Notes
Root distribution contributing to pushing resistance in soybean
2026 年 76 巻 4 号 p. 381-386
詳細
2026 年 76 巻 4 号 p. 381-386
To efficiently develop soybean (Glycine max) cultivars resistant to root lodging, which occurs in the late growth stages and causes significant yield losses, it is essential to understand the root traits that contribute to this resistance. However, the root zones most responsible for plant support remain unidentified, leaving uncertainty about which should be prioritized in root trait evaluation. This study aimed to identify root distributions associated with pushing resistance, which is closely related to root lodging, through vertical and horizontal root pruning combined with analyses of root distribution during the full pod to full seed stages. The results suggest that root length (≥0.5 mm diameter) within the top 15 cm of the soil and within 20 cm horizontally from the plant, especially directly beneath the plant, plays a crucial role in maintaining pushing resistance. The root zone identified in this study may provide a useful target for the detailed evaluation of root traits in field-grown soybean during the late growth stages, and approaches such as shovelomics and the basket method could facilitate such evaluations in future breeding programs.
Lodging during the reproductive stages of soybean (Glycine max) can cause yield losses of 21% to 34% (Cooper 1971, Saitoh et al. 2012), making lodging resistance a critical breeding target. Two major lodging types are recognized: stem-bending lodging, characterized by bending of the main stem, and root lodging, in which the plant tilts from the base during the late growth stages (Saitoh et al. 2001, Shimada et al. 2002). The latter is common in soybean (Konno and Homma 2024) and is especially problematic because plants rarely recover once lodged.
Pushing resistance provides an effective measure of root lodging potential in soybean (Shimada et al. 2002). It is defined as the maximum stress required to tilt a stem base to 45° using a prostrate tester after cutting the shoot at a fixed height (Kashiwagi and Ishimaru 2004). Previous studies identified root weight, root length, the basal diameter of the taproot, and the number and total cross-sectional area of primary lateral roots as root traits related to pushing resistance in soybean (Chen et al. 2011, 2017, Kitabatake et al. 2019, Terashima et al. 2001, Tsujioka et al. 2009). However, it remains unclear to what extent lateral roots distributed vertically and horizontally contribute to plant anchorage, and whether other root types (the taproot, basal roots, or shoot-borne roots) play any role in structural support. Furthermore, previous studies have primarily relied on correlation analyses between root traits and pushing resistance, lacking direct evidence for the contribution of specific root zones. This study aimed to directly identify spatial root distributions essential for anchorage through vertical and horizontal root pruning during reproductive stages.
Field experiments were conducted in 2024 at the University farm of Utsunomiya University, Japan (36.49°N, 139.98°E). Fertilizer was applied on June 17 (20 kg N, 80 kg P2O5, 80 kg K2O ha–1). Two seeds of soybean cultivars ‘Tamahomare’ (JP number: 29184, Origin: Nagano, Japan) and ‘Karasumame’ (JP number: 27584, Origin: China) were sown per hill on July 4. Seedlings were thinned to one plant at V2. Standard crop management was applied. To minimize lodging unrelated to treatments, garden stakes and vinyl strings were installed.
A randomized complete block design with four blocks was employed. To evaluate varietal differences in pushing resistance, two cultivars were randomly assigned within each block and grown in 11 rows per cultivar (row length: 4 m; row spacing: 70 cm; plant spacing: 20 cm) (Fig. 1a). Plants were supported with stakes and plastic strings to prevent lodging, but ‘Karasumame’ exhibited severe lodging during the late growth stage. Consequently, the experiment involving this cultivar was terminated at the R4 stage, and root pruning treatments were conducted only for ‘Tamahomare’. The unpruned plants were used as controls. For horizontal root pruning, trenches were excavated perpendicular to the planting row at a distance of 30 cm from the plant and to a depth of 60 cm (Fig. 1b, Supplemental Fig. 1a). At 15 cm below the soil surface on the trench wall, iron plates (25 cm × 47 cm × 0.2 cm) were inserted using a hammer to sever the roots (Supplemental Fig. 1b). For vertical root pruning, a square shovel (25.0 cm × 29.5 cm × 0.18 cm) was inserted parallel to the row at varying distances from the plant base to vertically sever roots on both sides of the plant (Fig. 1b, Supplemental Fig. 1c). At R4 (August 29), due to labor constraints, three of the four blocks were used, and four treatments were randomly assigned within each block: control, vertical pruning at 10 cm from the plant base, horizontal pruning at 15 cm below the soil surface, and combined vertical and horizontal pruning. At the R5 stage (October 10), vertical pruning treatments at distances of 5, 10, 15, 20, 25, and 30 cm from the plant base were randomly assigned within each of the four blocks. At the R6 stage (October 17), six treatments were randomly assigned within each block: control, vertical pruning at 10 and 20 cm from the plant base, horizontal pruning at 15 cm below the soil surface, and combined vertical and horizontal pruning treatments. Plants subjected to root pruning treatments were selected as follows. To minimize border effects, root pruning treatments were applied only to plants in the nine central rows of the 11-row plots (Fig. 1a). Due to spatial constraints associated with trench excavation, four plants per row (two at each end and two adjacent interior plants) were designated as candidate plants, resulting in 36 candidates in total. From these candidates, to minimize variation among individual plants across treatments, plants with similar first internode diameters were selected, and root pruning treatments at each growth stage were randomly assigned to the selected plants within each block.
Experimental design (a) and vertical and horizontal root pruning method (b). Although each cultivar was grown in 11 rows per block, only five rows are shown schematically (a). Root pruning treatments were randomly applied to the plants indicated by solid open circles. Solid filled circles indicate the treated and sampled plants. Dashed circles indicate plants that were neither treated nor sampled. Gray rectangles denote trenches. Horizontal root pruning was performed using a steel plate inserted from a trench (b). Vertical root pruning was conducted using a square shovel inserted perpendicular to the soil surface, parallel to the planting row, at varying distances on both sides of the target plant. The red arrows indicate the direction in which the plant was pushed for measuring pushing resistance.
For all the candidate plants, aboveground traits, including the diameters of the first internode and the cotyledonary node, were recorded prior to measuring pushing resistance. For the cultivar ‘Tamahomare’, which received root pruning treatments, aboveground traits were recorded before pruning, and pushing resistance was measured immediately after the pruning treatment. For pushing resistance measurements, the main stem was cut at 15 cm aboveground, and leaves and branches were removed. Pushing resistance was measured using a prostrate tester (DIK-7401, Daiki Rika Kogyo, Japan) by applying force perpendicularly to the stem at 10 cm above the soil surface. The maximum force required to tilt the stem to 45° was recorded. Roots beneath each plant were then excavated as 30 cm cubic soil blocks using the iron plates and square shovel described above. In the field, to collect only the roots of the plants used for pushing resistance measurements, the soil blocks containing roots were placed into a plastic container filled with water, and the soil was carefully washed away (Supplemental Fig. 1d). The washed roots were then suspended to preserve their spatial configuration and transported to the laboratory (Supplemental Fig. 1e). In the laboratory, the suspended root systems were sectioned from the bottom upward into 10 cm cubic segments while maintaining their original orientation (Fig. 2, Supplemental Fig. 1f). Within each 10 cm cube, root length ≥0.5 mm diameter was scanned following the method described by Koyama et al. (2021) in which the roots were spread out on a Plexiglas tray filled with enough tap water to cover the surface of the roots and analyzed using WinRhizo Pro 2017 (Regent Instruments, Canada). Statistical analyses were conducted using JMP 12.2 (SAS Institute, USA). Differences in pushing resistance between cultivars were evaluated using Student’s t-test. The effects of root pruning treatments were evaluated by ANOVA followed by Tukey’s HSD test. Relationships between pushing resistance and root length distribution were examined using Spearman’s rank correlation coefficient.
Root distribution used for correlation analysis with pushing resistance shown in Table 1. Root length of ≥0.5 mm in the gray areas was used. The arrow indicates the pushing direction.
At R4, ‘Tamahomare’ exhibited pushing resistance of 24.2 N, nearly twice that of ‘Karasumame’ (12.4 N), demonstrating significant varietal differences (p < 0.05, n = 3). The spatial distribution of root length (≥0.5 mm diameter) for each cultivar is shown in Supplemental Fig. 2a–2f. The mean root length (±SE, n = 3) of ‘Tamahomare’ at soil depths of 0–10 cm, 10–20 cm, and 20–30 cm (30 cm wide × 30 cm deep × 10 cm high) was 232 ± 45 cm, 104 ± 7 cm, and 64 ± 18 cm, respectively, whereas that of ‘Karasumame’ was 140 ± 14 cm, 143 ± 12 cm, and 37 ± 22 cm, respectively. Roots distributed within the upper 20 cm accounted for 84% and 89% of the total root length in ‘Tamahomare’ and ‘Karasumame’, respectively. Root length on the pushed side, directly beneath the plant row, and the side subjected to pulling during lodging (10 cm wide × 30 cm deep × 30 cm high) was 90 ± 33 cm, 188 ± 14 cm, and 122 ± 15 cm for ‘Tamahomare’, and 89 ± 10 cm, 154 ± 6 cm, and 77 ± 14 cm for ‘Karasumame’, respectively. Roots located directly beneath the plant row accounted for 47% and 48% of the total root length in ‘Tamahomare’ and ‘Karasumame’, respectively. Significant correlations were observed between pushing resistance and root length within the following soil volumes: (i) the region directly beneath the plant (10 cm wide × 10 cm deep × 20 cm high), (ii) the T-shaped region composed of a vertical section directly beneath the plant (10 cm wide × 10 cm deep × 30 cm high) and a horizontal section parallel to the row (10 cm wide × 30 cm deep × 10 cm high), and (iii) the region directly beneath the plant row (10 cm wide × 30 cm deep × 30 cm high) (Table 1, Supplemental Fig. 3a–3c).
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Two cultivars at R4a | 0.658 | 0.662 | 0.829* | 0.593 | 0.674 | 0.738 | 0.717 | 0.423 | 0.911* | 0.077 | 0.826* | 0.548 |
| Root pruning at R4b | 0.615* | 0.031 | 0.276 | 0.241 | 0.542 | 0.569 | 0.547 | 0.312 | 0.438 | 0.495 | 0.493 | 0.625* |
| Root pruning at R6c | 0.592** | 0.224 | 0.492* | 0.522** | 0.426* | 0.500* | 0.506* | 0.536** | 0.424* | 0.437* | 0.477* | 0.509* |
a Two cultivars at R4 used the data of the varietal differences (n = 6). b Root pruning at R4 used the data of Fig. 3a (n = 12). c Root pruning at R6 used the data of Fig. 3c (n = 24). * and ** indicate statistically significant correlations at the 5% and 1% level, respectively (Spearman’s rank correlation coefficient).
Because ‘Karasumame’ exhibited approximately half the pushing resistance of ‘Tamahomare’ and showed severe lodging during the later growth stages, all subsequent pruning experiments were conducted using ‘Tamahomare’. Horizontal root pruning at 15 cm depth had no significant effect on pushing resistance at either R4 or R6, indicating that deeper roots beyond this layer contribute little to anchorage (Fig. 3a, 3c). In contrast, at R5, vertical pruning performed beyond 20 cm had negligible effects, whereas pruning at distances less than 15 cm significantly decreased pushing resistance (Fig. 3b). Vertical pruning at 10 cm from the plant consistently reduced pushing resistance at R4 and 6 (Fig. 3a, 3c).
Effects of horizontal and vertical root pruning on pushing resistance of ‘Tamahomare’. (a) the full pod stage (R4) (n = 3), (b) the beginning seed stage (R5) (n = 4), (c) the full seed stage (R6) (n = 4). The lower x-axis indicates the depth of horizontal root pruning, whereas the upper x-axis indicates the distance from the plant for vertical root pruning. Different letters indicate statistically significant differences at the 5% level (Tukey’s HSD test).
After root pruning at R4, significant correlations were found between pushing resistance and root length in a 30 cm cubic volume and the region on the side subjected to pulling during lodging (10 cm wide × 30 cm deep × 30 cm high) (Table 1, Supplemental Fig. 3d, 3e). The distribution of root length (≥0.5 mm diameter) in the control plants at R4 was described above using Supplemental Fig. 2a–2c. At R6, the mean root length (±SE, n = 4) within the 0–10 cm, 10–20 cm, and 20–30 cm soil layers (30 cm wide × 30 cm deep × 10 cm high) was 139 ± 6 cm, 107 ± 17 cm, and 27 ± 8 cm, respectively, with roots within the upper 20 cm accounting for approximately 90% of the total root length (Supplemental Fig. 2g–2i). Root length on the pushed side, directly beneath the plant row, and the side subjected to pulling during lodging (10 cm wide × 30 cm deep × 30 cm high) was 53 ± 13 cm, 130 ± 5 cm, and 89 ± 16 cm, respectively, with roots beneath the plant row accounting for 48% of the total. At R6, correlations were significant at the 5% significance level except for root length within the region directly beneath the plant (a 10 cm cubic volume) (Table 1). Notably, strong correlations (p < 0.01) were observed for the 30 cm cubic volume, the region directly beneath the plant (10 cm wide × 10 cm deep × 30 cm high), and the T-shaped region composed of a vertical section directly beneath the plant (10 cm wide × 10 cm deep × 30 cm high) and a horizontal section perpendicular to the row (30 cm wide × 10 cm deep × 10 cm high) (Supplemental Fig. 3f–3h).
This study identified root zone critical for pushing resistance. Previous studies have shown that pushing resistance is associated with traits such as taproot basal diameter and the number of primary lateral roots (Kitabatake et al. 2019, Tsujioka et al. 2009), and the total cross-sectional area of lateral roots, particularly in the upper soil layers (Terashima et al. 2001). Consistent with these findings, our results demonstrated that root length (≥0.5 mm diameter) within the top 15 cm of soil and within 20 cm horizontally from the plant, particularly directly beneath the plant, plays a key role in maintaining pushing resistance (Table 1, Fig. 4). By clarifying the spatial distribution of roots associated with pushing resistance, our results provide information that cannot be obtained from pushing resistance measurements alone and may improve the efficiency of root trait evaluation in breeding programs aimed at improving resistance to root lodging.
Root distribution considered to contribute to pushing resistance (red-shaded areas), based on root pruning treatments and root length distribution analysis. Gray-shaded areas indicate root zones where pruning had no significant effect on pushing resistance. The area enclosed by the red dashed line represents the root length distribution that showed a correlation with pushing resistance. The arrow indicates the direction in which the plant was pushed for measuring pushing resistance.
Although the relative root distribution patterns were broadly similar between ‘Tamahomare’ and ‘Karasumame’ (Supplemental Fig. 2), pushing resistance differed markedly between the two cultivars. This suggests that varietal differences in pushing resistance cannot be explained solely by proportional root distribution. Rather, the greater absolute root length within the key root zones, particularly beneath the plant base and within the upper soil layers, is likely a more important determinant of anchorage strength. From a breeding perspective, these results indicate that increasing root amount within the critical root zone identified in this study may be more effective than modifying root distribution patterns alone for improving lodging resistance.
In the present study, root pruning treatments could be applied only to ‘Tamahomare’, because ‘Karasamame’ lodged during the later growth stages. ‘Tamahomare’ is a cultivar known for its strong lodging resistance and has been widely adopted as a recommended variety in many regions of Japan (Mikoshiba and Maruyama 1980), and it is still cultivated in some areas (MAFF 2025). Moreover, since the taproot typically elongates beneath the plant and primary lateral roots predominantly develop in the upper soil layers, the root distribution patterns identified in this study are consistent with general soybean root system architecture (Supplemental Fig. 2) (Tanaka 1977). Therefore, although our experimental results were mainly obtained from a single cultivar, evaluating the root distribution highlighted in this study represents a rational and biologically meaningful approach for assessing root traits associated with pushing resistance.
In this study, the distribution of root length (≥5 mm diameter) associated with pushing resistance varied with the growth stages (Table 1, Supplemental Fig. 3d–3h). At R4, roots oriented perpendicular to the pushing direction were associated with pushing resistance (Supplemental Fig. 3b, 3c). This relationship was likely attributable to the greater root length observed in the Front and Back positions of the Center at a depth of 0–10 cm during the R4 stage (Supplemental Fig. 2a). In contrast, at R6, roots oriented parallel to the pushing direction were associated with pushing resistance (Supplemental Fig. 3h). Previous studies have reported that lateral root length decreases after R5 (Tanaka et al. 1990), and a similar trend was observed in this study, where root length in the Front and Back positions of the Center decreased from R4 to R6 (Supplemental Fig. 2a, 2g). These results suggest that changes in root length distribution during plant development may alter the spatial distribution of roots contributing to pushing resistance.
Root lodging, which is the focus of this study, has been reported to occur during later growth stages when the physical strength of the stem becomes sufficiently large (Saitoh et al. 2001). Several methods have been used to evaluate root system traits in the field at later growth stages, including the monolith method, core sampling, and trench profile methods. However, these approaches require considerable labor for excavation, root washing, or trench construction, and the measurable traits are generally limited to traits such as root length or root diameter at different soil depths (Teramoto and Uga 2022). Therefore, practical approaches for evaluating root traits within the root zones associated with pushing resistance are needed. For a more detailed evaluation of the root traits within such confined zones, the use of shovelomics (Burridge et al. 2016), which allows assessment of multiple traits, including the diameter, number, and angle of different root types (tap, lateral, basal, and shoot-borne roots), and the basket method (Oyanagi et al. 1993, Sakata et al. 2003), which enables simple evaluation of root growth angle and number, are considered effective. Applying these approaches may facilitate breeding for root system architectures that strengthen pushing resistance and reduce root lodging.
TK: conceptualization, funding acquisition, investigation, methodology, project administration, supervision, validation, visualization, writing—review and editing. KA: data curation, formal analysis, investigation, writing—original draft, writing—review and editing. All authors have read and agreed to the published version of the manuscript.
We thank Mr. N. Yamaguchi for his assistance in conducting the experiments and Dr. T. Kashiwagi for his advice. We also thank Mr. T. Sekizawa, Mr. T. Akiyama, and the students of the School of Agriculture, Utsunomiya University, who assisted with the experiments. The soybean cultivars ‘Tamahomare’ and ‘Karasumame’ were obtained from the Genebank Project, NARO. This work was partly supported by JSPS KAKENHI Grants (No. 22K05595) from the Japan Society for the Promotion of Science and the C-Bio Collaborative Research Grant for Plant Molecular Agriculture at Utsunomiya University, funded by the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan.
