Soybean [Glycine max (L.) Merrill] is a self-pollinated crop that belongs to the family Leguminosae and is believed to have originated in eastern Asia and is native to China. In India, soybean has surpassed other major oilseed crops and has been cultivated for centuries under different names (Dupare et al., 2008). In the 2023-24 crop season, global soybean production reached a record 399 million tons, representing a significant increase of 10.8% compared to the previous year’s 370 billion tons, with the US, Brazil, Argentina, China and India being the leading producers of soybean in the world. In India, soybean was cultivated on an area of 123.85 lakh hectares, with a total production of 130.62 lakh tonnes, with Madhya Pradesh, Maharashtra, Rajasthan, Karnataka and Telangana being the top fivesoybean-producing states in India. Despite technological advancements in agriculture, global food security is increasingly threatened by climate change, which exerts significant pressure on India’s limited land and water resources (Godfray et al., 2010). In India, about 147 million hectares of land are affected by soil degradation, of which nearly 23 million hectares suffer from salinity, alkalinity, or acidification, posing a major constraint to agricultural sustainability (Bhattacharyya et al., 2015). Around 6.74 million hectares are salt-affected, accounting for about 2.1% of the country’s geographical area, including 2.95 million hectares of saline soils and 3.77 million hectares of sodic soils. It is estimated that by 2050, nearly 50% of arable land will be salt-affected (Kumar and Sharma, 2020). In South Asia, around 52 million hectares of land are affected by salt (Mandal et al., 2018). Osmotic stress caused by salinity disrupts several physiological processes, including nutrient balance, membrane function and the ability to detoxify reactive oxygen species (Essa, 2002).
       
Soybean is considered moderately salt-sensitive crop. Elevated salt levels exert toxic effect on the embryo, adversely influencing several aspects of plant development. Salinity reduces germination percentage, root length, seedling growth and callus formation (Agnihotri et al., 2006; Munns and Tester, 2008 and Ramana et al., 2012). Screening of salt-tolerant genotypes for enhanced yield is a key objective for plant breeders and agronomists, but it is often costly and time-consuming. However, assessing salt tolerance at the early stages of plant development can significantly save time and reduce costs. This is particularly relevant if the degree of salt tolerance at the early stage of development is correlated with other stages of growth and yield (Hu et al., 2005). The productivity of crops, including soybean, is critically determined by germination and early growth stages. The germination stage is particularly sensitive to salinity and is often used for selecting salinity tolerance in plants (Cirka et al., 2021). Seeds that exhibit rapid and normal germination in saline conditions are expected to be tolerant to salinity (Bybordi and Tabatabaei, 2009). However, limited work has been performed to screen the existing soybean germplasm accessions as well as a few promising varieties for salinity tolerance. Therefore, the present study aimed to decipher the genetic relationship between selection indices such as the salt tolerance index (STI), yield, morpho-physiological traits and seed quality parameters. The outcomes will contribute to improving the efficiency of breeding programmes under saline environments and suggest an effective screening protocol for soybean.

Seeds of 156 soybean germplasm accessions, comprising one-year stored seeds of 150 accessions (Table 1) and six freshly harvested varieties with yellow seed coat colour, viz. Pusa 9712, Pusa 12, SL 958, PS 1347, PS 24 and SL 1074 were procured from the Division of Genetics, ICAR-IARI, New Delhi and evaluated for salinity studies during the kharif season of two consecutive years, 2022 and 2023. The experiment was conducted in the Division of Seed Science and Technology, ICAR-IARI, New Delhi. The 150 accessions were categorized into three groups based on seed coat colour, viz., yellow-seeded (105 accessions), black- seeded (42 accessions) and brown-seeded (3 accessions). The study was broadly divided into two groups: (i) freshly harvested six released varieties and (ii) one-year stored seeds of 150 accessions. Among the 105 yellow-seeded accessions, only 84 accessions were considered for further evaluation, as the remaining failed to germinate due to reduced viability after storage. The classification by seed coat colour was done to account for the known variability in seed longevity associated with pigmentation (Kuchlan et al., 2010).


       
A standardisation experiment was conducted under laboratory conditions on three randomly selected genotypes using two techniques, viz. top-of-paper (petri plate) method and standard germination testing using the between-paper method. Four different NaCl solutions with salinity levels of 0, 3, 6 and 9 dS/m were used for the experiment. For the top-of-paper method, 15 cm diameter sterile petri dishes were lined with Whatman No.1 filter paper and soaked with 10 ml of distilled water (control) and NaCl solutions (3, 6 and 9 dS/m). Each petri dish contained 20 uniform seeds of soybean cultivars, arranged equidistantly in a randomised design with three replications. To standardise the top-of-paper method, three techniques were evaluated for the first time in soybean: (i) alternate day application of 2.5 ml solution, (ii) a single application of 10 ml solution at the beginning and (iii) alternate day application of 10 ml solution until the end of the experiment.For germination testing using between-paper method, the International Seed Testing Association (ISTA) recommended procedure was followed. Germination paper was soaked with distilled water (control) and NaCl solutions (3, 6, 9 dS/m) and seeds were placed between folds of paper before rolling.To assess the response of the soybean cultivars under different NaCl concentrations, seeds were placed on germination paper and the effect of salinity on various seed quality parameters viz. germination percentage, seedling length (cm), seedling fresh weight (mg) and seedling dry weight (mg) was observed on the final day of germination. Seedling vigour indices were calculated as per the standard procedure of Abdul-Baki and Anderson (1973) for all four salinity treatments. Seedling Vigour Index-I (SVI-I) was obtained by multiplying the germination percentage with the mean of total seedling length (cm) and Seedling Vigour Index-II (SVI-II) was determined by multiplying the germination percentage with the mean seedling dry weight (mg). Salt Tolerance Index (STI) was calculated as the ratio of total plant (shoot + root) dry weight obtained from 10 seeds grown under different salt concentrations (3, 6 and 9 dS/m NaCl) to the total plant dry weight obtained under control conditions (0 dS/m NaCl) (Khatun et al., 2013).

Standardization of protocol for salinity stress evaluation
 
The top-of-paper method was found superior to the between-paper method for evaluating seed quality parameters under varied salinity stress levels (0 to 9 dS/m), due to several advantages,including ease of handling, precise application of salt solutions at required concentrations and practical in execution. Among the different protocols tested for the top-of-paper method, the addition of 2.5ml NaCl solution every alternate day was proved to be most effective, maintaining adequate moisture without water stagnation during germination testing.

Analysis of variance for seedling quality parameters
 
The statistical analysis (ANOVA) revealed that treatment, accession and their interaction exerted highly significant effects (p<0.001) on all measured traits, including germination percentage, seedling length, fresh and dry weight (Table 2). These results underscore the strong influence of salt stress on seedling performance.The significant accession ×  treatment interaction further indicates that responses to salinity vary widely among accessions, reflecting a complex genetic basis for salt tolerance and emphasizing the role of both genetic background and environmental conditions for stress responses.


       
Germination percentage declined progressively with increasing salinity across all genotype groups (Fig 1A-1D). At 0 dS/m (control), black-seeded accessions recorded a mean germination of 75.51%, which declined to 67.81% at 3 dS/m, 52.77% at 6 dS/m and 36.46% at 9 dS/m. In contrast, yellow-seeded accessions recorded markedly lower germination values of 32.32%, 27.11%, 15.06% and 7.23% at 0, 3, 6 and 9 dS/m, respectively confirming that genotypes with yellow seed coat color were poor storers compared to black-seeded genotypes. For brown-seeded accessions, germination percentages were 68.33%, 56.66%, 38.33% and 23.33% at 0, 3, 6 and 9 dS/m, respectively. Similarly, for released varieties, values were 91.66%, 86.66%, 66.66% and 45.00% at corresponding salinity levels. Variation in germination percentage among genotypes may be attributed to differences in genetic makeup, metabolic activity and enzyme regulation (Bojovic et al., 2010; Kondetti et al., 2012 and Kandil et al., 2016).


       
A similar trend was observed for seedling length, fresh weight and dry weight, all of which decreased linearly with increasing salinity from 0 to 9 dS/m. This reduction may be attributed due to the osmotic stress limiting water uptake, restricted elongation and reduced enzyme activity involved in reserve mobilization from seeds to seedlings (Kandil et al., 2012; Kondetti et al., 2012; Ayed et al., 2014 and Tesfaye et al., 2014). Interestingly, few genotypes recorded highest germination percentage, seedling length and shoot and root dry weights at 3 dS/m, as low concentrations of NaCl may act as a nutrient, improving water uptake, promoting elongation, enhancing ion uptake, or facilitating better reserve mobilization under mild stress (Essa, 2002; Tamura and Chen, 2009 and Gogile et al., 2013). Both Seedling Vigour Index-I (SVI-I) and Seedling Vigour Index-II (SVI-II) followed the same declining trend with increasing salinity across all genotypic groups. The variation in vigour indices among soybean genotypes reflected differential genotypic responses to salinity stress, with the lowest values of SVI-I and SVI-II recorded at the highest salinity level (9 dS/m).
 
Formulation of screening criteria for selection of tolerant and susceptible genotypes
 
Earlier studies have reported that seedling dry weight, which is the ultimate result of all physiological and biochemical activities in plant, is markedly reduced under high salt concentrations (Akbarimoghaddam et al., 2011 and Rumena, 2006). Physiological and seed quality parameters, along with their relationship to salt tolerant indices, have been proved as an applicable tool for selecting salt-tolerant genotypes (Kagan et al., 2010; Khatun et al., 2013 and Putri et al., 2022). For black seeded genotypes, the present study revealed a progressive decline in the Salt Tolerance Index (STI) with increasing salinity (0–9 dS/m). At 6 dS/m, the highest STI value was 90.19% in accession no. 80, while the lowest was 36.43% in accession no. 71. At 9 dS/m, accession no. 3 recorded the highest STI (71.69%), whereas accession no. 9 recorded the lowest (23.81%) (Fig 2A). Similarly, for yellow and brown seeded accessions, the STI averaged across all genotypic treatments decreased with an increase in the salinity level from 0 - 9 dS/m (Fig 2B and 2C). Among freshly harvested varieties, the highest STI value at 6 dS/m was 74.55% in variety Pusa 9712, while the lowest was 56.78% in SL 958. At 9 dS/m, Pusa 9712 again recorded the highest value (48.60%), whereas SL 958 had the lowest (28.50%) (Fig 2D).


       
Hence, based on the significant correlation of STI with several seed quality traits, genotypes were categorized into three groups: highly tolerant, moderately tolerant and highly susceptible. At salinity level of 6 dS/m, seed quality parameters (germination percentage, SVI-I, SVI-II and related traits) showed a positive correlation with STI for both black-seeded accessions and released varieties (Fig 3 and 4), whereas in yellow-seeded accessions, significant correlations were recorded at 3 dS/m (Fig 5).






       
In contrast, no significant correlations were detected for brown-seeded accessions at any salinity level; therefore, no categorization was made for this group. Further selection of tolerant and susceptible genotypes within the three groups: black-seeded accessions, yellow-seeded accessions and released varieties was carried out based on STI values at their respective salinity levels. In total, 20 genotypes were selected: 7 black-seeded, 7 yellow-seeded and all 6 released varieties. The selection was carried out at 3 dS/m for yellow-seeded accessions and at 6 dS/m for black-seeded accessions and released varieties. The calculated STI values were then arranged in increasing order and the genotypes were classified into three categories based on their ability to withstand salinity stress: highly tolerant, moderately tolerant and highly susceptible (Table 3 and Fig 6-8).






 
Correlation analysis of seed quality traits and salt tolerance index
 
Principal component analysis (PCA)
 
PCA (Fig 9) revealed that the first principal component explained 79% of the variance, largely driven by shoot length, root length, total seedling length and seedling dry weight. The second component explained an additional 8.8%. The clustering of these traits indicates their interrelated roles in salinity tolerance. Control treatments clustered separately from salinity treatments, reflecting baseline physiological responses. Similar clustering of stress-related traits has been reported in drought and salinity studies (Lee et al., 2017; Gupta et al., 2020).These results suggest that PCA is effective in identifying key traits contributing to variation among genotypes under salinity stress. Traits that maintain their position closer to the control across increasing salinity levels might be more tolerant and hence valuable for breeding programs.


 
Multivariate selection using MGIDI
 
The Multi-Trait Genotype-Ideotype Distance Index (MGIDI) was employed to rank genotypes under control and salinity stress. By integrating multiple seedling traits, MGIDI enhanced selection efficiency and identified genotypes resilient under both optimal and stress conditions. A total of 156 genotypes were assessed for seed quality parameters under both normal and salt stress conditions for selecting the genotypes with high vigour. This approach identified eight superior soybean genotypes in each condition, highlighting varieties such as IC-81655, DS-76-35-4, Pusa -12, Pusa-9712, PS-1347, Himso-1681, DS-9703 and EC-14427 under control conditions and Pusa-9712, PS-1347, HIMSO-1574, Pusa-12, J-231, DS-76-37-2, EC-34141 and PS-1503 under salt stress (Fig 10). Genotypes near the cut-off threshold also displayed potential for tolerance and for further evaluation.

The present study standardised a reliable protocol for salinity stress evaluation in soybean, with the top-of-paper method using alternate-day application of 2.5 ml NaCl solution. Salinity stress significantly reduced germination, seedling length, vigour indices and biomass traits, with greater reduction at higher salinity levels. The Salt Tolerance Index (STI) emerged as a robust criterion for categorising genotypes into highly tolerant, moderately tolerant and highly susceptible groups. Black-seeded accessions and released varieties correlated strongly with STI at 6 dS/m, while yellow-seeded accessions showed significant associations at 3 dS/m and brown-seeded accessions showed no consistent response. Multivariate selection using MGIDI improved efficiency in identifying superior genotypes under saline conditions. Overall, the study demonstrated wide genetic variability for salinity tolerance in soybean. The identified tolerant genotypes provide valuable material for breeding programs. This integrated approach offers a practical framework for improving soybean resilience under saline environments.

The authors are grateful to the Division of Genetics, ICAR-IARI, New Delhi, for providing the soybean genotypes for executing the present research. The authors also acknowledge the support by The Graduate School, ICAR-IARI and ICAR, Department of Agricultural Research and Education, Government of India for facilitating the present research.

The authors declare that there is no conflict of interest regarding the publication of this article. No funding or sponsorship influenced the design of the study, data collection, analysis, decision to publish, or preparation of the manuscript.