ldentification of tolerance to high density and lodging in short petiolate germplasm M657 and the effect of density on yield-related phenotypes of soybean

GAO Hua-wei ,YANG Meng-yuan ,YAN Long ,HU Xian-zhong ,HONG Hui-long ,ZHANG Xiang,SUN Ru-jian,WANG Hao-rang,WANG Xiao-bo,LlU Li-ke,ZHANG Shu-zhenQlU Li-juan

1 Key Laboratory of Soybean Biology,Ministry of Education/College of Agriculture,Northeast Agricultural University,Harbin 150030,P.R.China

2 National Key Facility for Crop Gene Resources and Genetic Improvement,Institute of Crop Sciences,Chinese Academy of Agricultural Sciences,Beijing 100081,P.R.China

3 College of Life Science of Liaocheng University,Liaocheng 252059,P.R.China

4 Institute of Cereal and Oil Crops,Hebei Academy of Agricultural and Forestry Sciences/Hebei Laboratory of Crop Genetic and Breeding,Shijiazhuang 050035,P.R.China

5 Hulun Buir Institution of Agriculture and Animal Husbandry,Zhalantun 162650,P.R.China

6 Jiangsu Xuhuai Regional Institute of Agricultural Sciences,Xuzhou 221000,P.R.China

7 School of Agronomy,Anhui Agricultural University,Hefei 230036,P.R.China

Abstract Soybean yield has traditionally been increased through high planting density,but investigating plant height and petiole traits to select for compact architecture,lodging resistance,and high yield varieties is an underexplored option for further improving yield.We compared the relationships between yield-related traits,lodging resistance,and petioleassociated phenotypes in the short petiole germplasm M657 with three control accessions during 2017-2018 in four locations in the Huang-Huai region,China.The results showed that M657 exhibited stable and high tolerance to high planting density and resistance to lodging,especially at the highest density (8×105 plants ha-1).The regression analysis indicated that a shorter petiole length was significantly associated with increased lodging resistance.The yield analysis showed that M657 achieved higher yields under higher densities,especially in the northern part of the Huang-Huai region.Among the varieties,there were markedly different responses to intra-and inter-row spacing designs with respect to both lodging and yield that were related to location and density.Lodging was positively correlated with planting density,plant height,petiole length,and number of effective branches,but negatively correlated with stem diameter,seed number per plant,and seed weight per plant.The yield of soybean was increased by appropriately increasing the planting density on the basis of the current soybean varieties in the Huang-Huai region.This study provides a valuable new germplasm resource for the introgression of compact architecture traits that are amenable to providing a high yield in high density planting systems,and it establishes a high-yield model of soybean in the Huang-Huai region.

Keywords: soybean,short petiole,high density and lodging,yield-related phenotypes

1.lntroduction

Soybean is a major source of plant protein and fat,and it contributed 59.69% of the global oil crop production in 2017-2019 (Wilson 2010).Significant differences exist in the planted areas and yields among the major soybean producing countries.The United States,Brazil,and Argentina together accounted for 81-82.2% of the total yield globally,contributing 2.99,3.11,and 3.00 t ha-1,respectively;while China consumed a disproportionately high amount of soybean,with imports reaching 86.94% of demand in the most recent five years,and a national yield level of only 1.81 t ha-1despite recent increases (Lvet al.2018;Tanget al.2018).Thus,there is an urgent need to increase the yield of soybean,especially among the varieties favorable for cultivation in China.

In the 1990’s,Richard L.Cooper developed various solid-seeded semi-dwarf soybean cultivars,including Elf,Sprite,Hobbit,and his High-Yield-System-in-Place(HYSIP) Program demonstrated that the yields of semidwarf varieties with 17 cm intra-row spacing could be further increased by more than 20% compared with traditional cultivation using wide-row spacing (＞70 cm),by increasing the planting density from ＜2×105to 6-8×105plants ha-1(Cooper 1990,2003,2011;Cooperet al.2004).This innovation enabled the achievement of record yields through high planting densities,such as in the United States (7.62 t ha-1) (Holshouseret al.2002;Liuet al.2007),Brazil (5.05 t ha-1) (Knebelet al.2006;Spaderet al.2015),South Korea (4.19 t ha-1) (Choet al.2010),and Iran (4.33 t ha-1) (Daroishet al.2005).In contrast,the planting density of soybean in China has remained relatively low,typically only 1.8-3.6×105plants ha-1.However,there are only nine recorded trials of seven accessions in which the planting density exceeded 3×105plants ha-1(Duet al.2014;Luet al.2017;Guoet al.2019),suggesting that there is a great opportunity to considerably increase Chinese soybean yield through breeding for improved tolerance to planting density in the commonly grown Chinese cultivars.

To obtain any improvement in soybean yield with the increased planting density,the lodging resistance of plants is a prerequisite.The lodging properties of soybeans are determined by genotype (Chenet al.2017) and other traits,including plant height (Xieet al.2011),growth habit(Coberet al.2010),stem termination (Liuet al.2010),root morphology (Zhouet al.2007) and stem components (Liuet al.2016),as well as its relationship with environmental factors such as light and temperature (Board 2001;Zhanget al.2020;Caoet al.2021),shade level (Franklinet al.2005) and the intensity of wind and rain (Martinezet al.2016).

High yields of soybean depend on a balance between compact architecture and photosynthetic efficiency,both of which are partially determined by petiole length (Junet al.2012;Zhanget al.2015;Liuet al.2020).However,few studies have identified short petiole accessions that also exhibit adaptability to high planting densities and resistance to lodging with high yield-related phenotypes.In order to identify a candidate germplasm with lodging resistance and high yield-related phenotypes under dense cultivation conditions,we compared the short petiole germplasm M657 with Jihuang 13 (JH13) and two widely cultivated varieties (Jidou 12 (JD12) and Handou 5(HD5)) at four locations across the Huang-Huai region,China.We found that M657 exhibited strong resistance to lodging and high planting density,resulting in substantial yield increases.This study provides a valuable new germplasm resource for the introgression of compact architecture traits that are amenable to high yield in high density planting systems and technical support for the field production of soybean in the Huang-Huai region.

2.Materials and methods

2.1.Plant materials

A total of four varieties were used in this study,including the short petiole germplasm M657 (Gaoet al.2021),which was identified from an ethylmethylsulfone (EMS)-induced mutant library of JH13 (WT).JH13 and two control cultivars from the Chinese National Soybean Variety Approval Certification,JD12 (ZDD23040) and HD5 (ZDD23959),were obtained from the National Crop Germplasm Resources Platform (http://www.cgris.net/#).

2.2.Experimental design

A two year (2017-2018),multi-factor field experiment was set up in a randomized block design (RBD),in triplicate.The experimental factors were four germplasms (Acc.),four locations (Loc.),two years (Yea.),and 10 different densities (Den.).The germplasms were M657 and three soybean cultivars (JH13,JD12,and HD5) that are commonly planted varieties in the Huang-Huai region of China.The four experimental sites were located in the northern and middle parts of the Huang-Huai region,including Shunyi of Beijing in 2018 (Shunyi Experimental Station,Institute of Crop Sciences,Chinese Academy of Agricultural Sciences,2018BJ);Shijiazhuang of Hebei Province in 2018 (Mazhuang Experimental Station,Cereal Oil Crop Institute,Hebei Academy of Agricultural and Forestry Sciences,2018HB);Liaocheng of Shandong Province in 2017 and 2018 (Daokoupu Experimental Field,College of Life Science of Liaocheng University,2017SD and 2018SD);and Shangqiu of Henan Province in 2018 (Yucheng Experimental Field,Institute of Crop Sciences,Chinese Academy of Agricultural Sciences,2018HN) (Table 1).

The plot areas for the 10 densities were 5 m long and 5 rows wide,with 6.25-12.5 m2spacing (2×105-8×105plants ha-1) (Tables 1 and 2).The control values for interrow spacing and planting density (NPM,plants ha-1) were 50 cm,2.25×105plants ha-1and 40 cm,1.875×105plants ha-1,which are the respective planting densities for the northern and middle parts of the Huang-Huai region in the Chinese National Soybean Variety Regional Test Report (Seed Administration of Ministry of Agriculture and Rural Affairs &National Agricultural Technology Extension Service Center 2020).

Table 1 Field overview of the density testing at four sites

Table 2 Design of the planting density study

The crop preceding soybean was maize,and the fields were planted in both years as maize-soybean rotations.Three seeds were planted per hole and seedlings were removed or transplanted to ensure only one plant per hole at the first trifoliolate stage (V1) stage.The varieties were sown under rainfed conditions and received irrigation when necessary to avoid water stress.Weeds,insects,and diseases were chemically controlled as needed.

2.3.Measurement of plant and grain traits

The field investigation was carried out according toDescriptors and Data Standards for Soybean (Glycine spp.)(Qiuet al.2006).The phenotypic traits included petiole length (PL,middle node of 10 plants for each cultivar,measured in cm,from stem to leaf base at full bloom),lodging at flowering date (R2),podding date (R4),seed-filling date (R6),maturity date (LD-R8) (Fehret al.1977),and average lodging over the full growth period(LD-RA).

All plants in each experimental plot were assessed for lodging,and the proportion of lodging (defined as the angle of the main stem to the ground being less than 30°) was calculated across the whole plot.The lodging index value for each soybean variety was determined based on the plant lodging ratio and the following criteria:LD-1,no lodging (no lodged plants);LD-3,light lodging(0＜number of lodged plants≤25%);LD-5,much lodging(25%＜number of lodged plants≤50%);LD-7,severe lodging (50%＜number of lodged plants≤75%);and LD-9,prostrate (proportion of lodged plants＞75%).

At physiological maturity,all subsamples of a quadrant of 3 m×3 rows at ground level were harvested to determine the grain yield (YPM,kg ha-1) and yield components.The yield component traits,including plant height (PH,cm),the lowest pod height (BPH,cm),number of nodes on main stem (NN),effective branch number (EBN),pod number per plant (PN),seed number per plant (SNP),seed weightper plant (SWP,g),100-seed weight (SW,g) and stem diameter (middle position of hypocotyl,SD,cm),were tested at maturity for ten plants from each plot.

2.4.Statistical analysis

Microsoft Excel 2016 was used for statistical calculations of the phenotypic data,analysis of the phenotypic coefficients of variation,and analysis of the relationships between plant phenotypes and density.Hanabi (http://hanabi.data-viz.cn/index) was used to draw Sankey diagrams of lodging in the density experiments.The Aov function of R Software was used for multivariate analysis of variance (ANOVA) among the 10 yield-related phenotypes and Ger.,Den.,and environment (Env.) considering the combinations of location and year.Differences between varieties were analyzed using ANOVA with Ger.,Den.,and their interaction as fixed effects.Multiple comparisons of Ger.,Den.,and Env.were analyzed using LSD (least significant difference) in R Software (agricolae package,LSD test function,alpha=0.05,p.adj=”none”).The Z-score method was used to perform data normalization,and the SPSS version 19.0 for Windows statistical package was used for regression analysis,from “Analyze” process to“Regression” process to Stepwise of Linear.Correlations between NPM,LD-RA,and LD-RA with eight yield-related phenotypes and YPM were analyzed using the cor function in R Software,and the array of the correlation analysis was generated using the ggcorrplot function in R.

3.Results

3.1.Effects of accession,environment,and planting density on lodging

In order to assess the performance of short petiole M657 under high-density conditions in different locations,field trials were conducted across four sites in the Huang-Huai region in 2017-2018.The results of multiple comparisons and ANOVA analysis (Table 3) revealed that except for the BPH of Acc.×Env.,all three single factors,three twofactor interactions and one of the three-factor interactions were significant at aP-value of 0.05 for all nine traits.Compared with JH13,JD12,and HD5,M657 had significantly shortened PL,optimized internode length,increased PN and SNP,and resistance to lodging.PH and NN significantly declined with a decrease in latitude,and the trend of PH rose at first and then fell.The SD significantly declined with an increase in density.

3.2.Assessment of tolerance to high density and lodging for M657

Based on the results showing that M657 exhibited stability in field trials across locations and years,we also assessed its tolerance to density and lodging at four field sites in the Huang-Huai region.M657 exhibited higher resistance to both high density and lodging than the other three cultivars (Table 3;Fig.1-A;Appendices A and B).M657 plants remained standing at all sites and at planting densities from flowering date (R2) to podding date(R4),especially at the D3 density (800 400 plants ha-1)of 2017SD.In contrast,JH13,JD12,and HD5 stayed upright only at relatively low densities,and were generally completely prostrated at the R2 stage under the high densities of C1 (333 500 plants ha-1),C2 (500 250 plants ha-1),and C3 (667 000 plants ha-1),resulting in abnormal plant morphology.M657 only showed much lodging (index value LD-5) at densities of B3 (500 250 plants ha-1) and C3 in 2018BJ,while it exhibited light lodging (LD-3) at the B3 density in 2018HB and 2018HN at R8.Lodging in JH13,JD12,and HD5 varied among locations and densities.These varieties remained upright only at low densities(A1 (200 100 plants ha-1),A2 (300 150 plants ha-1),or B1 (250 125 plants ha-1)) and were severely lodging(LD-7) or prostrate (LD-9) when the density increased.

For the eight yield-related traits of PH,BPH,NN,EBN,PN,SNP,SWP,and SD,in the planting density trials in 2017SD,2018BJ,2018HB,2018SD,and 2018HN(Fig.1-B),the cumulative coefficients of variation for M657 were lower those of JH13,JD12,and HD5,indicating that M657 was more resistant to lodging under high density cultivation.

Fig.1 Lodging (A) and phenotypic coefficients of variation (B) for the four accessions at four locations.A1-D3,200 100,300 150,400 200,250 125,375 188,500 250,333 500,500 250,667 000,and 800 400 plants ha-1,respectively.HD5,Handou 5;JD12,Jidou 12;JH13,Jihuang 13.BJ,Beijing;HB,Hebei;SD,Shandong;HN,Henan.LD-1-9,the lodging index value.LD-1,no lodging (no lodged plants);LD-3,light lodging (0＜number of lodged plants≤25%);LD-5,much lodging (25%＜number of lodged plants≤50%);LD-7,severe lodging (50%＜number of lodged plants≤75%);and LD-9,prostrate (proportion of lodged plants＞75%).BPH,the lowest pod height (cm);PH,plant height (cm);EBN,effective branch number;NN,number of nodeson main stem;PP,pod number per plant;SNP,seed number per plant;SWP,seed weight per plant (g);SD,stem diameter (middle position of hypocotyl,cm).

3.3.Relationships among planting density,lodging,and yield-related traits

We next examined the effects of enhanced lodging resistance on the yield components collected at four sites in 2017-2018 (Fig.2).Lodging was significantly positively correlated with PL,NPM,PH,and BPH,indicating that plant lodging increased with increasing PL,NPM,PH,and BPH,and with decreasing SD.In contrast,lodging was significantly negatively correlated with EBN,PN,SNP,SWP,and YPM,indicating that increasing NPM,LDRA,and lodging could significantly reduce all of the above traits.In addition,we found a negative,but nonsignificant,correlation between NN and lodging (Fig.2).We also observed that yield-related traits and yield per unit area of the germplasms responded differently to the different density combinations of row spacing and plant spacing (Appendix C).In 2018BJ,compared with the conventional density,the PH of M657 showed an increasing trend with increasing density,while the control varieties JH13 and JD12 showed downward trends.PH showed inconsistent trends (Appendix C),while the other seven phenotypes of BPH,NN,EBN,SD,PN,SNP,and SWP (Appendix C) tended to decrease with decreasing space between the rows or between the plants,but they increased at several densities for some varieties when lodging occurred before R3.

Step-wise regression analysis was then performed to examine the relationships between lodging and 11 variables in M657 and JH13 across four locations in 2018.The lodging at maturity(R8) was set as the Y value (LD-R8),and variables were assigned for NPM,PL,PH,BPH,SD,NN,EBN,PN,SNP,SWP and YPM.This analysis yielded the regression equation:

Y=-2.853+8.692E-5NPM+0.171PL+0.125BPH-0.013YPM

The unstandardized coefficients of lodging resistance for each of the five traits were in the following order from the highest to lowest: PL＞BPH＞NPM,with three of these traits (PL,BPH,and NPM) contributing positive significance to lodging resistance (P＜0.01) when the other variables were removed.Specifically,compared to JH13,these effects resulted from shorter PL,lower BPH (lower plant center of gravity),and lower NPM.The standardized regression coefficients for each of the five traits indicated that lodging resistance was the most strongly correlated with PL,followed by BPH and then NPM.

3.4.Relationships between petiole length and yield-related phenotypes and its response to density

In light of the correlation and regression analyses showing strong relationships between petiole length and lodging resistance,we analyzed the relationships between petiole length and the other phenotypes as well as the response of petiole length to planting density (Fig.2).In this analysis,PL was significantly negatively correlated with EBN,PN and SNP,but it was not significantly correlated with SWP,NPM,or YPM.The change in planting density had no significant effect on petiole length,and the shorter petiole length could obviously improve the lodging resistance of the plants.These results indicated a significantly positive correlation between PL with both lodging (LD-RA and LD-R8) and PH.

Fig.2 Correlations between LD-RA (the average lodging over the full growth period) and LD-R8 (the lodging at maturity) and yield-related phenotypes.PL,petiole length (cm);NPM,planting density (plants ha-1);YPM,the grain yield (kg ha-1);BPH,the lowest pod height (cm);PH,plant height (cm);NN,number of nodes on main stem;SNP,seed number per plant;PN,pod number per plant;SWP,seed weight per plant (g);SD,stem diameter (middle position of hypocotyl,cm);EBN,effective branch number.＊ and ＊＊ indicate significant correlations at the 0.05 and 0.01 levels,respectively.

As the density increased,the petiole length of the main stem type JH13 showed a trend of first increasing and then decreasing,and the petiole length of the multibranched variety JD12 decreased only under the high density C3.For the density-tolerant germplasm M657,the petiole showed an increasing trend,but the difference was not significant (Appendix C).These results seem to suggest that inter-row spacing has a greater influence on petiole length.

3.5.Comparison of yields under different planting densities

Four accessions showed diverse yield performance in response to planting density at four locations (Fig.3).Compared with the conventional planting densities of the northern and middle parts of the Huang-Huai region,the yields of M657 showed a different but increasing trend;while JH13,JD12,and HD5 all showed increased densities with their highest yields at the four locations.The densities with the highest yield and lodging for M657,JH13,and JD12 were respectively B2 (3 016.09 kg ha-1,3),A2(2 462.96 kg ha-1,7),and B2 (2 540.44 kg ha-1,5) at 2018BJ;and B2 (4 048.60 kg ha-1,3),B2 (3 904.26 kg ha-1,7),and B1 (3 248.93 kg ha-1,1) at 2018HB.The densities with the highest yield and lodging for M657,JH13,and HD5 were respectively D3 (4 446.67 kg ha-1,1),A1 (2 910.56 kg ha-1,1),and A3 (4 203.33 kg ha-1,7) at 2017SD;C2 (3 394.53 kg ha-1,1),C2 (3 229.22 kg ha-1,7),and A2(5 169.25 kg ha-1,1) at 2018SD;and B2 (3 433.85 kg ha-1,3),B3 (2 090.06 kg ha-1,9),and B2 (3 824.12 kg ha-1,9) at 2018HN.

Fig.3 Yield analysis of the density test.Effects of intra-row spacing on the yields of 2018Beijing (BJ) (A),2018Shandong (SD)(C),2018Hebei (HB) (E),and 2018Henan (HN) (F);effects of inter-row spacing on the yields of 2018BJ (B),2018SD (D),and 2017SD (G).The quantitative level of the ordinate is yield per unit area (YPM,kg ha-1).JH,Jihuang 13;JD12,Jidou 12;HD5,Handou 5.A1-D3,200 100,300 150,400 200,250 125,375 188,500 250,333 500,500 250,667 000,and 800 400 plants ha-1,respectively.Bars mean SD (n=3).Lowercase letters indicate significant differences between locations for the same material;and the significance level is 0.05.

In 2018BJ,with an increase in density,the yields of the four germplasms increased at first and then decreased at inter-row spacings of 50 and 40 cm.Under the interrow spacing of 30 cm,only M657 still showed a trend of first rising and then falling,while the yield of the control variety showed a consistently downward trend.When the inter-row spacing was 10 cm,the yield of M657 showed an increasing trend as the row spacing decreased,while the control varieties showed either differences with no significance or downward trends.Under intra-row spacings of 7.5 and 5 cm,the yield of M657 first increased and then decreased,and the yields of the three control varieties showed either trends of decline or no significant difference.In 2018SD,2018HB,2018HN,and 2017SD,as the density increased,the yield of M657 showed an overall increasing trend,while the control varieties showed either no significant difference or a downward trend.

Compared with the conventional planting density,the high branch type variety,JD12,only showed an increase in yield at A2 in 2018BJ,but there was no significant difference with the yield of A1.The two main stem type varieties,JH13 and HD5,showed significantly higher yields at five densities in four locations and six densities in three locations,respectively.Among them,the highest yield was 5 169.25 kg ha-1for HD5 at A2 in 2018SD(Fig.3).

4.Discussion

4.1.Short petiole germplasm M657 has greater lodging resistance at high densities

Increases in planting density typically result in greater amounts of lodging,leading to thinner plant stems,which ultimately reduces the harvest index (Board 2001;Souzaet al.2010).Optimizing population density is a common practice for maximizing yield per unit area of cultivated land (Xieet al.1993;Placeet al.2009;Wegereret al.2016;Liebertet al.2017).This study showed that M657 exhibited greater tolerance to both high density cultivation and lodging (Table 3;Fig.1;Appendices A and B).Among 10 planting densities at four sites,M657 showed only slight lodging under the high densities of B3 and C3 in 2018BJ,largely due to the greater pressure of heavy wind and rain caused by three continuous typhoons,“Ampil” (Beijing,Hebei,Shandong),“Yagi” (Hebei,Henan,Shandong),and “Rumbia” (Hebei,Henan,Shandong),which were reported by the China Climate Bulletin of 2018(National Climate Center of the China Meteorological Administration,http://www.cma.gov.cn/).Varieties JH13,JD12,and HD5 only remained standing at low densities(A1,A2,or B1),and more severe lodging occurred when the density increased,indicating that the current planting density in the Huang-Huai area is relatively low,and their resistance to high density and lodging were not good,which may result in varying degrees of yield loss under higher wind and abnormal rain conditions.

The plant morphology of M657 (Appendix A) was unaffected by the increased planting from R5 to R6,and we observed abnormal morphology in the control accessions caused by severely prostrate lodging at high densities,which prevented normal harvesting at R8(Appendix A).The less pronounced changes in plant morphology of M657 could result in excellent lodging resistance under higher densities,heavy rain,and abnormal wind.

4.2.Shorter petiole length has positive effects on lodging resistance and canopy structure

Population density can affect the source-sink flow in cropping systems,and increasing the density can not only reduce the potential loss of yield caused by late sowing or drought,but it can also increase the yield by reducing the costs of weed control (Holshouseret al.2002;Placeet al.2009;Wegereret al.2016;Liebertet al.2017).However,increases in population density can also lead to fierce competition among plants for environmental resources,which can result in a lower plateau for productivity (Cooperet al.1971;Holshouseret al.2002;Andradeet al.2010).Short petioles will be conducive to dense planting and are associated with canopy structure and photosynthetic efficiency,and they also affect the final yield (Liuet al.2020).

The shorter petiole length of M657 is likely significant for the improvement of the soybean plant type (Table 3;Appendix A).Our regression analysis showed that PL had a significant effect on lodging,indicating that the short petiole of M657 was a key factor for its tolerance to high density and lodging.There was a significant positive correlation between petiole length and lodging,but no correlations with SWP or YPM (Fig.2).The short petiole of M657 leads to lighter shading between rows and plants,and a wider growth space generally results in excellent overall plant growth under better light and ventilation conditions(Appendix A).We therefore propose that this short petiole could be used to select for improved canopy structures which are associated with photosynthetic efficiency and air permeability in breeding programs that seek to develop high yielding plant varieties (Junet al.2002).

4.3.M657 can serve as a germplasm for high-yield breeding as it relates to high density and lodging resistance

Increasing the population density is an effective strategy for achieving high yield in soybean (Ethredgeet al.1989;Evensonet al.2003;Takedaet al.2008),and the high yield soybean grown in the United States is a typical example of this strategy,coupled with improvements in cultivation techniques (Egliet al.2008).Cooper (2004) and Boquetet al.(1982) demonstrated that the yields obtained with high planting density and narrow row spacing (17-18 cm)increased by 23.4 and 100% in northern and southern crops,compared with that of wide row spacing (50-80 cm).Liuet al.(2007) determined that a planting density of 8×105plants ha-1increased the grain yields for eight varieties by up to 92%,and among them,S26-H2 produced 7.62 t ha-1.Statistical data provided by the USDA-NASS in 2018 (USDA-NASS 2019) showed that among the 11 major soybean producing states,the planting frequency with a row spacing of less than 38.1 cm accounted for about 72.57% of the trials from 2014 to 2018,and among these,197 trials used row spacings below 19 cm.The widespread usage of such narrow spacings indicates that high-density planting of soybean in the United States is an important means of maximizing yield.

Compared with the mean yield of 51.6 bushels(approximately 3 470.25 kg ha-1) noted by the USDANASS,the soybean yield in China at comparable latitudes was only 1 858.18 kg ha-1in 2015-2019 (National Bureau of Statistics of China,http://www.stats.gov.cn/),indicating a large gap in soybean yield per unit area.At present,soybean researchers in China have bred some high-yielding varieties with improved yield potential,such as Zhonghuang 13,Zhonghuang 35,Jidou 17,Henong 71,Henong 91,and others,resulting in recent record yields (Appendix D).Among these records,three trials reported yields exceeding 6 t ha-1,and these three were among only nine that tested population densities exceeding 3×105plants ha-1,indicating that optimizing population density is a direct way to quickly and dramatically increase yield.

M657 had an excellent yield phenotype under high densities in 2018BJ and 2018HB,so it may be suitable for obtaining high yields with high planting density in the northern part of the Huang-Huai region (Fig.3).In contrast,M657 had relatively poor yields in the lower latitudes,and it only gave high yields at extremely high densities (2017SD) (Fig.3).Therefore,the ideal use of M657 would be to genetically adjust the petiole length of conventional soybean varieties to increase the resistance to high density and lodging in order to increase the yield of soybean in the Huang-Huai region.

The other three varieties in this study had several higher YPMs with increasing plant density,indicating that the full yield potential of soybean varieties could be maximized by optimizing the population density for the high-yield varieties (Fig.3).Compared with inter-row spacing,intra-row spacing may have a greater impact on lodging,and an intra-row spacing of 7.5 cm may be more beneficial for achieving the full yield potential(Table 1;Fig.3).These results indicated the possibility of increasing yield by optimizing the population density for high-yield varieties with more precise and sophisticated field management techniques,and this approach may be more suitable for the main-stem varieties.

4.4.Planting density and lodging strongly affect plant morphology,yield-related traits,and yield per unit area

Studies have shown that increasing the population density in soybean can not only reduce the potential losses of yield caused by late sowing or drought,but it can also increase yield by reducing the costs of weed control(Holshouseret al.2002;Placeet al.2009;Wegereret al.2016;Liebertet al.2017).However,increases in population density can also lead to fierce competition among plants for environmental resources,which results in a lower plateau of productivity due to the imbalance between dry matter accumulation and significantly reduced seed production (Cooper 1971;Andradeet al.2010;Zhaiet al.2018;Liu,et al.2021).Previous studies have shown that the yield-related traits of soybean are directly determined by four factors and their interactions,including growth environment (Bertramet al.2004;Corassaet al.2018),sowing date (Andradeet al.2010;Edwardset al.2010;Spader 2015;Masinoet al.2018),genotypic differences in germplasm (Pedersenet al.2003;Andradeet al.2010),and population density(determined by row spacing and plant spacing) (Bertramet al.2004;Janoviceket al.2006;Debruinet al.2008;Andradeet al.2010;Mauadet al.2010;Coxet al.2011).With increasing density,PH and BPH also increased,whereas the responses of PH to density varied among the genotypes.Six other traits,including SD,NN,EBN,PN,SNP,and SWP,have all been reported to decrease with increasing density (Drieveret al.2005;Placeet al.2009;Shamsiet al.2009;Mauadet al.2010;Souzaet al.2010;Zhouet al.2010).

The four varieties investigated in this study have diverse genotypes,and the four trial locations span a broad latitudinal range within the Huang-Huai region.These differences in location and latitude allowed us to observe the phenotypic impacts of different environments on each genotype,in addition to the effects of planting density.We found that with increasing plant density,PH(positive non-significant correlation),petiole length,and BPH (significantly positive correlation) also increased,while the other six traits noted above all showed significantly negative correlations (Fig.2).In this study,we found that the phenotypes of each soybean variety were affected by density and the stage in which lodging occurred.With increasing plant density,the PL of densitytolerant germplasm M657,which showed excellent resistance to both high planting density and lodging,showed an increasing trend and this may be the response of varieties with excellent lodging resistance to high density.In contrast,the PL of the control varieties,JH13 and JD12,showed downward trends with the weakening of the plants.

The response of PH to density is more complex than the other traits (Table 3;Appendix C).Light competition can increase PH within a certain density range,but the stem can become thinner under these conditions,resulting in a decreased NN (Appendix C),and so at high density,the PH will decrease rather than increase.Lodging will eventually occur due to the thinner and weaker stems caused by main stem elongation under insufficient light and ventilation conditions.When the plants become severely prostrate,the shoot apical meristem may lose dominance and the development of branch meristems increases,resulting in slower growth of the main stem and reduced plant height.At an increased density,M657 may be affected by the thinning of the stem,leading to weaker plants and resulting in a decrease in PH.In this study,M657 did not show an extreme situation similar to the significant increase in the PH of the other germplasms.This difference may represent the response of a density-tolerant germplasm to high planting density,which may be a manifestation of the lodging resistance of M657.

As planting density increases,BPH (Appendix C)can increase in varieties with less branching,while EBN(Appendix C) may decrease in varieties with high branch numbers,ultimately resulting in a lower BPH.When severe lodging occurs,the branches will continue to develop at a lower position,also resulting in decreased BPH.We also found that competition between plants under high planting density leads to the suppression of plant growth and thinner stems,thus resulting in a decreased NN.In general,the other five traits,EBN,PN,SNP,SWP,and SD,all showed downward trends with increasing density (Appendix C).However,these traits increased slightly if lodging occurred in the early developmental stages,due to the contributions to productivity from the lateral branches.

In suitable locations,different germplasms should be able to achieve their excellent yield potentials under the reasonable matching density of inter-and intra-row spacing.Compared with the three control germplasms,the YPM of M657 increased significantly when the density increased in 2018BJ,so it may be suitable for achieving high yields with high planting density in the northern part of the Huang-Huai region.Compared with inter-row spacing,intra-row spacing may have a greater impact on lodging,and intra-row spacing of 7.5 cm may be more conducive to obtaining the full yield potential (Table 3;Fig.3).The other three varieties had several higher YPMs with increasing plant density,indicating that the full yield potential of the soybean varieties in this region could be maximized by optimizing the population density for high-yield varieties.These results indicated the possibility of increasing yield by optimizing the population density for high-yield varieties with more precise and sophisticated field management techniques,and this approach may be more suitable for the main-stem varieties.

5.Conclusion

Planting density experiments in the Huang-Huai ecological region showed that compared with control germplasms,M657 has excellent tolerance to both high density and lodging.Short petioles play an important role in the tolerance to high density and lodging of M657,and M657 is thus an excellent germplasm resource for the development of new density-tolerant and high-yield soybean varieties.Our study discussed the yield-related phenotypes and yield changes for different germplasms under different locations and densities,and analyzed the effects of plant types and planting density on yield components and the feasibility of increasing yield by enhancing planting density based on the conventional planting density.Therefore,this study provides a germplasm and the theoretical basis for the use of realistic dense planting to increase and maintain high yields of soybeans in the Huang-Huai region.

Acknowledgements

This study was funded by the National Natural Science Foundation of China (31271753),the Central Publicinterest Scientific Institution Basal Research Fund,China(S2021ZD02) and the Agricultural Science and Technology Innovation Program (ASTIP) of the Chinese Academy of Agricultural Sciences (CAAS-ZDRW202003-1).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm