ZHENG Chang,WANG Yue-chao,XU Wen-ba,YANG De-sheng,YANG Guo-dong,YANG Chen,HUANG Jian-liang,PENG Shao-bing

National Key Laboratory of Crop Genetic Improvement/Hubei Hongshan Laboratory/MARA Key Laboratory of Crop Ecophysiology and Farming Systems in the Middle Reaches of the Yangtze River/College of Plant Science and Technology,Huazhong Agricultural University,Wuhan 430070,P.R.China

Abstract The border effect (BE) is widely observed in crop field experiments,and it has been extensively studied in many crops.However,only limited attention has been paid to the BE of ratoon rice.We conducted field experiments on ratoon rice in Qichun County,Hubei Province,Central China in 2018 and 2019 to compare the BE in the main and ratoon crops,and to quantify the contribution of BE in the main crop to that in the ratoon crop.The BE of two hybrid varieties was measured for the outermost,second outermost,and third outermost rows in each plot of both crops.To determine the contribution of BE between the two crops,portions of hills in the outermost and second outermost rows were uprooted during the harvest of the main crop so that the second and third outermost rows then became the outermost rows in the ratoon crop.Overall,the BE on grain yield was greater in the main crop than in the ratoon crop.In the main crop,the BE on grain yield was 98.3% in the outermost row,which was explained by the BE on panicles m–2,spikelets/panicle,spikelets m–2,and total dry weight.In the ratoon crop,the BE on grain yield was reduced to 60.9 and 27.6% with and without the contribution of the BE in the main crop,respectively.Consequently,55.1% of the BE on grain yield in the ratoon crop was contributed from the main crop.High stubble dry weight and non-structural carbohydrate (NSC) accumulation at the harvest of the main crop were responsible for the contribution of BE in the main crop to that in the ratoon crop.Our results suggest that increases in stubble dry weight and NSC accumulation at the harvest of the main crop could be important strategies for developing high-yielding cropping practices in the rice ratooning system.

Keywords: border effect,grain yield,non-structural carbohydrate,ratoon rice

1.lntroduction

The plant performance at the periphery of a plot is usually different from that at the center of the plot,and this difference can be defined as the border effect (BE) (Sato and Takahashi 1983).Such an effect is often induced by uncropped space,which is left between adjacent plots to facilitate crop management and to avoid seed mixing (Austin and Blackwell 1980;May and Morrison 1985).Typically,the border rows next to the uncropped space have higher grain yield than the center rows as a result of better ventilation,more light interception,and less nutrient competition (Tenget al.2008).Due to this yield advantage,discarding the border rows is recommended for accurately determining plot yield (Gomez 1972).On the other hand,maximizing the border advantage is important for increasing productivity in wide–narrow row or skip row planting patterns (Huet al.2020).Therefore,BE has been studied in many crops such as rice,maize,wheat,cotton,soybeans,and various others (Hartwiget al.1951;Romaniet al.1993;Du and Wang 1998).

For rice,which is one of the most important staple crops,previous studies have shown significant BE caused by uncropped space.One study reported that the yield increase of border rows over the center rows varied from 18.2 to 46.6% among 14 varieties (He 2011).Gomez and De Datta (1971) observed a BE of 63–159% by analyzing yield differences between border and center rows in eight rice yield trails.Similarly,Wanget al.(2013) found that the BE on grain yield ranged from 47.0 to 163.0% in two hybrid rice varieties.However,information on BE is currently limited for ratoon rice.

Ratoon rice involves the production of a second crop(ratoon crop) from the stubble left after the first crop (main crop) is harvested (Peng 2014;Wanget al.2021).The use of the ratoon crop is proposed as a promising system to increase rice production because additional grain yield can be obtained from the ratoon crop without the activities of land preparation,sowing,and transplanting (Douet al.2016;Peng 2017;Shenet al.2020).The grain yield of the ratoon crop is usually as high as 60% of the main crop yield (Wanget al.2020).The yield difference between the main and ratoon crops is the result of differences in growth duration and plant morphology.Compared with the main crop,the ratoon crop matures much earlier and the growth duration is shortened by nearly 40% (Donget al.2017).Also,the ratoon crop has lower panicle size,leaf area index,plant height,and biomass production than the main crop (Vergaraet al.1989;Wang Y Cet al.2019).However,the panicle number of the ratoon crop is generally higher than that of the main crop (Xuet al.2015).These differences in growth duration and canopy size may lead to differences in light interception and nutrient competition,resulting in different magnitudes of BE between the main and ratoon crops.

It is well known that the success of the ratoon crop largely depends on the performance of the main crop (Negaluret al.2017).For example,the yield performance of the ratoon crop is primarily determined by the number of panicles that develop from nodal buds sprouting on the stubble left from the main crop (Xie 2010).The nutrients,including carbohydrate and N stored in stubble,play major roles in supporting the growth of regenerated buds (Jianget al.2003;Torreset al.2020).The BE of the main crop may change stubble carbohydrate and N levels,which further influence the tiller growth of the ratoon crop,and consequently the BE on grain yield of the ratoon crop.Therefore,the BE of the ratoon crop would depend on the BE in the main season and the crop performance during the ratoon season.However,the quantitative contribution of BE in the main crop to that in the ratoon crop remains unclear.Therefore,the objectives of this study were to quantify the difference in BE between the main and ratoon crops as well as the contribution of BE in the main crop to that in the ratoon crop.

2.Materials and methods

2.1.Experimental sites and plant materials

A field experiment was conducted at Jiupu Village (30°14´N,115°25´E),Chidong Town,Qichun County,Hubei Province,Central China in 2018 and repeated in the same field in 2019.The soil texture of the experimental field was clay loam with pH 4.82,31.34 g kg-1organic matter,2.20 g kg-1total N,11.15 mg kg-1Olsen P,and 174.29 mg kg-1available K at the 0–20 cm layer.Anindicahybrid variety,Liangyou 6326(LY6326),and anindica/japonicahybrid variety,Yongyou 4949(YY4949),were used in this study.Both varieties have been widely grown for ratoon rice in Central China because of their high yield potential,strong ratooning ability,good grain quality,and suitable growth duration.

2.2.Experimental design and crop management

The two varieties were arranged in a randomized complete block design with four replications.Thirty-fiveday-old seedlings from wet bed nurseries were manually transplanted on 27 April in both years.Transplanting was done at a row spacing of 0.3 m and a plant spacing of 0.133 m with two seedlings per hill.Each plot contained thirteen 8.0-m rows (60 hills per row) with a north–south row direction.On either the east or west side of a plot,a 1.5-m-wide uncropped space (including a 0.3-m-wide alley)was left between adjacent plots and used for measuring BE.In the main crop,the outermost,second outermost,and third outermost rows in the BE measuring area of each plot were recorded as M1,M2,and M3,respectively (Fig.1-A).The main crop was harvested for both varieties on 31 July in 2018 and 8 August in 2019.During the harvest of the main crop,a portion of the hills in the outermost and second outermost rows were uprooted to prevent ratooning,so the second and third outermost rows became the outermost rows in the ratoon crop and the BE during the ratoon season with and without the contribution of the BE in the main crop could be determined.The remaining hills were cut for ratooning and the height of stubble left on the field was about 45 cm.The M1,M2,and M3corresponded to M1R1,M2R1,and M3R1in the ratoon crop,respectively (Fig.1-B).The center rows of a plot in the main and ratoon crops were recorded as MCand MCRC,respectively.The difference in BE between the main and ratoon crops could be determined by comparing the BE values between M1and M1R1,and the contribution of BE between the two crops could be determined by comparing the BE values between M1R1and either M2R1or M3R1.The growth and yield performance of M1R1,M2R1,and M3R1in the ratoon crop are illustrated in Fig.2.The ratoon crops were harvested for LY6326 and YY4949 on 27 October and 3 November,respectively,in 2018,and on 28 October and 2 November,respectively,in 2019.

Fig.1 Schematic illustrations of row placements in a plot in the main (A) and ratoon (B) crops.The blank spaces in a plot in the ratoon crop represent places where the plants were uprooted at the maturity of the main crop.The dotted boxes represent the sampling areas at the maturity of the two crops.M1,M2,and M3 are the outermost,second outermost,and third outermost rows of a plot in the main crop,respectively;M1R1,M2R1,and M3R1 are the outermost rows of a plot in the ratoon crop corresponding to M1,M2,and M3 in the main crop,respectively;MC and MCRC are the center rows of a plot in the main and ratoon crops,respectively.

Fig.2 Photographs of the ratoon crop taken at 5 days after the harvest of the main crop (A),15 days after the harvest of the main crop (B),and 3 days after the heading of the ratoon crop(C) in Qichun County,Hubei Province,Central China,in 2019.

Total nutrient inputs were 200 kg N ha–1,40 kg P ha–1and 60 kg K ha–1for the main crop,and 150 kg N ha–1and 60 kg K ha–1for the ratoon crop.In the main crop,40% of the N was applied and incorporated into the soil together with all of the P and K at 1 d before transplanting,and the remaining N was equally split at 7 d after transplanting and panicle initiation as topdressings.For the ratoon crop,50%of the N and all of the K were applied at 10–12 d after the heading of the main crop as bud-promoting fertilizer,and the other 50% of the N was applied at 3 d after the harvest of the main crop as tiller-promoting fertilizer.The sources of N,P,and K were urea,calcium superphosphate,and potassium chloride,respectively.The field was flooded with a water depth of 3–5 cm from transplanting to 10–12 d before maturity in the main crop,except that the water was drained at the mid-tillering stage.After harvesting the main crop,irrigation was immediately carried out with a water depth of only 1–2 cm until the regenerated buds grew to more than 10 cm,and then a water depth of 3–5 cm was maintained until the middle grain filling stage of the ratoon crop.Diseases,insects,birds,and weeds were intensively controlled to avoid yield losses.

2.3.Sampling and measurements

At the maturity of the main crop,12-hill samples were uprooted from M1,M2,M3,and MCin each plot.At the maturity of the ratoon crop,10-hill samples were taken from M1R1,M2R1,M3R1,and MCRCin each plot (Fig.1-B).Only 10 of the hills were sampled in the ratoon crop because there were 14 hills in M2R1and M3R1,and the two hills at each end of M2R1and M3R1were not sampled to avoid BE in the north and south sides.All plant samples were divided into stubble (45 cm),straw (the remaining culm,sheath,and leaves),and panicles.The panicles were hand-threshed after recording panicle number,and then the filled spikelets were separated from unfilled spikelets by submerging them in tap water.After air-drying for three days,the filled spikelets were weighed to determine the fresh weight of filled spikelets m–2,and three subsamples each with 30 g of filled spikelets were counted to calculate the average number of filled spikelets g–1.The number of filled spikelets m–2was then calculated (fresh weight of filled spikelets m–2×average number of filled spikelets g–1).To determine the number of unfilled spikelets m–2,all unfilled spikelets were counted.Spikelet number m–2(filled spikelet number m–2+unfilled spikelet number m–2),spikelet number per panicle (spikelet number m–2/panicle number m–2),and grain-filling percentage (100×filled spikelet number m–2/spikelet number m–2) were calculated.The dry weights of stubble,straw,rachis,and filled and unfilled spikelets were measured after oven-drying at 80°C to a constant weight for determining total dry weight,harvest index (100×filled spikelet weight/total dry weight),and 1 000-grain weight.Grain yield was calculated based on the dry weight of filled spikelets and expressed as grain yield per hectare (adjusted to 14% moisture).The BE was calculated as follows:

The contribution of BE in the main crop to BE in the ratoon crop (CBE) was calculated as follows:

The samples of stubble in the main crop were ground for measurements of non-structural carbohydrate (NSC) and N concentrations.The NSC concentration was determined as described by Yoshidaet al.(1976).The N concentration was measured using an elemental analyzer (Elementar vario MAX CNS/CN,Elementar Trading Co.,Ltd.,Germany).Stubble NSC or N content m–2was calculated as the product of stubble NSC or N concentration and stubble dry weight m–2.

2.4.Data analysis

Data were summarized as the means of four replications for each row.Mean comparisons for each variety within a year were made according to the LSD (P＜0.05) on the basis of the analysis of variance (Statistix 9.0,Analytical Software,Tallahassee,FL,USA).

3.Results

3.1.Crop growth duration

The ratoon crop reached physiological maturity in 81–95 d,which was 59–73% of the total growth duration of the main crop (Table 1).For the two varieties,although the total growth duration was the same in the main crop,YY4949 had a 5–7 d longer total growth duration than LY6326 in the ratoon crop.The total growth duration of the main crop was longer in 2019 than in 2018,but the reverse was true in the ratoon crop.The whole growth durations of the two crops ranged from 218 to 225 d.

3.2.Grain yield and its border effect

In the main crop,the grain yield of M1was significantly higherthan those of M2,M3,and MC(Table 2).Among M2,M3,and MC,the grain yield of M2was consistently lower than those of M3and MC,although the difference was only significant for LY6326 in 2018.There was small and insignificant difference in grain yield between M3and MC.Therefore,the BE on grain yield was positive for M1but negative for M2,and there was almost no BE on grain yield for M3.On average,the BE on grain yield values were 98.3,–10.3,and–1.9%for M1,M2,and M3,respectively.Inconsistent differences in BE on grain yield were found between the two varieties and between the two years.In the ratoon crop,M1R1producedthe highest grain yield,followed by M3R1,M2R1,and MCRC(Table 2).The BE on grain yield was positive for all three rows.On average,the BE on grain yield values were 60.9,9.0 and 27.6% for M1R1,M2R1,and M3R1,respectively.The BE on grain yield was higher in 2019 than in 2018,whereas a relatively small difference in BE on grain yield was observed between the two varieties.

Table 1 Growth durations in the main crop (from sowing to harvest of the main crop),the ratoon crop (from harvest of the main crop to harvest of the ratoon crop),and the two crops (from sowing of the main crop to harvest of the ratoon crop) for the two varieties grown in 2018 and 20191)

Table 2 Grain yield and border effect (BE) on grain yield of the main and ratoon crops for the two varieties grown in 2018 and 2019

3.3.Yield-related traits and their border effects

In the main crop,panicles m–2,spikelets/panicle,spikelets m–2,and total dry weight of M1were significantly higher than those of M2,M3,and MC(Table 3).The grain-filling percentage and harvest index of M1were consistently higher than the other rows in 2018,but not in 2019.There were inconsistent differences in 1 000-grain weight among the four rows.In the ratoon crop,panicles m–2,spikelets m–2,and total dry weight of M1R1were significantly higher than those of M2R1,M3R1,and MCRC(Table 4).On contrast,M1R1did not show significantly higher spikelets/panicle,grain filling percentage,1 000-grain weight,or harvest index values than the other rows.The spikelets m–2and total dry weight of M3R1were significantly higher than those of MCRC,but these two rows had insignificant differences in panicles m–2,spikelets/panicle,grain filling percentage,and 1 000-grain weight.The harvest index was higher in M3R1than in MCRC,but the difference was significant only in 2019.

The values for BE on the yield-related traits of both crops are shown in Table 5.In the main crop,there were consistently positive BE on panicles m–2,spikelets/panicle,spikelets m–2,and total dry weight.However,the BE on grain-filling percentage,1 000-grain weight,and harvest index were relatively small and inconsistent across varieties and years.In the ratoon crop,there were consistently positive BE on panicles m–2,spikelets m–2,and total dry weight in M1R1and M3R1,and the BE on these three traits were higher in M1R1than in M3R1.In both M1R1and M3R1,the BE on panicles m–2,spikelets m–2,and total dry weight were higher in 2019 than in 2018,while relatively small differences in BE on these traits were observed betweenthe two varieties.The BE on spikelets/panicle,grain filling percentage,1 000-grain weight,and harvest index were small and inconsistent across varieties and years in both M1R1and M3R1.

Table 3 Yield-related traits of the different rows in a plot for the two varieties grown in the main crop of 2018 and 2019

Table 4 Yield-related traits of different rows in a plot for the two varieties grown in the ratoon crop of 2018 and 2019

3.4.lmpact of main crop on border effects of ratoon crop

The contribution of BE in the main crop to that in the ratoon crop (CBE) was determined by comparing the BE between M1R1and M3R1.The CBE values were calculated for grain yield and the three yield-related traits which had consistent BE in both M1R1and M3R1(Table 6).The CBE for grain yield ranged from 53.3 to 58.6%,with an average of 55.1%.The CBE for panicles m–2,spikelets m–2,and total dry weight were greater than those for grain yield.There were inconsistent differences in the CBE for grain yield and the three yield-related traits between the two varieties and between the two years.

Table 5 Border effects (%) on yield-related traits of different rows for the two varieties grown in the main and ratoon crops of 2018 and 2019

Table 6 Contributions of border effects in the main crop to that in the ratoon crop (CBE) for the two varieties grown in 2018 and 2019

At the maturity of the main crop,stubble dry weight perstem and per m2of M1were significantly higher than those of M2,M3,and MC(Fig.3).Similar results were observed for the stubble NSC concentration and content m–2(Fig.4).The differences in stubble dry weight and NSC content per m2between M1and the other three rows were greater than those in stubble dry weight per stem and NSC concentration,respectively (Figs.3 and 4).The stubble N content m–2of M1was also significantly higher than those of M2,M3,and MC,whereas there were small and inconsistent differences in stubble N concentrations among the four rows (Fig.5).Regression analysis showed that the BE on grain yield of the ratoon crop was significantly and positively related to stubble dry weight,NSC concentration and content,and N content (Fig.6).The coefficients of determination were the highest for stubble NSC concentration and content,followed by stubble dry weight and N content.There was an insignificant relationship between the BE on grain yield of the ratoon crop and stubble N concentration.

Fig.3 Stubble dry weight per stem (A and B) and per m2 (C and D) of different rows in a plot for the two varieties grown in the main crop of 2018 (A and C) and 2019 (B and D). Data are mean±standard error(SE,n=4).Bars followed by the same letters represent nonsignificant differences according to LSD (0.05) among different rows within the same variety.LY6326 and YY4949 are Liangyou 6326 and Yongyou 4949,respectively;M1,M2,M3,and MC are the outermost,second outermost,third outermost,and the center rows of a plot in the main crop,respectively.

Fig.4 Stubble non-structural carbohydrate (NSC) concentration (A and B) and content m–2 (C and D) of different rows in a plot for the two varieties grown in the main crop of 2018 (A and C) and 2019 (B and D).Data are mean±standard error (SE,n=4).Bars followed by the same letters represent nonsignificant differences according to LSD (0.05)among different rows within the same variety.LY6326 and YY4949 are Liangyou 6326 and Yongyou 4949,respectively;M1,M2,M3,and MC are the outermost,second outermost,third outermost,and the center rows of a plot in the main crop,respectively.

Fig.5 Stubble N concentration (A and B) and content m–2 (C and D) of different rows in a plot for the two varieties grown in the main crop of 2018 (A and C) and 2019 (B and D).LY6326 and YY4949 are Liangyou 6326 and Yongyou 4949,respectively;M1,M2,M3,and MC are the outermost,second outermost,third outermost,and the center rows of a plot in the main crop,respectively.Data are mean±standard error (SE,n=4).Bars followed by the same letters represent nonsignificant differences according to LSD (0.05) among different rows within the same variety.

Fig.6 Relationships of border effect (BE) on grain yield of the ratoon crop with stubble dry weight per stem and per m2 (A and B),stubble non-structural carbohydrate (NSC) concentration and content per m2 (C and D),and stubble N concentration and content m–2 (E and F) at the harvest of the main crop (n=12).Data were pooled from the experiments conducted in 2018 and 2019.M1R1,M2R1,and M3R1 are the outermost rows of a plot in the ratoon crop corresponding to M1,M2,and M3 in the main crop,respectively.The symbols of ns and ＊＊ represent insignificance at the 0.05 the probability level and significance at the 0.01 probability level,respectively.

4.Discussion

4.1.Comparison between two crops in border effect

Our study demonstrated a large positive BE on grain yield in the outermost row (M1) of the main crop.On average,the BE on grain yield in M1was 98.3%,which was similar to the values of 97.0–98.0% observed for rice variety IR8 growing in wet and dry seasons by Gomez and De Datta (1971).The BE on grain yield was also reported in previous research on ordinary rice (Liuet al.2006;Wanget al.2013),although the BE on grain yield varied between 9.9 and 163.0% in those studies.There was also a positive BE on grain yield in M1R1in the ratoon crop for the plants regrown from the stubble left from M1in the main crop,but the magnitude was reduced to 60.9%.Interestingly,the BE on grain yield in the ratoon crop was 62.0% of the BE in the main crop in the outermost row,which is consistent with the fact that the grain yield of the ratoon crop can reportedly reach as high as 60% of that of the main crop in a well-managed ratoon crop (Wang Y Cet al.2019).

The BE on grain yield in M1of the main crop was attributed to higher spikelets m–2and total dry weight.Both panicles m–2and spikelets/panicle contributed to higher spikelets m–2of M1,although the contribution from panicles m–2was larger than that from spikelets/panicle.These findings on the contributions of yield-related traits to the BE on grain yield were also reported in other studies (Sato and Takahashi 1983;Wanget al.2013).In the ratoon crop,the BE on grain yield in M1R1was also attributed to higher spikelets m–2and total dry weight.However,panicles m–2contributed mainly to the higher spikelets m–2in M1R1.The greater BE on grain yield in the main crop than in the ratoon crop was supported by the facts that the BE on spikelets m–2and total dry weight in M1were larger than those in M1R1.Smaller BE in the ratoon crop could be due to a smaller difference in the ventilation and light environment between border and inner rows compared with the main crop.This is because the ratoon crop has a smaller canopy with a lower leaf area index than the main crop (Yiet al.2009).Moreover,the shorter growth duration of the ratoon crop could also explain the smaller BE in the ratoon crop compared with the main crop (Donget al.2017).However,the BE on panicles m–2of M1R1was generally larger than that of M1.This confirmed the fact that more panicles are of great importance for achieving higher yield in the ratoon crop (Xie 2010).

4.2.Border effects of ratoon crop as influenced by main crop

Positive BEs on grain yield were also observed in M2R1and M3R1in the ratoon crop,although they were much smaller than that in M1R1.The BE on grain yield was smaller in M2R1than in M3R1,which was caused by a more negative BE on grain yield in M2than in M3(–10.3%vs.–1.9%) in the main crop.The negative BE in M2and M3in the main crop probably resulted from the competition from the very vigorous plants in M1(Gomez and De Datta 1971).Because the BE on grain yield in M3in the main crop was close to zero,the BE on grain yield in M3R1in the ratoon crop (27.6%)could be considered as the BE during the ratoon season alone,without any contribution of BE in the main crop.Thus,the difference in the BE between M1R1and M3R1was used to quantify the contribution of BE in the main crop to that in the ratoon crop (CBE) in this study.The average CBE for grain yield,panicles m–2,spikelets m–2,and total dry weight were 55.1,76.5,64.3,and 69.9%,respectively,suggesting that the main crop had a substantial contribution to the BE of the ratoon crop in the outermost row.To identify the underlying mechanisms of CBE,the differences in dry matter accumulation and nutrient status of the stubble left in the field among different rows were determined in this study.At the maturity of the main crop,M1showed significantly higher values for stubble dry weight per stem and per m2and stubble NSC concentration and content m–2than the inner rows.A similar result was observed for stubble N content m–2but not for stubble N concentration.The substantial improvements in stubble dry weight and NSC accumulation could be attributed to a stronger photosynthetic capability of the outermost row during the main season (Romaniet al.1993;Wanget al.2017).The regression analysis indicated that stubble dry weight and NSC accumulation were strongly related to the BE on grain yield of the ratoon crop.Previous studies stated that high stubble dry weight at the maturity of the main crop could provide an abundant carbohydrate source for bud growth,thus increasing panicles m–2and enhancing the grain yield of the ratoon crop (Nakanoet al.2020;Torreset al.2020).We also observed that the buds of the outermost row sprouted earlier and grew faster than those of the inner rows after the harvesting of the main crop (data not shown).

4.3.Variations in border effects and their implication for crop management of ratoon rice

The magnitudes of BE for rice crops may differ with variety and year (Austin and Blackwell 1980;Chenet al.2006).In this study,however,there were no differences in BE on grain yield between the two varieties in either the main or ratoon crops.Both varieties showed great BE on grain yield because they had a strong tillering ability or large panicles(Chenet al.2006).In addition,a consistent difference in BE on grain yield between the two years was observed in the ratoon crop,but not in the main crop.The greater BE on grain yield of the ratoon crop in 2019 could be explained by its higher levels of grain yield,stubble dry weight,and NSC accumulation in the main crop compared with 2018(Table 2;Figs.3 and 4).

The BE is usually studied for two purposes: (1) avoiding the over-estimation of crop yield in field experiment plots(Gomez 1972) and (2) improving crop productivity by using skip-row and rectangular planting patterns (Huet al.2020).Skip-row planting is a common practice to facilitate mechanical operations,so that the tracks or wheels do not damage plants (Hoodet al.1991;Wang F Yet al.2019).In mechanized rice ratooning practices,the grain yield of the ratoon crop could be greatly reduced by the crushing damage to the stubble from mechanical harvesting of the main crop (Lin 2019;Zhenget al.2021).The skip-row planting of the main crop is proposed as a way to avoid the crushing damage.However,one concern is that there would be yield reductions in both the main and ratoon crops due to the decreased hill density in skip-row planting.At issue is whether these yield reductions of the two crops could be compensated by the BE of the rows next to the traffic paths.Our results indicated that the BE on grain yield existed in both the main and ratoon crops,and the BE in the ratoon crop could be substantially increased by the contribution of the BE in the main crop.This implies that it is possible to adopt skip-row planting in the mechanized rice ratooning system for increasing the grain yield of the ratoon crop,especially by coupling it with suitable varieties and proper water and N management practices.Rectangular or skip-row planting was recommended for rice varieties with strong tillering ability and large panicles to maximize grain yield (Liuet al.2006).Reports indicate that proper water and N management practices can be effective in improving the BE on grain yield in both the main and ratoon crops by increasing tiller number and panicle size (Yaoet al.2012;Jianget al.2015;Wanget al.2022).Further research is needed to confirm whether the BE on grain yield can over-compensate for the yield loss due to the reduced hill density from skip-row planting in mechanized rice ratooning systems.

5.Conclusion

The BE on grain yield was greater in the main crop than in the ratoon crop.In the main crop,the BE on grain yield was 98.3% in the outermost row,which was explained by the BE on panicles m–2,spikelets/panicle,spikelets m–2,and total dry weight.In the ratoon crop,the BE on grain yield was reduced to 60.9 and 27.6% with and without the contribution of BE in the main crop,respectively.Consequently,55.1% of the BE on grain yield in the ratoon crop was contributed from the main crop.High stubble dry weight and NSC accumulation at the harvest of the main crop were responsible for the contribution of BE in the main crop to that in the ratoon crop.Our results suggest that increases in stubble dry weight and NSC accumulation at the harvest of the main crop could be important strategies for developing high-yielding cropping practices in the rice ratooning system.

Acknowledgements

This work was supported by the Major International(Regional) Joint Research Project of National Natural Science Foundation of China (32061143038),the earmarked fund for China Agriculture Research System (CARS-01-20),the National High Technology Research and Development Program of China (the 863 Project,2014AA10A605),and the Fundamental Research Funds for the Central Universities,China (2662020ZKPY015).

Declaration of competing interest

The authors declare that they have no conflict of interest.