Characterization of a Soybean ( Glycine max L. Merr.) Population for Germination and Seedling Root Traits under Water Stress

: Dry soil conditions at soybean planting results in poor stand establishment, which often necessitates replanting. We conducted a study to identify soybean genotypes that can maintain germination rates and possess better root morphology under water stress. We tested 41 Plant Introductions (PI) for germination and seedling root traits under controlled environmental conditions at ﬁve water potentials: 0.00, − 0.27, − 0.54, − 0.82, and − 1.09 MPa (no, low, mild, severe, and extreme water stress, respectively). The same genotypes were tested for emergence and seedling root traits under ﬁeld conditions in South Carolina (2021 and 2022) and North Carolina (2022). Among the 41 genotypes evaluated, PI 398566 and PI 424605A maintained higher germination percentages ( ≥ 63%) under water stress. The same genotypes were ranked among the top 15 genotypes for root traits (total-root and ﬁne-root (diameter between 0.25 and 0.50 mm) length, surface area, and/or volume) under water stress. Furthermore, they had relatively higher emergence percentages under ﬁeld conditions ( ≥ 35% under dry soil conditions). The superior genotypes identiﬁed in this study (PI 398566 and PI 424605A) that had better germination and root morphology under water-stress and no-stress conditions and better emergence would be useful for developing varieties with drought tolerance during the emergence phase. grown at Florence, SC in 2021 and 2022 and Jackson Springs, NC in 2022. Emergence percentage was measured at 7 days after planting (DAP) at both locations. Root traits were measured at Florence at 14 DAP in 2021 and at 7 and 14 DAP in 2022. Root traits were not measured at Jackson Springs.


Introduction
Soybean (Glycine max L. Merr.) is an important legume crop globally with Brazil (38% of total global production in 2021), USA (31%), and Argentina (13%) being the major producers [1]. Soybean is widely grown as an oilseed and a protein source worldwide [1][2][3]. Because of its high nutritional quality and protein content, soybean is a raw material in a variety of food products and is used as an important feed material. Recently, soy foods and their bioactive compounds have received significant attention because of their human health benefits, and they are also known as the 'functional foods of the century' [4]. In USA, soybean is the second most planted field crop after maize (Zea mays L.) [1], and the second most important crop that contributes to total crop cash receipts [5]. Changing precipitation patterns and rising temperatures create significant challenges to soybean production, especially under rainfed production conditions worldwide [6,7].
Soybean is the most drought-sensitive species among the legumes [8]. Drought stress negatively affects the growth and metabolism of the shoot and root tissues in soybean [9,10]. The soybean crop is most sensitive to drought stress at two stages: at emergence and from flowering through pod-fill [11][12][13]. Limited knowledge is available about the effects of water stress during the emergence phase as most drought research has been conducted Thresher. In between each genotype, the combine was cleaned using an air compressor. After harvest, the soybean seeds were cleaned and stored in a cold room at 7.2 • C and 50% relative humidity and maintained around 12% moisture content until use. The seed quality of all genotypes was determined using the protocol given in the 'Uniform Soybean Tests-Southern States' [36]. Visual ratings were used to assess seed quality according to the amount and degree of wrinkled, cracked, greenish, or moldy seeds, using a scale of 1 = very good, 2 = good, 3 = fair, 4 = poor, and 5 = very poor (only 'very good'-quality seeds were used for the present study). The ragdoll method was used to determine the germination rate of each genotype [37]. Briefly, non-toxic paper toweling was obtained from the Seedburo Equipment Company. The paper towel was moistened and placed on a clean, flat surface and 100 seeds were placed on one half of the towel. The towel was folded over and rolled into a moderately tight tube. The rolled-up ragdoll was then placed upright in a plastic container with no lid. The cylinder was placed in the greenhouse, which was kept between 23 and 30 • C. After five days, the towel was unrolled and the number of seeds that germinated was counted. The germination rate of genotypes varied between 92 and 98%.

Laboratory Experiment to Measure Seed Germination
To measure seed germination, a laboratory experiment was conducted following the "Rules for Testing Seeds" of the Association of Official Seed Analysts (AOSA) [38]. Clean, healthy, and undamaged seeds of each genotype were selected to test for germination. The selected seeds were first disinfected by treating them with 2% (v/v) sodium hypochlorite solution for 10 min. After that, seeds were rinsed with deionized water and airdried on filter papers. Twenty seeds of each genotype were placed on top of two layers of Whatman No.1 filter paper in a sterile petri dish with an internal diameter of 9 cm. Seed germination was tested under five different water potentials: 0.00, −0.27, −0.54, −0.82, and −1.09 MPa using polyethylene glycol (PEG) 6000 at a concentration of 0, 5, 10, 15, and 20% (w/v), respectively [39]. Deionized water was used to prepare the PEG solutions. Zero percent PEG concentration represented the control treatment (no-water stress) for which just 10 mL of deionized water was added to petri dishes containing seeds. Similarly, for the other water stress treatments [−0.27 MPa (low water stress), −0.54 MPa (mild water stress), −0.82 MPa (severe water stress), and −1.09 MPa (extreme water stress)], 10 mL of corresponding PEG solutions was added to respective petri dishes containing the seeds. There were three Petri dishes (replications) per genotype in each water stress treatment; thus, a total of 60 seeds were sampled per genotype in each water stress treatment. Petri dishes with seeds were placed at 25 • C in an incubator under dark conditions. Petri dishes were arranged in an incomplete block design. Seeds with a radicle length, ≥2 mm were considered germinated [39]. Each petri dish was scored for germination from day 3 to day 7 after placing the seeds in it. Germination percentage, germination potential [40], and germination rate index [41] were calculated as given below: Germination percentages = (Number of seeds germinated in 7 days)/(Total number of seeds) × 100 Germination potential (%) = (Number of seeds germinated in 3 days)/(Total number of seeds) × 100 Germination rate index (GRI) = G1/1 + G2/2 + . . . + G7/7; where G1 is the germination percentage on day 1, G2 is the germination percentage on day 2, and so on.

Greenhouse Experiment to Measure Root Morphological Traits
A greenhouse experiment was conducted to measure root morphological traits of soybean genotypes under different levels of water stress. Thirty seeds of each genotype were germinated in a Petri dish containing Whatman No.1 filter paper, moistened with 0, 5, 10, 15, or 20% (w/v) PEG 6000 solution (corresponding to water potentials of 0.00 MPa (no water stress), −0.27 MPa (low water stress), −0.54 MPa (mild water stress), −0.82 MPa Agronomy 2022, 12,1944 4 of 17 (severe water stress), and −1.09 MPa (extreme water stress)). Deionized water was used to prepare the PEG solutions. Petri dishes with seeds were placed at 25 • C in an incubator under dark conditions. Germinated seeds with a radicle length, ≥1 cm were carefully placed in opaque plastic cups (0.5 L) containing full-strength Hoagland's solution with PEG 6000 at one of the five concentrations (0, 5, 10, 15, and 20% (w/v)) [42,43]. Three seedlings of the same genotype were grown in each plastic cup. The plastic cups were arranged in a split-plot design in a greenhouse with water stress as the main plot factor and genotype as the sub-plot factor. There were four cups (replications) per genotype in each water stress level. The environmental conditions in the greenhouse included daytime maximum/night-time minimum temperatures of 30/20 • C, 13 h light/11 h dark, ambient CO 2 levels (414 ppm), and relative humidity of 70%. Fourteen days after starting the seeds, one of the three seedlings (subsamples) in each cup (replication) was randomly chosen to measure root traits, and its root system was gently separated from the shoot using sharp scissors. Each root system was washed with distilled water to remove the growth solution and then blotted with tissue paper [44]. The cleaned roots were scanned using an Epson Perfection V600 scanner at 300 dpi resolution (Epson, Long Beach, CA, USA) [45,46]. The scanned images were analyzed using WinRHIZO Pro image analysis system (Regent Instruments, Inc., Quebec City, QC, Canada) to estimate the length, surface area, and volume of the whole root system and that of fine roots (diameter <0.25 mm or between 0.25 and 0.5 mm) [ The soil type at Florence was Emporia loamy fine sand (Fine-loamy, siliceous, sub-active, thermic Typic Hapludults) and Bonneau sand (Loamy, siliceous, sub-active, thermic Arenic Paleudults) in the experimental fields used in 2021 and 2022, respectively. The soil type in the experimental field at Jackson Springs was Candor sand (Sandy, kaolinitic, thermic Grossarenic Kandiudults). The previous crop in the 2021 experimental field at Florence was corn during the summer seasons in 2020 and 2019. The same field was covered by a wheat (Triticum aestivum L.) cover crop during the winter seasons in 2020-2021 and 2019-2020. The previous crop in the 2022 experimental field at Florence was winter wheat in 2020. In the same field, no crop was grown in 2021. At Jackson Springs, cotton and corn were grown during the summer seasons in 2020 and 2021, respectively, and no cover crops were grown during the winter seasons in both years. A soil test was conducted prior to planting in all experimental fields. The results of soil analysis are presented in Table S2. At Florence, no fertilizer was added to the experimental fields in 2021 and 2022. At Jackson Springs, K 2 O and S were applied at a rate of 112 kg ha −1 and 28 kg ha −1 in 2022 as recommended. The land was strip-tilled using a disk plow on the day of planting at Florence in 2021 and 2022. At Jackson Springs, the land was prepared using a disc harrow five days before planting in 2022. At Florence, soybean genotypes were planted on May 21st in 2021 and May 31st in 2022 using a four-row dynamic disc planter (Wintersteiger, Salt Lake City, UT, USA). At Jackson Springs, soybean genotypes were planted on 7 June 2022 using a four-row cone plot planter (Almaco, Nevada, IA, USA). Planting depth was about 4 cm in all fields. Each genotype was planted in four plots (four replications). Plots were arranged in a randomized complete block design. Each plot was of 3 m × 3 m size and consisted of four rows. Row length was 3 m and row spacing was 0.76 m. A total of 100 seeds were sown in each row.
In the 2021 experimental site at Florence, the cumulative precipitation for a 21-day period prior to planting was 3.4 cm, which was just half that compared to historical precipitation data (30-year climate normal) ( Figure S1). No rain was recorded in the 7-day period after planting ( Figure S1). A total of 5 cm of rain was recorded in the 7-day period  Figure S1). In the 2022 experimental site at Florence, a total of 1.9 cm of rain was recorded in the last 8 d prior to planting. In the same site, a total of 0.84 cm rain was recorded within 3 d after planting (DAP), but no rain was recorded thereafter until the emergence was counted at 7 DAP. At Jackson Springs in 2022, no rain was recorded in the last 9 d prior to planting, however, a total of 1.12 cm rain was recorded within 5 DAP. No irrigation was provided in any fields as we wanted to measure emergence and root morphology under rainfed conditions. Weeds were controlled through hand-weeding whenever needed. No pest or pathogen problems were observed during the experiments.

Measurement of Seedling Emergence and Root Morphology
A seedling was considered to have emerged when cotyledons were visible over the soil surface [48]. Emerged seedlings in each of the four rows in each plot were counted at 7 DAP. In each plot (where a total of 400 seeds were sown), emergence was estimated as the number of emerged seedlings divided by 400 and was expressed in percentage.
Root morphology of the seedlings was measured at Florence at 14 DAP in 2021 and at 7 and 14 DAP in 2022. Root morphology was not measured at Jackson Springs. To measure root morphology, root systems of 10-15 consecutive seedlings were excavated using a shovel from the second row of each plot to a depth of 25 to 30 cm and a distance of 25 to 30 cm on either side of the row [49]. The excavated root system was carefully shaken to remove adhering soil. After that, three intact root systems from each plot were selected for further analysis. The selected root systems were washed and placed between two paper towels before storing them at 4 • C in sealed Ziploc bags (S.C. Johnson & Sons, Inc. Racine, WI, USA). Afterward, the root systems were gently separated from the shoots using sharp scissors, and each of the three root systems was scanned separately using an Epson Perfection V600 scanner (Epson, Long Beach, CA, USA) at 300 dpi resolution [45,46] in 2021 and at 800 dpi resolution in 2022. The scanning process was completed within three days after root excavation. The scanned root images were analyzed using WinRHIZO Pro image analysis system to estimate the length, surface area, and volume of the whole root system and that of fine roots (diameter <0.25 mm or between 0.25 and 0.5 mm). Root traits of all three seedlings (subsamples) in each plot (replication) were measured and averaged to get a single value per replication.

Statistical Analyses
A statistical model was developed for germination, emergence, and root trait data including genotype, water stress level, and the interaction as fixed effects, and replication as a random effect. Analysis of variance to test the effects of genotype, water stress level, and the interaction was performed using the GLIMMIX procedure in SAS (Version 9.4, SAS Institute, Cary, NC, USA). There were three replications per genotype in each water stress treatment for germination percentage, germination rate index, and germination potential measured under laboratory conditions; four replications per genotype in each water stress treatment for root morphological traits measured under greenhouse conditions, and four replications per genotype for emergence percentage and root morphological traits measured under field conditions. The probability threshold level (α) for statistical significance was set at 0.05. Separation of least-squares means was performed using Fisher's LSD test. Ward's hierarchical cluster analysis was performed using the JMP program (Version 16.0.0, SAS Institute, Cary, NC, USA) to cluster the genotypes based on germination traits at different water stress levels in the laboratory experiment and emergence percentage under field conditions.

Germination and Emergence of Soybean Genotypes
Water stress, genotype, and their interaction had significant effects on germination percentage, germination potential, and germination rate of soybean genotypes evaluated in the laboratory experiment (Table 1). Water stress significantly reduced soybean germination  (Figures 1 and S3 and Table S3). Germination percentage was reduced by 53 to 97% under extreme water stress of −1.09 MPa, compared to no-water stress conditions, except for genotypes PI 398566, PI 398532, and PI 424605A (Table S3). Genotypes PI 398566, PI 398532, and PI 424605A demonstrated only 7, 32, and 37% reductions in germination percentage under extreme water stress. Five genotypes (PI 398237, PI 398201, PI 407848, PI 408169D, and PI 507382) did not germinate under extreme water stress. At −0.82 MPa of water potential (severe water stress), the germination percentage was less than 60% for half of the genotypes. However, genotypes PI 398319, PI 398566, PI 424605A, and PI 442012B demonstrated >95% germination under −0.82 MPa of water potential. Genotypes PI 398532, PI 398566, PI 424605A, which had better germination percentages under extreme and severe water stress levels, also possessed high values for germination percentage under mild, low, and no water stress conditions (−0.54, −0.27, and 0.00 MPa, respectively) ( Figure 1). Germination of genotypes PI 398532, PI 398566, and PI 424605A were 95, 98, and 98%, respectively, under mild water stress and 95, 100, and 100%, respectively, under low and no water stress (Table S3).
Genotypic differences were significant for emergence percentage at Florence, SC and Jackson Springs, NC ( Table 2). The emergence percentage of genotypes ranged between 4 and 39 at Florence in 2021, 2 and 65 at Florence in 2022, and 23 and 77 at Jackson Springs in 2022 ( Figure 2 and Table S3). Genotypes PI 398566, PI 398532, and PI 424605A, which demonstrated a higher germination percentage under severe and extreme water stress (−0.82 and −1.09 MPa, respectively) in the laboratory experiment, also exhibited a relatively greater emergence percentage (≥35% at Florence and ≥54% at Jackson Springs) under field conditions (Table S3). Similarly, genotypes PI 398237 and PI 603171, which demonstrated poor germination percentage under severe and extreme water stress in the laboratory experiment, also exhibited a poor emergence percentage (≤4% at Florence and ≤22% at Jackson Springs) under field conditions (Table S3).        (Table S3). A similar trend was observed when the genotypes were clustered based on germination index and germination potentials (both traits demonstrate the germination rigor) under laboratory conditions (Figure 3b,c).   Table S1 for other genotype names corresponding to the numbers. The color bars right to the dendrograms represent values for the traits presented in the dendrograms, with colors changing from blue (lowest) to red (highest).

Root Morphological Traits of Soybean Genotypes at the Seedling Stage
The main effect of water stress was significant on all measured root traits (length, surface area, and volume of the total root system, fine roots with diameter <0.25 mm, and fine roots with diameter between 0.25 and 0.5 mm) under controlled environmental conditions (Table 1). Similarly, the main effect of genotypes was significant on all root traits except total root volume under controlled environmental conditions (Table 1). However, the 'water stress-by-genotype' interaction effect was not significant on any root traits. This suggests that the relative ranking of genotypes for root traits was similar at different levels of water stress (no stress, low, mild, severe, and extreme). Under field conditions at Florence, SC, genotypic differences were observed for total root surface area and total root volume ( Table 2).
Under controlled environmental conditions, the magnitude of all measured root traits decreased with increasing water stress (Figure 4). When genotypes were ranked based on the numerical values of root traits under extreme and severe water stress conditions, PI 424605A and PI 398566 that showed better germination under water stress conditions (extreme, severe, mild, and low stress) were ranked among the top 15 (Table S4). Between those two genotypes, PI 424605A was superior; it was ranked among the top three under extreme water stress conditions and among the top 10 under severe water stress conditions for all measured root traits. PI 398532, another genotype with better germination under water stress conditions (extreme, severe, mild, and low stress), was ranked among the top 15 for all measured root traits under extreme water stress conditions.  Table S1 for other genotype names corresponding to the numbers. The color bars right to the dendrograms represent values for the traits presented in the dendrograms, with colors changing from blue (lowest) to red (highest).

Root Morphological Traits of Soybean Genotypes at the Seedling Stage
The main effect of water stress was significant on all measured root traits (length, surface area, and volume of the total root system, fine roots with diameter <0.25 mm, and fine roots with diameter between 0.25 and 0.5 mm) under controlled environmental conditions (Table 1). Similarly, the main effect of genotypes was significant on all root traits except total root volume under controlled environmental conditions (Table 1). However, the 'water stress-by-genotype' interaction effect was not significant on any root traits. This suggests that the relative ranking of genotypes for root traits was similar at different levels of water stress (no stress, low, mild, severe, and extreme). Under field conditions at Florence, SC, genotypic differences were observed for total root surface area and total root volume ( Table 2).
Under controlled environmental conditions, the magnitude of all measured root traits decreased with increasing water stress (Figure 4). When genotypes were ranked based on the numerical values of root traits under extreme and severe water stress conditions, PI 424605A and PI 398566 that showed better germination under water stress conditions (extreme, severe, mild, and low stress) were ranked among the top 15 (Table S4). Between those two genotypes, PI 424605A was superior; it was ranked among the top three under extreme water stress conditions and among the top 10 under severe water stress conditions for all measured root traits. PI 398532, another genotype with better germination under water stress conditions (extreme, severe, mild, and low stress), was ranked among the top 15 for all measured root traits under extreme water stress conditions. The correlation analysis between the germination percentage and root traits under controlled environmental conditions showed that no root traits were correlated with germination percentage in the absence of water stress (Figure 5a). However, with the increasing severity of water stress, the correlation between root traits and germination became more prominent. For example, under low water stress (−0.27 MPa), germination was correlated with length, surface area, and volume of fine roots with diameter between 0.25 and 0.5 mm. Under mild water stress (−0.54 MPa), germination was correlated with all measured root traits, except total root volume. However, under low and mild water stress, the values of the correlation coefficient among root traits and germination were always less than 0.40. Under severe and extreme water stress (−0.82 or −1.09 MPa, respectively), germination was correlated with all measured root traits with a correlation coefficient greater than 0.40 in most cases. The strongest correlation of germination with root traits was observed under severe water stress (correlation coefficient ≥0.60). Though the values of correlation coefficient between germination and any root trait steadily increased with increasing severity of water stress, they decreased under extreme water stress (Figure 5a). This may be because of the apparent damage of water stress on germination and root morphology when the stress intensity is too high. We also evaluated the relationship of root traits with the emergence percentage of the soybean genotypes under field conditions at Florence, SC in 2021 and 2022 (Figure 5b). We found that fine root (diameter between 0.25 and 0.5 mm) length had a positive relationship with emergence percentage (correlation coefficient ≥0.32) in both years.  The correlation analysis between the germination percentage and root traits under controlled environmental conditions showed that no root traits were correlated with germination percentage in the absence of water stress (Figure 5a). However, with the increasing severity of water stress, the correlation between root traits and germination became more prominent. For example, under low water stress (−0.27 MPa), germination was correlated with length, surface area, and volume of fine roots with diameter between 0.25 and 0.5 mm. Under mild water stress (−0.54 MPa), germination was correlated with all measured root traits, except total root volume. However, under low and mild water stress, the values of the correlation coefficient among root traits and germination were always less than 0.40. Under severe and extreme water stress (−0.82 or −1.09 MPa, respectively), germination was correlated with all measured root traits with a correlation coefficient greater than 0.40 in most cases. The strongest correlation of germination with root traits was observed under severe water stress (correlation coefficient ≥0.60). Though the values of correlation coefficient between germination and any root trait steadily increased with increasing severity of water stress, they decreased under extreme water stress (Figure 5a). This may be because of the apparent damage of water stress on germination and root morphology when the stress intensity is too high. We also evaluated the relationship of root traits with the emergence percentage of the soybean genotypes under field conditions at Florence, SC in 2021 and 2022 (Figure 5b). We found that fine root (diameter between 0.25 and 0.5 mm) length had a positive relationship with emergence percentage (correlation coefficient ≥0.32) in both years. Agronomy 2022, 12, x FOR PEER REVIEW 11 of 17

Discussion
Significant genetic variability was observed in germination percentage across all water stress levels under laboratory conditions and emergence percentage under field conditions. Overall, based on germination percentage under no, low, mild, severe, and extreme water stresses (0.00, −0.27, −0.54, −0.82, and −1.09 MPa) and emergence percentage under three environments (Florence, SC in 2021 and 2022 and Jackson Springs, NC in 2022), PI 398566, PI 424605A, and PI 398532 were the best genotypes, and PI 398237, PI 603171, PI 408169D, PI 408175, and PI 398704 were the poor genotypes. Among those genotypes, PI 398566 and PI 424605A were the top two, and PI 398237 and PI 603171 were the bottom two according to ranking based on numerical values of the above traits. A principal component analysis (PCA) using germination percentage at different water stress conditions in the laboratory experiment and emergence percentage under field conditions confirmed these results (PI 398566 and PI 424605A as the best and PI 398237 and PI 603171 as the worst) ( Figure 6). On the PCA biplot, PI 398566 and PI 424605A formed a cluster (Cluster-1), and PI 398237 and PI 603171 formed another cluster (Cluster-4), which were clearly distinct from other clusters ( Figure 6). Since it would be advantageous to develop new varieties with good germination percentages, not only in the presence of water stress but in the absence of it, PI 398566 and PI 424605A which had high germination percentages under no, low, mild, severe, and extreme water stress would be good materials to include in variety development programs.

Discussion
Significant genetic variability was observed in germination percentage across all water stress levels under laboratory conditions and emergence percentage under field conditions. Overall, based on germination percentage under no, low, mild, severe, and extreme water stresses (0.00, −0.27, −0.54, −0.82, and −1.09 MPa) and emergence percentage under three environments (Florence, SC in 2021 and 2022 and Jackson Springs, NC in 2022), PI 398566, PI 424605A, and PI 398532 were the best genotypes, and PI 398237, PI 603171, PI 408169D, PI 408175, and PI 398704 were the poor genotypes. Among those genotypes, PI 398566 and PI 424605A were the top two, and PI 398237 and PI 603171 were the bottom two according to ranking based on numerical values of the above traits. A principal component analysis (PCA) using germination percentage at different water stress conditions in the laboratory experiment and emergence percentage under field conditions confirmed these results (PI 398566 and PI 424605A as the best and PI 398237 and PI 603171 as the worst) ( Figure 6). On the PCA biplot, PI 398566 and PI 424605A formed a cluster (Cluster-1), and PI 398237 and PI 603171 formed another cluster (Cluster-4), which were clearly distinct from other clusters ( Figure 6). Since it would be advantageous to develop new varieties with good germination percentages, not only in the presence of water stress but in the absence of it, PI 398566 and PI 424605A which had high germination percentages under no, low, mild, severe, and extreme water stress would be good materials to include in variety development programs.  Table S1 for other genotype names corresponding to the numbers.
With increasing intensity of water stress, germination percentage decreased, and the range widened (i.e., the difference between the lowest and highest values of germination percentage increased) (Figure 1). This result indicates that the germination of soybean is significantly affected by water stress. The above result is supported by our previous research that demonstrated decreasing emergence rate with increasing levels of water stress when soybean plants were grown at 100, 80, 60, 40, and 20% pot-water holding capacity [50]. In the present study, while most genotypes suffered from a drastic reduction in germination percentage under water stress conditions, the genotypes PI 398566 and PI 424605A could maintain better germination under water stress. These genotypes were more rigorous in germination which is reflected in their better germination potential and germination rate index ( Figure 3).
Often, seed size is believed to influence soybean emergence, though there are contradicting reports regarding this [51][52][53]. In the present study, we did not find any significant relationship between the seed size and germination percentage under different water stress levels and between seed size and emergence percentage under three field environments ( Figure S2). These results are supported by our previous research in which we did not find any correlation between soybean seed size and emergence percentage when seeds were sown in pots maintained at 100, 80, 60, 40, and 20% pot-water holding capacity [50]. A lack of the relation between seed size and emergence percentage has also been reported for cereal grains such as wheat, oat, and barley [54][55][56][57][58].  Table S1 for other genotype names corresponding to the numbers.
With increasing intensity of water stress, germination percentage decreased, and the range widened (i.e., the difference between the lowest and highest values of germination percentage increased) (Figure 1). This result indicates that the germination of soybean is significantly affected by water stress. The above result is supported by our previous research that demonstrated decreasing emergence rate with increasing levels of water stress when soybean plants were grown at 100, 80, 60, 40, and 20% pot-water holding capacity [50]. In the present study, while most genotypes suffered from a drastic reduction in germination percentage under water stress conditions, the genotypes PI 398566 and PI 424605A could maintain better germination under water stress. These genotypes were more rigorous in germination which is reflected in their better germination potential and germination rate index ( Figure 3).
Often, seed size is believed to influence soybean emergence, though there are contradicting reports regarding this [51][52][53]. In the present study, we did not find any significant relationship between the seed size and germination percentage under different water stress levels and between seed size and emergence percentage under three field environments ( Figure S2). These results are supported by our previous research in which we did not find any correlation between soybean seed size and emergence percentage when seeds were sown in pots maintained at 100, 80, 60, 40, and 20% pot-water holding capacity [50]. A lack of the relation between seed size and emergence percentage has also been reported for cereal grains such as wheat, oat, and barley [54][55][56][57][58].
Emergence percentage was measured at Florence and Jackson Springs, two major soybean growing areas in SC and NC, respectively. Though germination percentage ranged between 87 and 100% under no-water stress conditions in the laboratory experiment ( Figure 1 and Table S3), emergence was considerably lower for most genotypes under field conditions (between 3 and 39% at Florence in 2021; 2 and 65% at Florence in 2022; and 22 and 77% at Jackson Springs in 2022) (Figure 2 and Table S3). This may be because of the dry soil conditions that existed during the planting time. The experimental sites did not receive any rain in the last 7 to 8 d prior to planting. The cumulative precipitation values at Florence in a month prior to planting in 2021 and 2022 were lower than what is typical for this region (based on the 30-year cumulative normal precipitation) ( Figure S1). In 2021, only 3.5 cm rainfall was recorded in a month prior to planting at Florence. Emergence was recorded at 7 DAP in all experimental fields. Florence experimental field did not receive any rainfall till 7 DAP in 2021, and it received only 0.84 cm rainfall in the first week of planting in 2022. Jackson Springs experimental field received 1.12 cm rainfall in the first week of planting in 2022. The dry soil conditions at Florence in 2021 might have led to the relatively worst emergence percentage in that environment, compared to that in the other two environments.
We also evaluated whether emergence percentage was related to root traits, and found that the length of fine roots with a diameter between 0.25 and 0.5 mm was positively related to emergence percentage in 2021 and 2022. Interestingly, the finest root (diameter <0.25 mm) traits did not appear to influence emergence (no correlation between emergence and L1, SA1, or V1 at 14 DAP in 2021 and at 7 DAP in 2022; Table 2). In other dicot species, it has been reported that instead of the finest parts in the root system (such as root tip), the first 2-3 inches behind the root tip represents the most active site of water and nutrient uptake [59]. Our results provide a background to investigate the role of root morphology in supporting emergence.
When root traits measured at Florence in 2022 were further evaluated, the length, surface area, and volume of the total root system and fine roots with diameters between 0.25 and 0.5 mm measured at 7 DAP were correlated with the same traits measured at 14 DAP with correlation coefficient values ≥0.38 (Table S5). Furthermore, in 2021, at the same location, the length, surface area, and volume of the total root system and fine roots (diameters between 0.25 and 0.5 mm) measured at the seedling stage (14 DAP) were correlated with the same traits measured at the late vegetative stage (74 DAP), with correlation coefficients ≥0.37 (Table S5). This implies that genotypes that possess a better root system in the early seedling stage could maintain doing so later in their life cycle. However, it should be noted that after root emergence, further development of the root system and the final root system architecture could be influenced by soil physical conditions and biotic and abiotic factors (e.g., temperature, water, oxygen, soil nutrient availability, and soil biota) [60].
Germination and emergence can be affected by different biotic and abiotic factors. Among the biotic factors, soil-borne pathogens and pests are major concerns, though in some cases birds and other vertebrate pests also impact soybean emergence [61,62]. In the present study, we did not evaluate the effect of biotic factors on germination and emergence. However, we dug up the ungerminated seeds in the experimental fields and did not find any disease symptoms on those seeds. Abiotic factors such as soil temperature, soil compaction, and soil surface crusting may also affect soybean germination and emergence [63,64] and should be evaluated in future studies.

Conclusions
In the present study, we found that soybean germination and root morphology are significantly affected by water stress. Among the 41 soybean PI's evaluated, PI 424605A and PI 398566 could maintain germination percentage and better root morphology (greater length, surface area, and volume of the whole root system and those of fine roots) under water stress. PI 424605A and PI 398566 had better germination and root morphology not only under water-stress but under no-stress conditions as well. Additionally, the same genotypes possessed better emergence. These genotypes would be useful for breeding programs in developing new varieties with drought tolerance at the emergence phase. The best (PI 424605A and PI 398566) and worst (PI 398237 and PI 603171) genotypes identified in this study could be further characterized to identify any seed-related or root-related compounds that may improve germination and emergence from dry as well as optimally moist soils. Furthermore, the above contrasting genotypes could be used to develop mapping populations to elucidate the genetic mechanism behind better germination and emergence and the genetic loci regulating traits related to improved germination and emergence.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/agronomy12081944/s1. Figure S1: Cumulative precipitation and daily average temperature from one month prior to planting to one month after planting at Florence, SC (a,b) and Jackson Spring, NC (c,d) in comparison with the historic weather data. Figure S2: Relationship of seed size (measured through 200-seed weight) with germination percentage under controlled environmental conditions (a-e) and with emergence percentage under field conditions (f-h). Figure S3: Distribution of germination percentage values among 41 soybean genotypes under controlled environmental conditions (a-e) and distribution of emergence percentage among the same genotypes under field conditions (f-h). Table S1: List of soybean genotypes used in the study, country of origin, and 200-seed weight. Table S2: Results of soil analysis conducted before soybean planting. Table S3: Germination traits of soybean genotypes under controlled environmental conditions and emergence percentage under field conditions. Table S4: Ranking of soybean genotypes for root traits under severe and extreme water stress. Table S5