Effects of Mechanically Transplanting Methods and Planting Densities on Yield and Quality of Nanjing 2728 under Rice-Crayﬁsh Continuous Production System

: Rice–crayﬁsh continuous production system offers high economic and ecology beneﬁts, which developed rapidly in China. To investigate the effects of different mechanical transplanting methods and planting densities on rice yield and quality, Nanjing 2728 was used to determine rice growth performance under mechanically transplanted carpet seedling (MTCS) with equal row spacing (30 cm) at ﬁve spacings and mechanically transplanted pot seedling (MTPS) with wide and narrow rows (23 + 33 cm) at ﬁve spacings. The results showed that MTPS presented signiﬁcantly higher rice yields than MTCS as more spikelets per panicle. Rice yields of both mechanical transplanting methods ﬁrst increased and then reduced with decreasing planting density, and its highest value was obtained at 77.9 × 10 4 seedlings ha − 1 . Compared with MTCS at the same stage, rice tiller dynamics of MTPS ﬁrst increased and then decreased. Additionally, its dry matter accumulation per stem at jointing, heading, and maturity stages, leaf area index, photosynthetic potential, crop growth rate, and net assimilation rate were all signiﬁcantly higher relative to MTCS. For each mechanical transplanting method, dry matter accumulation per panicle, leaf area index, photosynthetic potential, crop growth rate, and net assimilation rate from the sowing to jointing stages declined with decreasing planting density, while dry matter accumulation per stem and net assimilation rate from the heading to maturity stages increased. Compared with MTCS, MTPS signiﬁcantly improved rice milling and appearance quality, decreasing density was also beneﬁcial to rice milling and appearance quality, while grain content of amylose and protein were not sensitive to both transplanting method and planting density. Consequently, MTPS with 13.8 cm plant spacing is a suitable mechanical transplanting method for Nanjing 2728 to obtain better yield and quality under rice–crayﬁsh continuous production system.


Introduction
Rice (Oryza sativa L.) is one of the most important grain crops in China, its yield performance affects national food security [1,2]. However, traditional rice cultivation could not bring high profits to farmers in China due to its general purchasing price, which negatively affected farmers' enthusiasm for planting rice. Rice-crayfish (Procambarus clarkii) integrated farming is an integrated agricultural production system that uses the wetland resources of paddy field to plant rice and raise crayfish in the same field, which could increase farmers' income relative to traditional rice farming system as the booming

Experiment Design
The experiment was arranged in a split plot design, with the mechanical transplanting method as the main plot and the planting densities as split plots. The experiment was conducted with four replications. The size of each subplot was 20 m 2 . Two mechanical transplanting methods were tested: Mechanically transplanted carpet seedling (equal spacing, 30 cm, A) and mechanically transplanted pot seedling (wide narrow row, 23 cm + 33 cm, B). Five plant spacing treatments were set for each mechanical transplanting method. The two mechanical transplanting methods and their plant spacing combinations are shown in Table 1.
The carpet seedling method uses plastic floppy disk to cultivate seedling. They were sown on June 7 in both years, when about 120 g seeds were mechanically sown in each tray. The pot seedling method also adopts plastic floppy disk dry seedling raising. They were sown on May 28 in both years, using the 2bd-600 (Ispe-60am) pot seedling seeder, which sows 4-5 seeds per hole, ensuring at least three basic seedling per hole. For both mechanically transplanting methods, the seedlings were transplanted on June 28. The carpet and pot seedling were harvested on November 3, and October 29, respectively.

Experiment Design
The experiment was arranged in a split plot design, with the mechanical transplanting method as the main plot and the planting densities as split plots. The experiment was conducted with four replications. The size of each subplot was 20 m 2 . Two mechanical transplanting methods were tested: Mechanically transplanted carpet seedling (equal spacing, 30 cm, A) and mechanically transplanted pot seedling (wide narrow row, 23 cm + 33 cm, B). Five plant spacing treatments were set for each mechanical transplanting method. The two mechanical transplanting methods and their plant spacing combinations are shown in Table 1. The carpet seedling method uses plastic floppy disk to cultivate seedling. They were sown on June 7 in both years, when about 120 g seeds were mechanically sown in each tray. The pot seedling method also adopts plastic floppy disk dry seedling raising. They were sown on May 28 in both years, using the 2bd-600 (Ispe-60am) pot seedling seeder, which sows 4-5 seeds per hole, ensuring at least three basic seedling per hole. For both mechanically transplanting methods, the seedlings were transplanted on June 28. The carpet and pot seedling were harvested on November 3, and October 29, respectively. Nitrogen fertilizer application rate was 225 kg ha −1 for all treatments, which was applied at a ratio of 4:3:3 as base, tillering, and panicle initiation, respectively. Compound fertilizer (N:P 2 O 5 :K 2 O = 15%:15%:15%) was used as the base fertilizer. The prevention and control of diseases, insect pests, and weeds followed the green prevention and control plan for local rice-crayfish comprehensive cultivation. Biological pesticides are mainly used to prevent and control rice diseases and insect pests, some chemical pesticides with high efficiency and low toxicity were applied, solar insect killing lamps are installed at the edges of the fields, and sex attractants are arranged in the field to control rice stem borer. Vetiver grass was planted around the experimental fields to control borers and other pests. Weeds were removed manually.

Rice Tiller Dynamic
A series of 20 consecutive plants were marked in the same growth of each plot to count the number of tillers for every 7 days after transplanting.

Dry Matter Accumulation and Leaf Area Index
To assess the average number of stems and tillers per plot at the jointing, heading, and maturity stages, plants from five representative hills in each plot were sampled. The length-width coefficient method was used to measure the leaf area of the jointing and heading stages. After that, each sample was separated into the green leaves, stems plus sheaths, and panicles and placed in kraft paper bags in an oven; rice plants were oven-dried separately at 105 • C for 30 min and then at 80 • C in the bags until they reached a constant weight. The dry matter weight of each part of the ground was measured.

Yield and Its Components
The number of panicles was determined from 100 representative squares randomly sampled from each plot at the maturity stage. According to the average number of panicles per plant, the rice panicles of three whole plants were collected and placed into net bags to determine grain number per panicle, seed setting rate, and thousand grain weight. At maturity stage, 50 holes were harvested continuously in each plot, except for three rows. An LDS-1G grain moisture meter was used to measure the grain moisture content, remove impurities, and convert the actual yield to 14.5% moisture content.

Rice Quality
The harvested rice plants from each plot were threshed, air-dried, and then stored indoors for 90 days. They were air-selected with an NP-4350 winnowing machine and mixed with three samples, each weighing 200 g. Brown rice rate, milled rice rate, head rice rate, chalky rate, chalky area, chalkiness, and amylose content were measured according to the NY/T83 guidelines [14]. Protein content was measured with a FOSS Kjeldahl nitrogen analyzer.

Statistical Analysis
Tables and graphs were conducted using Microsoft Excel 2013 and Origin 9.1, and the analysis of the relevant data was performed using SPSS 20.0. Differences were considered significant at p < 0.05 using least significant difference.

Rice Yield and Its Components
As shown in Table 2, mechanical transplanting methods and planting densities exhibited significant effects on rice yield and its components. At the same planting density, the pot seedling showed significantly higher grain yields, by 3.9-11.4% in 2018 and 3.5-7.0% in 2019, than the carpet seedling. With decreasing planting density, rice yield of carpet seedling first increased and then decreased. The highest yields of A2 were 10.17 t ha −1 in 2018 and 10.69 t ha −1 in 2019, respectively. Rice yield of the pot seedling was consistent with that of the carpet seedling. In 2018 and 2019, the yields of B2 were 10.62 t ha −1 , and 11.24 t ha −1 , respectively, 4.4% and 5.1% higher than those of the carpet seedling at the same density. According to a general survey of lodging in the field plots, the lodging rates of A1 and B1 differed (10-37%) between the two years, leading to a decrease in yield. According to a yield components analysis at the same density, there were no significant differences in panicle or 1000-grain weight between the two mechanical transplanting methods. The pot seedling had significantly more spikelets per panicle than the carpet seedling. The seed setting rate also seemed to be slightly higher, but the difference was not significant. The results indicate that the higher number of spikelets per panicle was the main factor for increasing the yield of the pot seedling relative to the carpet seedling.
The ANOVA results showed significant differences in the panicle, spikelets per panicle, and filled grain rate between years ( Table 2). Except for the number of panicles, which did not differ significantly between the two mechanical transplanting methods, the differences in yield and its components between the mechanical transplanting methods and among the different planting densities were extremely significant. The interaction of year and planting density, the mechanical transplanting method, and the planting density had significant effects on the panicle. The interaction between mechanical transplanting method and planting density had no significant effect on the filled grain rate, but the interactions of the other two or three factors significantly affected filled grain rate.

Rice Tiller Dynamic
Tiller dynamics of both two mechanical transplanting methods first increased and then decreased ( Figure 2). For Mechanical transplanting with carpet seedling, there was no obvious tillering at 7 days after transplanting, and the number of tillers increased rapidly after 14 days, reaching a peak at 28 days after transplanting and then decreased rapidly. The overall curve shows an initial flat trend, followed by a rapid rise and drop. For Mechanical transplanting with pot seedling, there was no obvious slow seedling stage in the first 7 days after transplanting, the tillers did not increase obviously during the first 28 days after transplanting, and the number of pot seedling was slightly lower than the number of carpet seedling. With decreasing planting density, the number of tillers at each stage decreased. The peak seedling number of the two mechanical transplanting methods showed the same trends as their numbers of panicles at maturity.

Leaf Area Index and Decreasing Rate of Leaf Area at Grain-Filling Stage
As shown in Table 3, the leaf area index of the carpet seedling was low the pot seedling at the jointing, heading, and maturity stages, and it reduced

Leaf Area Index and Decreasing Rate of Leaf Area at Grain-Filling Stage
As shown in Table 3, the leaf area index of the carpet seedling was lower than that of the pot seedling at the jointing, heading, and maturity stages, and it reduced with decreasing density. The differences in leaf area index between the high-density and lowdensity treatments were significant under the same mechanical transplanting method, but the difference between the two methods at the same density were not significant, indicating that planting density had a greater effect on the leaf area index than mechanical transplanting method. The effect of the mechanical transplanting method on the leaf area attenuation rate was small; the attenuation rate of the pot seedling seemed to be slightly higher than that of the carpet seedling, but the difference was not significant at the same density. Under the same mechanical transplanting method, the leaf area attenuation rate at the grain-filling stage first increased and then decreased with decreasing density, but there were no significant differences between adjacent density treatments.

Photosynthetic Potential
As displayed in Table 4, with continued rice development and growth, the photosynthetic potential of both carpet and pot seedling increased. The results showed that the photosynthetic potential of the pot seedling was higher than that of the carpet seedling at three key growth stages. At the same density, the photosynthetic potential of the pot seedling was significantly higher than that of the carpet seedling at the sowingjointing stage. With decreasing planting density, the photosynthetic potential of both the carpet and pot seedling decreased at the sowing-jointing, jointing-heading, and heading-maturity stages.

Dry Matter Accumulation
Generally, mechanical transplanting method and planting density presented significant effects on rice dry matter accumulation ( Table 5). The dry matter accumulation of the pot seedling at the jointing, heading, and mature stages was higher than that of the carpet seedling, laying a material foundation for ears with larger and more grains. Compared with A2 and B2 treatments, dry matter accumulation of the pot seedling was 24.7% and Agronomy 2021, 11, 488 8 of 15 19.2% higher at the jointing stage, 11.1% and 8.8% higher at the heading stage, and 7.8% and 3.4% higher at the maturity stage in 2018-2019. Under the same mechanical transplanting method, dry matter accumulation at jointing and heading stages decreased as planting density lowered, and the change of dry matter accumulation at the maturity stage was the same as that of yield, which first increased and then decreased.

Per Panicle Dry Matter Accumulation
As shown in Table 6, the single-stem dry matter accumulation of the pot seedling was higher than that of the carpet seedling at each stage. In the B2 treatment, the singlestem dry matter accumulation at the jointing, heading, and maturity stages were 0.10, 0.21, and 0.36 g higher than those in the A2 treatment. The results showed that the pot seedling could form stronger individuals at the jointing stage, and this advantage became more obvious as development progressed. With decreasing density, the single-stem dry matter accumulation of the plants in both mechanical transplanting methods increased significantly. The results indicated that the maximum differences in the single-stem dry matter accumulation of the carpet seedling were 0.10 g, 0.70 g, and 1.56 g among densities at the same growth stage, respectively, while the maximum differences in the pot seedling were 0.19 g, 0.74 g, and 1.91 g, respectively.

The Crop Growth Rate and Net Assimilation Rate
As shown in Table 7, mechanical transplanting method and planting density had significant effects on crop growth rate at each key growth stage, and presented significant influence on net assimilation rate at the sowing-jointing stage. No significant effects on net assimilation rate at the jointing-heading and heading-maturity stages were detected. Except for the rice growth rate and net assimilation rate at the sowing-jointing stage, rice cultivated by pot seedling was superior to carpet seedling at the other growth stages. In both years, crop growth rate of the rice was the highest in the jointing-heading stage, followed by the heading-maturity stage, and lowest in the sowing-jointing stage. Under the same mechanical transplanting method, the crop growth rate declined with decreasing density at sowing-jointing, jointing-heading, and heading-maturity stages. In terms of the net assimilation rate, there was a significant difference between the two mechanical transplanting methods at the sowing-jointing stage, but there were no significant differences at the jointing-heading and heading-maturity stages. Under the same mechanical transplanting method, net assimilation rate at the sowing-jointing stage decreased as planting density reduced, but there were no significant differences between adjacent densities. There was no significant change in the net assimilation rate in the jointingheading stage, but in the heading-maturity stage, it first increased and then declined with decreasing density.

Rice Grain Quality
As presented in Table 8, there were no significant differences in brown rice rate between the carpet seedling and the pot seedling. However, the pot seedling exhibited better performance in their milled and head rice rates. The head rice rate increased by 3.3% on average, indicating that the rice processing quality of the pot seedling was better than that of the carpet seedling. Under the same mechanical transplanting method, the brown, milled, and head rice rates increased with decreasing density, and the head rice rate was significantly different between the high-density and low-density treatments in 2019.
In the analysis of rice appearance quality, chalky rate and chalkiness of the pot seedling were lower than those of the carpet seedling at the same density, but there was no obvious difference in chalky area. With decreasing planting density, the chalky rate and chalkiness of the pot seedling decreased, but the chalky area showed no obvious trend. This implies that increasing its transplanting distance could improve the appearance quality of rice.
Amylose content of the pot seedling seemed to be lower than that of the carpet seedling, but the difference was not significant. With decreasing density, the amylose content showed a downward trend. At the same planting density, rice protein content was not sensitive to the mechanical transplanting method. Under the same mechanical transplanting method, the protein content of the rice generally decreased with decreasing density, but the differences between adjacent densities were not significant.

Effect of Mechanical Transplanting Method and Planting Density on Grain Yield
Mechanically transplanted pot seedling and mechanically transplanted carpet seedling were two main mechanically transplanting methods, they had deveploped for more than 40, and 60 years, respectively. Carpet seedling transplanter had a lower one-time investment and was easier for famers to master, making it be adopted more widely, while mechanically pot seedling transplanter required higher one-time investment and more complex seedling cultivation technique, but this method was thought to have better adaptability to unfavourable growth environment and higher yield potential [15][16][17][18]. Our recent practices in rice cultivation under rice-crayfish integrated farming indicated that mechanically transplanted pot seedling had an advantage than mechanically transplanted carpet seedling in adapting the special working environment, in which carpet seedling transplanted were trapped into the soil because of long-term deep water-flooding.
Compared with the carpet seedling, the pot seedling showed a reduced number of effective panicles and a significantly increased number of spikelets per panicle, thus, increasing the spikelet number of rice crop and stabilizing the filled grain rate and 1000grain weight. Pot seedling is more suitable to obtain high rice yield through the route of "proper population, strong individual and high yield" [16,17]. Our results suggested that more dry matter could be accumulated by pot seedling during the early growth period after rice transplanting, and leaf area index at the middle and late stage was also higher than that of the carpet seedling, providing basis for obtaining larger rice panicles. According to the analysis of the dynamic changes in rice stems and tillers, rice seedlings reached the peak values at 28 days after transplanting in each density treatment for both carpet seedling and pot seedling, which was marked earlier than in a conventional rice farming system, such as rice-wheat rotation. This phenomenon may be linked to the high basic soil fertility of crayfish paddy fields with long-term feed application. Therefore, farmers occupied in rice-crayfish integrated farming should minimize ineffective tillers in time by draining, rather than estimating the draining time according to the experiment of rice monoculture. The carpet seedling produced tillers occurred rapidly once recovery from transplanting injury, and the number of tillers rapidly exceeded that of the pot seedling. The peak number of seedling at the same density was also higher under carpet seedling than pot seedling. The tillers increased rapidly at the tillering stage by mechanical transplanting of carpet seedling, while higher nodal position tillering accounted for a large proportion of these, resulting in an increase of ineffective tillers, and a high death rate of ineffective tillers after drying the field, leading the "Sharp rise and fall" phenomenon. The tiller number of pot seedling was lower than that of carpet seedling, and the dynamic their stem tillers presented the trend of "slow rise and slow decline" [18]. Therefore, the rice population from mechanically transplanted pot seedling, not only showed high dry matter accumulation, but also controlled ineffective tillers easily, which could avoid an obvious waste of nutrient by non-effective tillers. Previous studies have indicated that mechanically transplanted pot seedling yield can be higher than those of mechanically transplanted carpet seedling [9,19]. Meanwhile, mechanically transplanted pot seedling performed higher photosynthesis potential, crop growth rate and net assimilation rate comparing with mechanically transplanted carpet seedling, which was helpful to provide more sufficient assimilates for obtaining more spikelets per panicle and grain filling. Our results suggested that the highest yield was achieved when the basic seedling density was 77.9 × 10 4 ha −1 . At this seedling density, yield of pot seedling were 4.4% and 5.1% higher than those of carpet seedling in 2018-2019, respectively.
Planting density plays an important role in regulating rice yield. Various academics have performed researches on the effect of density on rice yield, and the research results were controversial to some extent [20][21][22][23]. Most previous studies reported that rice grain yield can be increased by improving the quality of individual plants, thereby, enhancing yield components in panicle, including spikelets per panicle, seed setting rate, or 1000grain weight [15,16]. Huang et al. [24] reported that high densities and nitrogen loss could significantly increase grain yield of machine-transplanted double-cropped rice. However, our results showed that, although the dense planting could make up for part of the yield through the high number of panicles, the contradiction for growth and development between individual plants was aggravated, and single-stem dry matter weight decreased significantly. Meanwhile, the lodging risk caused by thin and weak stems increased, the number of grains per panicle and seed setting rate decreased, and yield was slightly lower than that at a more appropriate density. Under sparse planting, individual development was sufficient, the single-stem dry matter accumulated greater, the number of grains per spike was sufficient, and the seed setting rate was improved, but the ear number was seriously insufficient, causing a waste of temperature and light resources, and it would be difficult to improve rice yield. Therefore, the population deterioration caused by dense planting and the insufficient panicles numbers under sparse planting resulted in different degrees of yield reduction. Only with the appropriate planting density can the yield potential of rice be fully exploited. In terms of the influence of planting density on rice population dynamics, Zhu et al. [23] thought that, by decreasing planting density, the individual growth potential of rice could be fully developed. However, they found that the number of tillers, leaf area index and dry matter weight of the population showed downward trends with decreasing planting density. Similarly, our results showed that the number of tillers, dry matter accumulation and leaf area index decreased with decreasing density. Although the number of tillers per hole was high, each single stem was of high quality, and the stalks were plump under low-density conditions, it is difficult to compensate for the shortage of effective panicles and the decrease in dry matter accumulation of the population caused by the low number of basic seedling. Therefore, the yield was reduced. Therefore, a suitable density is conducive to the formation of reasonable tiller dynamics and improves leaf area index and dry matter accumulation properly, which lays a matter foundation for the construction of high-yield and high-quality rice populations, thereby, improving the yield.

Effect of Mechanical Transplanting Method and Planting Density on Rice Quality
Besides yield, rice quality could also be affected by mechanically transplanting methods. Hu et al. [25] believed that compared with carpet seedling machine transplanting, pot seedling machine transplanting can improve rice milling quality and appearance quality, and reduce amylose content and protein content. Some studies have also suggested that the early sowing or suitable sowing pot seedling machine transplanting [26,27], due to the rice grain filling rate and heading date ahead of time lead to the increase of daily average temperature during grain-filling stage. This results in a higher chalkiness rate and chalkiness than carpet seedling machine transplanting, and reduce the appearance quality of rice. This study showed that rice milling and appearance quality under pot seedling machine transplanting mode were significantly improved compared with the carpet seedling machine transplanting mode. This may be related to the large leaf area index, strong leaf photosynthetic capacity, single stem weight, high dry matter accumulation and transportation rate. In terms of amylose content and protein content, pot seedling machine transplanting is slightly lower or equal to carpet seedling machine transplanting under the same density, which indicated that pot seedling machine transplanting can improve rice cooking and eating quality and nutritional quality to a certain extent, but the effect is not significant. Compared with milling and appearance quality, the improvement degree of pot seedling machine transplanting is not obvious.
Most previous studies [15,16,19] have reported that strong individuals formed under sparse planting owned favorable temperature and light conditions, high single stem plumpness, sufficient grain-filling in the panicles, which significantly improved rice milling and appearance quality. Hu et al. [25] reported that, with a reduction in density, amylose and protein content of rice varieties with small panicle, planted using different mechanical transplanting methods both decreased. Lv et al. [28] suggested that the effect of sparse planting on amylose content is not obvious, but the effect on protein content varies with varieties. Our results showed that, with increasing plant spacing, milling quality and appearance quality of rice were significantly improved, which was consistent with the above results, indicating that appropriately sparse planting played an important role in promoting the commodity value of rice. The main reason was that density reduction promoted the development of individual rice plant, and the storage of photosynthetic products in the stem sheath of sparsely planted rice was higher than that of more densely planted rice at heading stage. Dry matter accumulation during early growth stage can not only enrich the "source" that used by spikelets during grain-filling stage, but also facilitate the formation of smooth "flow", and promoted the development of high-quality photosynthetic organs and the construction of photosynthetic structure. This can meet the high-intensity demand of rice grain-filling, which could improve the grain plumpness and compactness, resulting in that the space between the starch bodies in the endosperm could not form easily, thus, reducing chalkiness, thereby lowering the possibility of grain breakage during rice milling. Rice cooking and eating quality connected closely with amylose and protein content, within a certain range, lower amylose and protein content reflects the softer texture and the higher viscosity, indicating better rice cooking and eating quality [29,30]. The results of our research showed that amylose and protein content decreased with decreasing density, while there were no significant differences among most of the density treatments, indicating that protein and starch accumulation were not very sensitive to transplanting density variation in Nanjing 2728. Therefore, transplanting density may not significantly affect rice cooking and eating quality for Nanjing 2728.

Conclusions
The grain yield and quality of Nanjing 2728 were generally significantly affected by the different mechanical transplanting methods and planting densities under rice-crayfish continuous production system conditions. The optimum planting density for mechanically transplanted carpet seedling was A2 (30 cm × 12.8 cm), and the optimum density for mechanically transplanted pot seedling was B2 ((23 cm + 33 cm) × 13.8 cm). The pot seedling yields were 4.4% and 5.1% higher than those of the carpet seedling due to more spikelets per panicle. In addition, the processing and appearance qualities of the pot seedling rice were also better than those of the carpet seedling rice. Therefore, mechanically transplanted pot seedling with 13.8 cm plant spacing showed good application prospects for rice production in rice-crayfish continuous production systems.