm{i}\mathrm{l}\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}\left(\mathrm{g}\right)}}$$where, ES is the exchangeable sodium (milliequivalent per 100 g of soil), Na concentration refers to the Na⁺ concentration of extract in milliequivalent per litre from the standard curve, and soil sample (g) refers to the weight of soil sample (5 g) used.

Exchangeable sodium percentage (ESP; %) was computed using the formula:

$$\mathrm{E}\mathrm{S}\mathrm{P}\left(\mathrm{\%}\right)={\displaystyle \frac{\mathrm{E}\mathrm{S}\mathrm{x}100}{\mathrm{C}\mathrm{E}\mathrm{C}}}$$where ESP (%) is the exchangeable sodium percentage, ES refers to the exchangeable sodium (milliequivalent per 100 g of soil), and CEC stands for cation exchange capacity (milliequivalent per 100 g of soil).

Tree growth and leaf area

Six trees (n = 6) of each jamun cultivar within both control and sodicity treatments were tagged for recording the observations. Tree height, canopy spread and trunk diameter were recorded during the first week of September, 2020; after 2 years after planting. Trunk diameter was measured in the east-west (E-W) and north-south (N-S) directions 20 cm above the graft union using a digital caliper (Mitutoyo, Kanagawa, Japan). Trunk cross sectional area (TCSA) was calculated using the formula: TCSA = π(d/2)²; where d = mean of E-W and N-S trunk diameters. Canopy volume (CV) was calculated as: CV = (w² × h)/2; where w= canopy diameter in E-W and N-S directions, and h= tree height. Four fully developed leaves in the outer canopy of each replicate tree were tagged for measuring the unit leaf area using a handheld laser leaf area meter (CI-203; CID Bio-Science). There values were averaged and treated as a single replication in data analysis.

Leaf phenolics, relative chlorophyll and gas exchange traits

The leaf phenolics, including anthocyanins (anth) and flavonols (flav), relative chlorophyll, and gas exchange attributes, were measured on the same leaves that had previously been used for leaf area measurement. The leaf phenolics were measured using the Dualex 4 Scientific leaf-clip sensor (FORCE-A, Orsay Cedex, France). This leaf-clip sensor allows rapid and non-destructive measurements of leaf flavonols and anthocyanins. It first measures near-infrared chlorophyll fluorescence under a first reference excitation light not absorbed by polyphenols. It then measures the specific light absorbed by polyphenols including the green light for anthocyanins and ultraviolet light for flavonols. The dual-light excitation strategy enables quantification of polyphenols without damaging the leaves, and helps understand the changes in their levels in response to stress conditions (ForceA, 2019). A Soil Plant Analysis Development (SPAD) meter (SPAD-502Plus) was used to determine the relative leaf chlorophyll content. A portable infrared gas analyzer attached to a 6 cm² cuvette (LI-COR 6400 XT system; LI-COR, Lincoln, NE, USA) was used to assess the gas exchange attributes: net photosynthesis (P_n), transpiration (E), stomatal conductance (g_s), and internal CO₂ concentration (C_i). Before being enclosed in the leaf chamber, the leaves were wiped-off with a muslin cloth. A photosynthetic photon flux density of 1,000 µmol m⁻² s⁻¹, CO₂ concentration of 400 ppm, and a leaf temperature of 25 °C of were maintained during the measurements. The ratio of net photosynthetic to transpiration rate (P_n/E) was used to compute the instantaneous water use efficiency (WUE). These measurements were performed on 7 and 8 September, 2020 between 9.0 and 11.0 h.

Anti-oxidant enzymes and proline

Fresh leaf sample (300 mg) was homogenized in 0.1 M phosphate buffer (pH 7.5) supplemented with 5% (w/v) polyvinyl polypyrrolidone, 1 mM EDTA, and 10 mM b-mercapto-ethanol to assess ascorbate peroxidase (APX, EC 1.11.1.11) and superoxide dismutase (SOD, EC 1.15.1.11) activities. APX activity was measured subsequent to the oxidation of ascorbic acid (Nakano & Asada, 1981). The reaction mixture had 2.25 ml of 100 mM phosphate buffer (pH 7.0), 0.2 ml of 0.5 mM ascorbate, 0.2 ml of 0.1 mM H₂O₂, and 0.05 ml of the enzyme extract. Measurements using a UV-VIS spectrophotometer (Specord 210 Plus; Analytik Jena, Jena, Germany) revealed a drop in absorbance at 290 nm, indicating the oxidation of ascorbic acid. The enzyme activity was determined using ascorbic acid’s molar extinction coefficient of 2.8 mM⁻¹ cm⁻¹. One enzyme unit refers to one μ mole of ascorbic acid oxidized per minute at 290 nm. The ability of SOD to prevent nitro blue tetrazolium (NBT) from being reduced photochemically was used to measure its activity (Beauchamp & Fridovich, 1971). The reaction mixture (3.0 ml) consisted of 2.5 ml of 50 mM Tris-HCl (pH 7.8), and 100 μl each of 14 mM L-methionine, 10 µM NBT, 3 µM riboflavin, 0.1 mM EDTA, and the enzyme extract. The tubes were carefully shaken and positioned 30 cm below three 20 W fluorescent light bulbs (Phillips, India). Reaction was stopped after 40 min of incubation. The tubes were covered with black cloth to shield them from light. A non-irradiated colourless reaction mixture was used as control. The reaction mixture without enzyme extract developed a bright color and showed a decrease in absorbance following the addition of enzyme (Specord 210 Plus; Analytik Jena, Jena, Germany). The amount of enzyme needed to prevent one mmol of NBT from being photo-reduced was determined to be one unit of SOD. Catalase (CAT; EC 1.11.1.6) activity was determined using the procedure given in Aebi (1984). Reaction mixture comprised 0.5 ml of 0.1 M phosphate buffer (pH 7.0), 0.4 ml of 0.2 M hydrogen peroxide, and 0.1 ml of the diluted enzyme extract. The reaction was stopped by adding 3 ml mixture of glacial acetic acid (1:3 v/v) and potassium dichromate (5% w/v) after incubation at 37 °C for 3 min. The tubes were heated in a bath of boiling water for 10 min. A control was run under similar conditions. The absorbance of the test and control tubes was measured at 570 nm following their cooling (Specord 210 Plus). The absorbance of the test samples was subtracted from the absorbance of the control to calculate the amount of residual H₂O₂. One unit of enzyme activity refers to the enzyme quantity required to catalyze the oxidation of 1.0 µmole H₂O₂ per minute. The rate of guaiacol oxidation in the presence of H₂O₂ at 470 nm was used to measure the activity of peroxidase (POX, EC 1.11.1.7) (Rao, Paliyath & Ormrod, 1996). The reaction mixture consisted of 50 mM phosphate buffer (pH 6.5; 2.5 ml), 0.5% hydrogen peroxide (0.1 ml), 0.2% 0-dianisidine (0.1 ml), and the enzyme extract (0.1 ml). H₂O₂ (0.1 ml) was added to initiate the reaction. The reaction mixture devoid of H₂O₂ served as the blank. The change in absorbance was monitored at 430 nm for 3 min. One unit of peroxidase is represented by a change of one optical density in a minute. For proline extraction, the leaf tissue (200 mg) was homogenized in 10 ml of 3% sulphosalicyclic acid (Bates, Waldren & Teare, 1973). Acid-ninhydrin and glacial acetic acid (2 ml) were added to the extract (2 ml). The reaction was stopped in a water bath after 1 h of incubation at 100 °C. Subsequent to extraction using 4 ml of toluene, reaction mixture was rapidly stirred for 15–20 s. The absorbance was measured at 520 nm using toluene as blank (Specord 210 Plus, Germany).

Leaf ions

Leaf samples, dried to a constant weight at 60 °C in a hot air oven (NSW, India), were finely ground and 100 mg tissue sample was digested in nitric acid (HNO₃). Na⁺ and K⁺ were determined using a flame photometer (Systronics India), Cl⁻ using a chloride ion-selective electrode (Thermo Fisher Scientific, Mumbai, India), and Ca²⁺ and Mg²⁺ using an Atomic Absorption Spectrophotometer (Analytik Jena, Jena, Germany). All ion contents are in mg/g dry weight.

Statistical analysis

The experiment was laid out in a randomized block design with six replications (n = 6) within both control and sodicity stress treatments. Each replication consisted of a single biological replicate. The growth observations were recorded on these six trees (n = 6) of each cultivar under both control and sodic treatments. Leaf area, leaf phenolics, relative chlorophyll (SPAD) and gas exchange attributes were measured on four leaves of each replicate tree; their values were averaged and treated as a single replication in data analysis (n = 4). A two-way analysis of variance (ANOVA) was used to examine the independent and interaction effects of cultivar and sodicity on different traits. The means were compared by Tukey test (p < 0.05). Pearson’s correlations between the measured traits and the corresponding significance levels were computed. These analyses were performed using the MV App (Julkowska et al., 2019). Principal component analysis (PCA) was applied on replicated observations to discern the major patterns in the data (JAMOVI 2.3.28).

Membership function analysis

Membership function analysis, a comprehensive index based on multiple traits was used to rank the jamun cultivars for sodicity tolerance. This approach is particularly useful for assessing traits such as salt tolerance which are influenced by multiple traits and cannot be reliably measured using a single variable (Tian et al., 2024). This methodology uses membership functions based on the theory of fuzzy mathematics. Membership functions are used to define fuzzy sets like ‘salt stress’. Fuzzy logic accepts partial membership, assigning values between 0 and 1, in contrast to classical logic, which stipulates that an element is either a member (1) or not a member (0) of a set (Ross, 2004). The membership function values (MFV) for each measured trait were calculated using the sodicity tolerance coefficients (STCs). The STCs were calculated as the ratio of a trait’s value under sodicity stress to its value under control (non-sodic) conditions, expressed as percentage. Next, the MFV1 for each trait was calculated (Quamruzzaman et al., 2022). The traits inversely related with sodicity tolerance (leaf Na⁺ and Cl⁻) were measured using MFV2. Leaf Na⁺ and Cl⁻ are considered to have inverse relationships with sodicity tolerance; their exclusion and/or compartmentalization and thus their lower concentrations in leaves would improve tolerance to sodicity stress. To calculate MFV2, the MFV1 values for leaf Na⁺ and Cl⁻ were subtracted from 1. Each trait has its own MFV, ranging between ‘0’ and ‘1’. Higher salt tolerance is correlated with higher mean MFV, and vice versa. The MFVs of each measured trait were averaged to determine the mean MFV for jamun cultivars (Gyanagoudar et al., 2024). The sodicity tolerance levels of jamun cultivars were divided into five classes based on the mean value ( $\overline{\text{X}}$) and standard deviation (SD) of the mean MFV (X_i) at 90% and 68% confidence intervals (Z score = 1.64 and 1, respectively). These five classes included ‘highly tolerant’: X_i ≥ $\overline{\text{X}}$ + 1.64 × SD; ‘tolerant’: $\overline{\text{X}}$ + 1.64 × SD > X_i ≥ $\overline{\text{X}}$ + 1 × SD; ‘moderately tolerant’: $\overline{\text{X}}$ + 1 × SD > X_i > $\overline{\text{X}}$ − 1 × SD; ‘sensitive’: $\overline{\text{X}}$ − 1 × SD > X_i > $\overline{\text{X}}$ − 1.64 × SD; and ‘highly sensitive’: $\overline{\text{X}}$ − 1.64 × SD > X_i (Gyanagoudar et al., 2024).

Soil properties

The soil pH_s and ESP increased with depth, and were invariably significantly higher at different depths in sodic than in normal soils. The EC_e values were 0.58, 0.73 and 0.92 dS/m in control, and 0.66, 0.95 and 1.14 dS/m in sodic treatment in 0–30, 30–60, and 60–100 cm depths. The corresponding ESP was 8.35, 10.82, and 16.58%, respectively, in control and 18.34%, 29.79%, and 39.60%, respectively, in sodic treatment. The CEC was significantly lower in sodic than in control treatment across different depths. The organic carbon (OC) declined significantly with increase in sodicity across all the soil depths (Table S1). Pearson’s correlation analysis indicated significant positive relationships of CEC (r = 0.946) and ESP (r = 0.918) with the respective concentrations of Na⁺ measured during these determinations. Soil ESP had significant positive correlations with pH (r = 0.943), and negative correlations with CEC (r = −0.871) and OC (r = −0.914) (Table S2).

Tree growth

The effects of cultivar (C), sodicity (S), and their interaction (C × S) were significant (p < 0.05) on trunk cross sectional area (TCSA), canopy volume (CV), and leaf area (LA). Cultivars GP and KB had the highest (18.31 cm²), and the lowest (10.49 cm²) TCSA under control conditions. When compared to controls, sodicity stress reduced TCSA by 18.03, 31.29, 21.68, and 32.60% in J-37, J-42, GP, and KB, respectively. Under control conditions, CV was the largest (2.37 m³) in GP, and the smallest (0.77 m³) in J-42. Sodicity-induced reductions in CV were significant in cultivars J-42 (25.97%), GP (22.79%), and KB (32.60%). Sodicity stress caused significant reductions in LA only in cultivars J-37 (6.22%), GP (5.36%), and KB (8.19%) (Table 1).

Relative chlorophyll and gas exchange attributes

Compared to controls, sodicity-induced reductions in SPAD were relatively large (~17.0%) in J-42 and GP. Sodicity-induced declines in P_n were fairly similar in cultivars J-37 and J-42 (~18.0%). Cultivar GP exhibited the largest drop (37.16%) in P_n under sodicity stress. The tested cultivars displayed varying reductions in E [J-37 (28.42%), J-42 (21.34%), GP (37.23%), and KB (35.58%)] in response to sodicity. The g_s also declined in a similar fashion; decreases relative to controls were 30.77, 20.0, 47.06, and 42.86% in J-37, J-42, GP, and KB, respectively. Interestingly, there were indiscernible differences among the cultivars [J-37 (15.80%), J-42 (19.97%), GP (19.21%), and KB (16.31%)] for sodicity-induced reductions in C_i. While sodicity stress did not significantly affect WUE in J-42 and GP, cultivars J-37 (13.49%) and KB (21.91%) exhibited significant increases in WUE (Table 2).

Leaf flavonols, anti-oxidant enzymes and proline

Sodicity-induced upticks in anthocyanins (Anth) differed substantially among the cultivars; KB displayed the biggest increase (80.0%) while J-42 showed the lowest uptick (52.18%) in comparison to controls. Except for KB (0.55 units), the flavonol (Flav) levels in other tested cultivars were statistically similar (0.63–0.67 units) in control. While both J-37 (31.75%) and J-42 (29.23%) showed relatively less increases in Flav, GP (40.30%) and KB (70.91%) exhibited moderate and high increases in Flav, respectively, under sodic conditions (Table 3). While the constitutive levels of APX were hardly different, sodicity stress increased the APX activity by 47.09, 8.49, 106.33, and 68.18% in J-37, J-42, GP, and KB, respectively, relative to controls. The CAT levels varied considerably among cultivars under both control and sodicity treatments; GP (108.81%) and KB (6.46%) exhibited the highest and lowest spikes in CAT activity under sodic conditions (Table 3). These cultivars also showed the highest (63.09%) and the lowest (1.12%) increases in POX activity in response to sodicity stress. Sodicity-induced increases in SOD activity were remarkable only in J-37 (92.84%) and GP (119.75%), and rather weak in J-42 (2.12%). The tested cultivars also differed remarkably in leaf proline levels under sodic conditions; increases in proline were substantially greater in J-37 (202.34%) and GP (122.40%) than in J-42 (46.39%) and KB (50.71%) (Table 3).

Leaf ions

Leaf Na⁺ increased markedly in response to sodicity stress; cultivars KB (170.99%) and J-37 (86.36%) displayed the largest and smallest upticks in leaf Na⁺. Although leaf K⁺ dropped under sodicity stress, declines were rather subtle: 5.66% in J-37, 7.04% in J-42, 16.17% in GP, and 1.37% in KB. While J-37 exhibited modest increase (11.96%) in leaf Ca²⁺ when exposed to sodicity stress, leaf Ca²⁺ levels did not differ significantly between control and sodicity treatments in J-42, KB and GP (Table 4). Under sodicity stress, leaf Mg²⁺ decreased significantly in J-37 (9.75%), J-42 (27.11%), and KB (9.69%). Despite having the highest leaf Cl⁻ under control (1.91 mg/g DW), cultivar J-37 showed the lowest increase (33.51%) in leaf Cl⁻ under sodic conditions. In contrast, J-42 (131.82%), GP (172.09%), and KB (121.01%) showed noticeably larger increases in leaf Cl⁻ in response to sodicity stress. Sodicity stress caused significant reductions in leaf K⁺/Na⁺ ratio; varying between 49.55% (J-37) and 63.68% (KB) (Table 4).

Correlation analysis

Pearson’s bivariate correlations and the corresponding p-values are given in Table S2 and Fig. 1. TCSA had highly significant (p = 0.000) positive correlations with CV (r = 0.821), LA (0.549), P_n (0.625), and Ca²⁺ (0.638), and significant negative correlations with Anth (0.609), POX (0.721), SOD (0.626), and Cl⁻ (0.747). Similarly, CV had significant positive correlations with P_n (0.414), E (0.292), C_i (0.318), Ca²⁺ (0.445), and Mg²⁺ (0.437), and significant negative correlations with Anth (0.418), POX (0.358), SOD (0.734), and Cl⁻ (0.547). Presumably, trees with bigger trunk diameters can translocate more water and nutrients for the enhanced growth of leaves and canopies. Significant positive correlations of TCSA and CV with P_n were likely because more leaves and a bigger canopy in trees with larger trunk diameters increase the surface area for photosynthesis. Calcium and magnesium play structural and functional roles in the leaves, and their increased accumulation may support the metabolic demands of a larger canopy. The negative correlations of TCSA and CV with anthocyanins and antioxidant enzymes (POX and SOD) highlight trade-off between tree growth and stress-related secondary metabolism. Interestingly, both TCSA and CV exhibited stronger negative correlations with leaf Cl⁻ (0.747 and 0.547, respectively) than with leaf Na⁺ (0.439 and 0.282, respectively), implying that increased leaf Cl⁻ levels may more adversely impact jamun tree growth under sodic conditions. Leaf area (LA) correlated positively with P_n (0.460), WUE (0.433), K⁺ (0.582), Ca²⁺ (0.611), and K⁺/Na⁺ ratio (0.482). Larger leaves probably contribute more to photosynthesis because of their larger surface area for light interception, higher density of mesophyll cells, and increased gas exchange capacity. Both K⁺ and Ca²⁺ play diverse physiological functions in plants, including the maintenance of cell osmotic pressure which may explain their strong positive correlations with leaf area. LA showed significant negative correlations with flavonols, anti-oxidant enzymes, proline, and Na⁺ and Cl⁻; its correlations with Anth (0.544), Flav (0.485), CAT (0.547), and POX (0.552) were notably negative. Plants under stress often accumulate more secondary metabolites including anthocyanins, flavonols, and antioxidant enzymes at the expense of growth. While gas exchange attributes (P_n, g_s and C_i) had highly significant positive correlations with each other, WUE exhibited highly significant negative correlations with E (0.751), g_s (0.623), and C_i (0.523) (Table S2, Fig. 1).

Principal component analysis

The results of Bartlett’s test (χ² = 2,293.54, p < 0.001) and a reasonably high Kaiser-Meyer-Olkin score (0.754) suggested that PCA would efficiently reduce the dimensionality, producing distinct principal components. The first four principal components (Eigen value >1.0) accounted for 90.50% of the cumulative variance in data. The PC1 (Eigen value = 10.57, variance = 48.10%) was strongly correlated with gas exchange parameters (P_n, E, g_s and C_i), flavonols (Anth and Flav), APX, Na⁺ and Cl⁻. The PC2 (Eigen value = 4.16, variance = 18.90%) was largely a construct of SPAD, POX and Ca²⁺. While PC3 (Eigen value = 3.01, variance = 13.70%) had leaf K⁺ and WUE as the highly loaded variables, PC4 (Eigen value = 2.17, variance = 9.90%) had CV and leaf proline as the most prominent variables (Table S3). A glance at the PCA biplot indicated that the grouping of traits was largely consistent with their properties and functions. For instance, gas exchange traits, leaf K⁺ and K⁺/Na⁺ ratio were clustered in tandem (lower left quadrant). The placement of leaf flavonols, proline, antioxidant enzymes, and harmful ions along PC2 indicated that osmotic and oxidative stress defenses were probably triggered in response to excess Na⁺ and Cl⁻. PC1 was able to differentiate between the control and sodicity treatments fairly clearly. The fact that PC2 could distinguish KB from other cultivars in a measurable way implied divergent responses to sodicity stress unique to KB. Interestingly, while cultivars responded quite differently under control treatment, J-37, J-42, and GP seemed to possess some shared reactions to deal with sodicity stress (Fig. 2).

Ranking for sodicity tolerance

The membership function value (MFV) approach was used to analyze the sodicity tolerance of the tested cultivars. This approach provides a comprehensive assessment of cultivar’s ability to withstand stress conditions by considering multiple traits simultaneously. The MFV for each cultivar was calculated using the sodicity tolerance coefficients (STC) for all the measured traits (Table S4). It was found that none of the cultivars exhibited high tolerance, sensitivity, or extreme sensitivity to sodicity stress. With a mean MFV of 0.76, cultivar J-37 performed far better than others and was ranked as ‘tolerant’ to sodicity stress. The remaining cultivars were categorized as ‘moderately tolerant’ because their MFV values were all less than 0.50 (0.36 for J-42, 0.45 for GP, and 0.42 for KB).