Reducing N Application by Increasing Plant Density Based on Evaluation of Root, Photosynthesis, N Accumulation and Yield of Wheat

: (Aims) To clarify the mechanisms though which dense planting could alleviate the negative effect of the reducing N rate on yield, (Methods) an experiment with four nitrogen levels—0 (N0), 120 (N1), 180 (N2) and 240 (N3) kg N ha − 1 —and three plant densities—180 (D1), 240 (D2) and 300 (D3) × 10 4 basic seedlings ha − 1 —was conducted. (Results) Increasing plant density decreased the root length, root volume, root surface area and root tips of individual plant while it enhanced the aforementioned root traits in population. The chlorophyll content, photosynthetic rate, stomatal conductance and transpiration rate of the individual plants were decreased with the increase in plant density and enhanced with the increase in N level. The increasing density and N application rate enhanced the leaf area index, photosynthetic high-efﬁciency leaf area and canopy photosynthetically active radiation of population. N accumulation per plant was decreased with increasing density and was enhanced with an increasing N application level. Within the same N level, the N accumulation in the population, N production efﬁciency and N recovery efﬁciency were consistently D3 > D2 > D1. A high N application rate with high density was not conducive to improving the NR (nitrate reductase), GS (glutamine synthetase) and GOGAT (glutamate synthase) activities. The yield could be maintained as stable or improved if decreasing by 60 kg N ha − 1 with increasing 60 × 10 4 basic seedlings ha − 1 within the range of N application in this experiment. (Conclusions) These results indicated that the yield of wheat could be improved with less N application by adjusting the compensatory effects from the plant density in populations.


Introduction
Wheat is the most important crop for human nutrition and caters to one-fifth of the calorie requirement of the global population, and is a major staple food in China and in other temperate regions worldwide. Plant density and nitrogen application are two critical agronomic practices in the wheat cropping system for high yield [1]. In traditional wheat production, low density with a high nitrogen application rate was widely regarded as a good cropping mode for a high wheat yield in China, due to many farmers believing that the reduced tillers from the low plant density could be compensated for by applying more N fertilizer. Since excessive quantities of N fertilizers were applied to agricultural crops, more and more agricultural researchers have discovered that high N input not only increased the production cost, but also resulted in severe environmental pollution, low nitrogen use efficiency and a limitation on crop yield and quality improvement [2]. However, reducing the N application rate would increase the risk of grain yield loss due to the reductions in both the number of spikes and grains [3], and evidence has shown that increasing plant density could obviously enhance the wheat population, number of spikes and grains for high yield [4,5]. Thus, we hypothesized that increasing plant density with a reduction in the N application rate might benefit wheat production for high yield with high N use efficiency.
The root traits are critically important for the acquisition of water and soil nutrients, and some studies have shown that root growth and the root system architecture were obviously affected by plant density and N application rate [2,6]. Therefore, a comprehensive understanding of the relationship between root traits in addition to plant density and N application rate may improve crop growth, N efficiency and increase grain yield. Increasing plant density and the N application rate could also obviously affect the development of plants and the total green leaf area, chlorophyll content, photosynthetic rate as well as the interception of canopy photosynthetically active radiation [7][8][9], which is an important premise for the formation of dry matter and yield. Thus, how to effectively improve crop photosynthesis and increase the green leaf area by optimizing plant density and N application rate is a major step toward improving crop yield. N accumulation, N efficiency and N assimilatory enzymes' activities not only have a significant relationship with crop growth and yield but are also closely related to N application rate and plant density, and some studies have suggested that N fertilizer can be used at a moderately lower rate than they have been traditionally used by improving N efficiency and the capacity of N uptake and accumulation [10,11]. N accumulation in the population and N efficiency could be enhanced by increasing appropriately plant density [12]. Therefore, in order to achieve a high yield with a low N application rate, it is necessary to have a better understanding of the N accumulation, N efficiency and N assimilatory enzymes' activities in response to different plant density and N application rate.
It is well known that reducing the amount of N application will usually decrease crop yield to a certain extent, however, the excessive application of N fertilizer will lead to a reduction in N use efficiency and a series of environmental problems, and even cause a decline in crop yield [7,10]. How to reduce the N application rate without yield reduction is extremely important for modern agricultural production and has become a new research arena. Some studies have shown that increasing plant density could alleviate the negative effect of reducing the N application rate on yield by enhancing the panicles per m 2 or spikelets per m 2 , and also predicted that dense planting with a reducing N application rate might produce not only a higher or similar amount of yield compared with sparse planting with a high N rate, but also improve N use efficiency [10,13,14]. In view of the fact that the specific compensation mechanism of increasing density to decreasing N application is not well documented, a three-year experiment was conducted to (i) detect the combined impacts of plant density and N application rate on root traits, photosynthetic indexes, N accumulation, N efficiency, N assimilatory enzymes activities and yield of wheat; (ii) clarify the mechanisms (from wheat population and individuals) why increasing plant density could alleviate the negative effect of reducing the N rate on yield; (iii) determine the optimum plant density that can compensate for the loss caused by reduced nitrogen rate.

Experimental Design and Management
The experiment was carried out in Baihu farm, Lujiang County, Hefei City (31 • 53 N, 117 • 14 E; 29.8 m a.s.l.), Anhui Province in China from 2017 to 2020 (only two-year test data was selected for analysis). The region is classified as having a subtropical monsoon climate. The annual mean temperature is 16.4 • C and accumulated temperatures above 10 • C were 4800-5400 • C. Annual mean precipitation is 1050-1250 mm, and 50% occurs from April to August. The frost-free period is 255-270 days each year. The basic chemical and physical properties of 0~20 cm soil in the initial stage of experiment were total N 1.36 g·kg −1 , total P 0.74 g·kg −1 , available N 96.48 mg·kg −1 , available P 8.12 mg·kg −1 , organic matter 17.13 g·kg −1 , and pH 5.8.
The experiment was carried out in a randomized block design with four N application levels (N0, N1, N2, and N3) and three plant densities (D1, D2 and D3) as the treatment variables. This experimental plan generated 12 treatments and each treatment was replicated three times. The four N levels were N0 (0 kg ha −1 ), N1 (120 kg ha −1 ), N2 (180 kg ha −1 ), and N3 (240 kg ha −1 ); and the three plant densities were D1 (180 × 10 4 basic seedlings ha −1 ), D2 (240 × 10 4 basic seedlings ha −1 ) and D3 (300 × 10 4 basic seedlings ha −1 ), and the topdressing N was applied at the jointing stage of wheat. All plots were given a basal application of 120 kg P ha −1 and 90 kg K ha −1 . N was applied as urea (46.4% N), and p and K were applied as calcium superphosphate (12% P 2 O 5 ) and potassium chloride (60% K 2 O), respectively. A local wheat variety of 'Ningmai 13' was selected and sown on 6 November with row spacing of 20 cm, which was harvested on 25 May (the next year). Each experimental plot was 12 m 2 (3 m × 4 m) with 50 cm row spacing between neighboring plots.

Sampling and Measurements
Root traits: the root samples were carefully taken from a single plant in each treatment group at the flowering stage, the sampling depth was 0~50 cm, the root samples were kept in an icebox and transported to the laboratory within 3 h of collection, and the root system of the target plant was separated from other roots by carefully extracting the roots under running water on a sieve (mesh size 0.2 mm). The separated samples were then scanned with an Expression 10,000 XL 1.0 scanner (dpi = 400; Epson Telford, Ltd., Telford, UK), and the images were analyzed using the WinRhizo software (Instruments Regent Co., Ville de Québec, QC, Canada) to determine the root length (RL), root surface area (RSA), root volume (RV) and root tips (RT). Five root samples of wheat were taken from each treatment for measurement (Table 1). Chlorophyll content was measured by using a hand-held chlorophyll meter (SPAD-502, manufactured by the Konica Minolta Company, Tokyo, Japan, measuring area: 2 mm × 3 mm), the same parts of flag leaves of the wheat plants (ten flag leaves in each treatment) were selected and measured at jointing, flowering and the middle of filling stages in 2019 and 2020, respectively. Leaf area index: leaf area was measured by passing the leaves through a LI-3100C leaf area meter (USA), and then the leaf area index (LAI) was calculated as a leaf area per unit land area. Ten wheat plants were sampled from each treatment at jointing, flowering and middle of filling stages in 2019 and 2020, respectively.
Photosynthetic characteristics: leaf photosynthetic rate, stomatal conductance and transpiration rate were measured with a portable photosynthesis system (LI-6400, Li-Cor, Lincoln, NE, USA) at 9:00-11:30 h local time in middle of the filling stage in 2019 and 2020, respectively. Five flag leaves of wheat in each treatment were selected for the leaf measurements.
The photosynthetic high-efficiency leaf area of population was calculated using the sum of the top three leaves' area of the plant population. The area of the flag leaf, the second and the third leaves from the top of wheat plant (ten wheat plants were selected from each treatment) were measured by a portable leaf area meter (Model Li-3000C, Lincoln, NE, USA) at middle of the filling stage in 2019 and 2020, respectively.
Canopy photosynthetically active radiation (PAR): photosynthetically active radiation was measured by SUNSCAN Canopy Analysis System (Delta company, Cambridge, UK) at the flowering and middle of the filling stages of wheat in 2019 and 2020, respectively. The PAR of the canopy was calculated by the difference value between the PAR of the top (approximately 1.5 m above ground level) and bottom (the transmitted PAR measured at the base of the canopy) of wheat canopy, and the measurement time was 9:00-11:30 in the morning and 13:00-16:00 in the afternoon, respectively. The spot measurements were limited to clear days to avoid the poor quality of incident PAR (e.g., multiple sources of PAR due to cloud refraction and reflection) that influence the light interception measurements. Three measurements were taken for the 'top' and 'bottom' canopy position at locations selected randomly in each treatment in rapid succession.
N accumulation and efficiency: for the determination of total N uptake, the dried grain and straw samples of each plant part (five wheat plants were selected from each treatment) were milled with a Wiley mill and screened through a 0.5 mm sieve. The N contents of the grain and straw samples were determined by Continuous Flow Analysis (AA3, Seal Analytical Inc., Southampton, UK). N accumulation per plant (mg/per plant) = N content of grain and straw × dry matter weight per plant; N production efficiency (NPE, kg/kg) = grain yield/N application rate; N recovery efficiency (NRE, %) = (total N accumulation of the population with N fertilization-total N accumulation of the population without N fertilization) /N application rate.
Nitrogen assimilatory enzymes activities: the GS (glutamine synthetase) activity, NR (nitrate reductase) activity determination method was determined by the method of by Lang et al. [15]; and used the method of Li et al. [16] for the determination of GOGAT (glutamate synthase) activity. Green leaves of five wheat plants were selected from each treatment at flowering stage for determination.
Yield: wheat grains were harvested by the plot harvester at the maturity stage (each treatment was repeated three times), and the grain yield was calculated after sun drying (the average moisture content of grain was 12.5%).

Statistical Analysis
ANOVA was performed by the general linear model-univariate (mixed model, N level and plant density were considered as fixed factor and random factor, respectively) procedure from SPSS 20.0 software (IBM, Armonk, New York, NY, USA). ANOVAs were performed with the N level and plant density as the main effects and including their interactions. All treatment means were compared for any significant differences by the LSD's multiple range tests at the significant level of p = 0.05 with the SPSS 21.0 software package for Windows.

The Interaction of Increasing Density and Decreasing N on Root Traits of Wheat
The root length (RL), root volume (RV), root surface area (RSA) and root tips (RT) per plant of wheat decreased with the increase in planting density (Table 2), and the RL, RV, RSA and RT per plant of D1 were significantly higher than that of D3 under the same N level. Within the same N level (N0, N1, N2 or N3), the RL, RV, RSA and RT of population were significantly enhanced by increasing planting density. Within the same density, the RL, RV, RSA and RT of individual plant and population were consistently N3 > N2 > N1 > N0, and the differences between N1 and N0, and N3 and N1 were significant. The RL, RV, RSA and RT of population in N1D3 were significantly higher than that in N2D1, and in N2D3 were significantly higher than that in N3D1, indicating that the RL, RV, RSA and RT of wheat population could be still significantly increased when decreasing 60 kg N ha −1 with increasing 120 × 10 4 basic seedlings ha −1 .

The Interaction of Increasing Density and Decreasing N on Chlorophyll Content
The chlorophyll content in the treatments with N application was significantly higher than that in N0 (Table 3). Within the same N level, the chlorophyll content of wheat at each growth stage decreased with increasing density. Within the same density (D1, D2 or D3), the chlorophyll content at the same growth stage was consistently N3 > N2 > N1 > N0. Compared to N0, N3 for D1 significantly increased the chlorophyll content at the jointing, flowering and middle filling stages by 29.29%, 27.97%, 20.01% and 24.78%, 20.12%, 27.15% in 2019 and 2020, respectively. The results showed that the chlorophyll content of wheat was significantly enhanced by increasing 120 kg N ha −1 under the same density (within the range of N application in this experiment). The differences in chlorophyll content at each growth stage between N1D1 and N2D3, N2D1 and N3D3 were insignificant, indicating that a high N application rate with high density could not obviously improve the chlorophyll content of wheat as compared to a low N application rate with low density.

The Interaction of Increasing Density and Decreasing N on Leaf Area Index
The leaf area index (LAI) increased with the increase in density at each growth stage (Table 4), under the same N application rate, and the LAI in D3 was significantly higher than that in D1, and in D3, it was higher than that in D2. Within the same density (D1, D2 or D3), the LAI in N1 was significantly higher than that in N0, and in N2, it was significantly higher than that in N1, and in N3, it was significantly higher than that in N2 (except between N2D3 and N3D3 at middle of filling stage in 2019), indicating that the LAI of the population could be significantly improved by the increasing N application rate from 120 kg ha −1 to 180 kg ha −1 or from 180 kg ha −1 to 240 kg ha −1 . The LAI in N1D3 was higher than that in N2D1, and in N2D3 was higher than that in N3D1, indicating that the LAI could be still enhanced when decreasing 60 kg N ha −1 with increasing 120 × 10 4 basic seedlings ha −1 .

The Interaction of Increasing Density and Decreasing N on Photosynthetic Characteristics and Photosynthetic High-Efficiency Leaf Area
The photosynthetic rate (Table 5), stomatal conductance and transpiration rate in N1, N2, and N3 were significantly higher than that in N0. The increasing density decreased the photosynthetic rate, stomatal conductance, and transpiration rate, but enhanced the photosynthetic high-efficiency leaf area (PHLA) of the population. Compared to D1, D3 for N2 decreased the photosynthetic rate, stomatal conductance, transpiration rate by 4.43%, 7.32%, 9.41% and 9.64%, 9.96%, 8.10% in 2019 and 2020, respectively. Under the same N level, the PHLA of the population in D2 was significantly higher than that in D1, and in D3 was significantly higher than that in D2, indicating that increasing plant density with 60 × 10 4 basic seedlings ha −1 could significantly enhance the PHLA of the population. Within the same density (D1, D2 or D3), the photosynthetic rate, stomatal conductance, transpiration rate and PHLA of the population in N1, N2 and N3 were significantly higher than that in N0, and in N3 were significantly higher than that in N1. In addition, the PHLA of population in N0D3 was significantly higher than that in N1D1, and in N1D3 was significantly higher than that in N2D1, and in N2D3 was significantly higher than that in N3D1, indicating that the PHLA of population could be still enhanced when decreasing 60 kg N ha −1 with increasing 120 × 10 4 basic seedlings ha −1 (within the range of N application in this experiment). Table 5. The interaction of increasing density and decreasing N on photosynthetic characteristics and photosynthetic high-efficiency leaf area of population.

The Interaction of Increasing Density and Decreasing N on Canopy Photosynthetically Active Radiation
The canopy photosynthetically active radiation (PAR) in N application treatments were significantly higher than that in N0 (Table 6). Increasing density enhanced the canopy PAR of wheat, within the same N level (N1, N2 or N3), the canopy PAR in D2 was significantly higher than that in D1, and in D3 was significantly higher than that in D2. Within the same density, the canopy PAR in N2 was significantly higher than that in N1 (except between N1D1 and N2D1 in 2020), and in N3, it was significantly higher than that in N2, indicating that increasing or decreasing 60 kg N ha −1 had a significant effect on canopy PAR within the range of N application in this experiment. The canopy PAR in N1D3 was significantly higher than that in N2D1, and in N2D3, it was significantly higher than N3D1; when compared to N3D1, N2D3 significantly increased canopy PAR at the flowering and middle of the filling stages by 9.63% and 10.95% in 2019, and by 9.96% and 6.41% in 2020, respectively. The results indicated that the canopy PAR was significantly enhanced when decreasing 60 kg N ha −1 with increasing 120 × 10 4 basic seedlings ha −1 within the range of N application in this experiment.

The Interaction of Increasing Density and Decreasing N on N Accumulation and N Efficiency
Within the same N level (Table 7), increasing density decreased N accumulation per plant but enhanced the N accumulation of the population, N production efficiency and N recovery efficiency. When compared to D1, D3 for N0, N1, N2, N3 significantly decreased N accumulation per plant by 15.75%, 15.43%, 8.97%, 6.62% and 12.85%, 9.63%, 7.80%, 7.55% in 2019 and 2020, respectively. Within the same N level, the N accumulation of the population, N production efficiency and N recovery efficiency were consistently D3 > D2 > D1, and the differences between D3 and D1 were significant. Increasing the N application enhanced the N accumulation of individual plants and the population, while decreasing N production efficiency and N recovery efficiency. Within the same density, the N accumulation of individual plants and population in N2 were significantly higher than that in N1, and in N3, were significantly higher than that in N2, while the N production efficiency in N2 was significantly lower than that in N1, and in N3, it was significantly lower than that in N2. In addition, the N accumulation of the population in N1D2 was significantly higher than that in N2D1, and in N2D2, it was significantly higher than that in N3D1, indicating that the N accumulation of the population was significantly enhanced when decreasing 60 kg N ha −1 with increasing 60 × 10 4 basic seedlings ha −1 within the range of N application in this experiment.

The Interaction of Increasing Density and Decreasing N on Nitrogen Assimilatory Enzymes Activities
The activities of NR, GS and GOGAT in the treatments with N application were significantly higher than that in N0 (Table 8). Within the same N level (N0, N1, N2 or N3), the NR, GS and GOGAT activities of leaf decreased with increasing density, and sometimes the differences in the activities of NR, GS and GOGAT between D3 and D1 were significant. Within the same density (D1, D2 or D3), the NR, GS and GOGAT activities at flowering stage were consistently N3 > N2 > N1 > N0, and the differences in the activities of NR, GS and GOGAT between N1 and N0, and N3 and N1 were significant. The results showed that the NR, GS and GOGAT activities of leaf were significantly enhanced by increasing 120 kg N ha −1 under the same density (within the range of N application in this experiment). The activities of NR, GS and GOGAT in N2D1 were higher than that in N3D3, indicating that a high N application rate with high density could not always improve the NR, GS and GOGAT activities as compared to a low N application rate with low density.

The Interaction of Increasing Density and Decreasing N on Yield
Within the same N level (Figure 1), increasing density could help enhance yield, compared to D1, as D3 for N1 and N3 significantly increased yield in 2019 and 2020 by 15.77%, 16.03% and 13.09%, 11.44%, respectively, indicating that increasing plant density with 120 × 10 4 basic seedlings ha −1 could significantly enhance yield under the same N level. Within the same density, the yield in N1 was significantly higher than that in N0, and in N2 was significantly higher than that in N1, and in N3 was significantly higher than that in N2, indicating that increasing 60 kg N ha −1 could significantly enhance yield (within the range of N application in this experiment). In addition, the difference in yields in 2019 and 2020 between N1D3 and N2D1 were insignificant, and the yields in 2019 and 2020 of N2D3 were slightly higher than that of N3D1, indicating that the yield could be still maintained as stable or improved if decreasing 60 kg N ha −1 with increasing 60 × 10 4 basic seedlings ha −1 under this experimental condition.

The Interaction of Increasing Density and Decreasing N on Yield
Within the same N level (Figure 1), increasing density could help enhance yield, compared to D1, as D3 for N1 and N3 significantly increased yield in 2019 and 2020 by 15.77%, 16.03% and 13.09%, 11.44%, respectively, indicating that increasing plant density with 120 × 10 4 basic seedlings ha −1 could significantly enhance yield under the same N level. Within the same density, the yield in N1 was significantly higher than that in N0, and in N2 was significantly higher than that in N1, and in N3 was significantly higher than that in N2, indicating that increasing 60 kg N ha −1 could significantly enhance yield (within the range of N application in this experiment). In addition, the difference in yields in 2019 and 2020 between N1D3 and N2D1 were insignificant, and the yields in 2019 and 2020 of N2D3 were slightly higher than that of N3D1, indicating that the yield could be still maintained as stable or improved if decreasing 60 kg N ha −1 with increasing 60 × 10 4 basic seedlings ha −1 under this experimental condition.

Discussion
The changes of N application rate and plant density can lead to the differences in root parameters which were closely related to crop growth and yield. Liu et al. [17] indicated that the root growth of super hybrid rice cultivars tended to decrease when subjected to high N fertilization treatment N4 (390 kg ha −1 ), while N3 (300 kg ha −1 ) had greater root traits when compared with N4, N2 (210 kg ha −1 ) and N1 (0 kg ha −1 ). Jia et al. [

Discussion
The changes of N application rate and plant density can lead to the differences in root parameters which were closely related to crop growth and yield. Liu et al. [17] indicated that the root growth of super hybrid rice cultivars tended to decrease when subjected to high N fertilization treatment N4 (390 kg ha −1 ), while N3 (300 kg ha −1 ) had greater root traits when compared with N4, N2 (210 kg ha −1 ) and N1 (0 kg ha −1 ). Jia et al. [18] found that the root surface area density and average root length density of maize in the 0-100 cm soil layers increased under high and medium densities compared with low density, while the root weight under high density significantly decreased compared with that under low density. Dai et al. [12] reported that increasing wheat density from 135 to 405 m −2 significantly enhanced root length density and total root numbers per unit area at 0.2, 0.6 and 1.0 m soil depth. In this study, we found that increasing the N application rate enhanced the RL, RV, RSA and RT of individual plants and the population together, though we did not find the inhibitory effect of nitrogen fertilizer on root traits, which was different from the research result of Liu et al. [17] to a certain extent, this might be due to our nitrogen application rate was not too high. Within the same N level, increasing plant density decreased the RL, RV, RSA and RT of individual plants while it enhanced the RL, RV, RSA and RT of population, and our findings were consistent with those of Dai et al. [12] and Chen et al. [19]. In addition, we also found that the RL, RV, RSA and RT of the wheat population could be still significantly increased when decreasing 60 kg N ha −1 with increasing planting density 120 × 10 4 basic seedlings ha −1 . Adebayo et al. [20] also reported that the root system architecture traits of a crop were obviously affected by the increasing or decreasing N application rates and plant densities. The main reason why decreasing 60 kg N ha −1 with increasing density 120 × 10 4 basic seedlings ha −1 could significantly increase the RL, RV, RSA and RT of the wheat population was due to (i) the increase in plant density enhanced the total root population (as the results of Table 2), and the negative effects of the reducing N rate on root traits could be obviously compensated by increasing the planting density; and (ii) the increase in total root population enhanced the total absorption and utilization amount of soil N which could effectively compensate for the decrease in nitrogen application [10,12].
The photosynthetic capacity of a crop has a close positive correlation with the N application rate and plant density, as the optimization of the planting density and nitrogen level is an important step toward improving the radiation interception, radiation use efficiency, and photosynthetic rate of the plant [21]. Fang et al. [8] reported that the net photosynthetic rate, stomatal conductance, transpiration rate, chlorophyll content and leaf area index of wheat increased and then decreased with the increase in planting density and nitrogen fertilizer, and their maximum values appeared at the plant density of 90 plants m −2 and N fertilization of 90 kg ha −1 treatment. In this study, we observed that the chlorophyll content, photosynthetic rate, stomatal conductance and transpiration rate of the individual plant were decreased with the increase in plant density, and increased with the increase in N level; the LAI, PHLA and canopy PAR of population were increased with the increase in N level and plant density, and the similar results were also observed in previous studies of Hou et al. [7] and Yao et al. [21]. We did not find the inhibitory effect of density and nitrogen fertilizer on the aforementioned photosynthetic indexes, which might be due to the fact that our nitrogen application rate and density were not high enough. In addition, we also found that the LAI, PHLA and canopy PAR of the population could obviously still be enhanced when decreasing 60 kg N ha −1 with increasing 120 × 10 4 basic seedlings ha −1 within the range of N application in this experiment. Some studies also proved that different plant densities and N application rates had an obvious effect on adjusting leaf photosynthetic capacity and the canopy photosynthetic use efficiency of light [22][23][24]. Increasing plant density could compensate for the decrease in the photosynthetic efficiency caused by reducing N application, which was mainly because that plant density could significantly influence the light environment in an individual plant and the population [9,21,25], as increasing density obviously enhances the LAI, PHLA, and canopy PAR of population.
In order to promote early growth and yield for wheat, excessive N fertilizers are usually applied; however, high N input might lead to low N use efficiency resulting from N loss. Xie et al. [10] reported that dense planting with reducing N rate significantly improved N use efficiency for grain production, N recovery rate and N retention rate compared with sparse planting with a high N rate. Tian et al. [14] found that the highest N accumulations in rice at the high (32.5 × 10 4 hills ha −1 ) and low (25.5 × 10 4 hills ha −1 ) densities were 205 and 207 kg ha −1 under the 180 and 270 kg ha −1 N levels, respectively, indicating that dense planting with low N rate can obtain N accumulation equivalent to sparse planting with a high N rate. In this study, we found that increasing density decreased the activities of NR, GS and GOGAT but enhanced the N accumulation of population, N production efficiency and N recovery efficiency, and the research results of us were consistent with the previous findings of Chen et al. [19] and Xu et al. [26]; within the same density, increasing the N application level enhanced nitrogen assimilatory enzymes activities and the N accumulation of the population while decreased N production efficiency and N recovery efficiency, and similar conclusions were also drawn by Mrid et al. [27] and Hou et al. [7]. In addition, we also proved that the N accumulation of the population could still be significantly enhanced when decreasing 60 kg N ha −1 with increasing 60 × 10 4 basic seedlings ha −1 , and a high N application rate with high density did not help to improve the NR, GS and GOGAT activities. Luo et al. [28] also found that the N rate could be reduced to 20-30% from the traditionally recommended rate without sacrificing yield under high plant density by enhancing N efficiency. Increasing plant density could enhance N production efficiency and N recovery efficiency, and compensate for the decrease in the N accumulation of a population caused by reducing the N application rate, which was mainly because increasing plant density could (i) obviously enhance N uptake and the accumulation capacity of crop population, although reducing the N application rate decreased the N accumulation of an individual plant, the increase in the number of wheat population could significantly improve the N accumulation capacity of the population [8,29]; (ii) enhance the total amount of roots per unit soil area (the ability of crop N uptake was significantly positively correlated with plant root system); and (iii) enhance the N uptake from deep soil, the increase in planting density enhanced root length density, root biomass per unit soil area and aboveground N uptake; therefore, it could help to efficiently recover N leached to deep soil [12].
Recently, dense planting with less N application rate is widely recommended as an appropriate approach to achieve high yield. Zhu et al. [1] observed that an increase in planting density by about 50% with correspondingly reduction in basal N rate by about 30% enhanced rice yield by 0.5-7.4% over the conventional sparse planting. Hou et al. [7] reported that it is difficult to improve crop yield when a high N level combined with a high plant density or a low plant density, because of the lower number of panicles under a low plant density, or too many unproductive panicles under a high plant density due to fierce intraspecific competition. In this study, we found that increasing plant density could significantly enhance wheat yield under the same N level, and a similar conclusion was also drawn by Liu et al. [30]. The reason that we did not find the inhibitory effect of density and nitrogen fertilizer on yield might be due to our nitrogen application rate and density not being too high. In addition, the two-year results also showed that the yield could still be still maintained as stable or improved if decreasing 60 kg N ha −1 with increasing 60 × 10 4 basic seedlings ha −1 . Dense planting with reduced N rate could produce a higher or approximate yield compared with sparse planting with a high N rate, which was mainly because increasing plant density could (1) enhance the total amount of roots of crop population, and then the crop population could absorb and utilize more kinds and quantities of soil nutrients [31]; (2) improve the photosynthetic capacity of the population, and the improvement of photosynthetic capacity of the crop population has always been considered as a very important way to increase yield [5,21]; (3) enhance N use efficiency and the N accumulation of the population, therefore the limitation of reducing the N application rate on yield increase can be effectively avoided [32]; (4) increase the spike number of population in per unit area, and the sufficient spike number is one of the essential factors to ensure high and stable yield [14], in addition, some new cultivars having improved tolerance to plant density stress is also one of important reasons [33].

Conclusions
N application rate and plant density significantly affected the development of the wheat root traits, photosynthesis, N assimilatory enzymes activities and N accumulation which in turn influenced the nutrient absorption capacity, photosynthetic capacity, N efficiency and yield of wheat. Although increasing density or reducing N application decreased the RL, RV, RSA, RT, chlorophyll content, photosynthetic rate and N accumulation of the individual plant, the increase in plant density could obviously enhance the RL, RV, RSA, RT, LAI, PHLA, canopy PAR, N accumulation, N production efficiency, N recovery efficiency and spike number of the population. The negative effect of reducing N rate on wheat could be offset by improving plant density, such as decreasing 60 kg N ha −1 with increasing 120 × 10 4 basic seedlings ha −1 which obviously enhanced the RL, RV, RSA, RT, LAI, PHLA and canopy PAR in population, and decreasing 60 kg N ha −1 with increasing 60 × 10 4 basic seedlings ha −1 not only significantly enhanced the N accumulation of population but also kept yield stable or improved. Therefore, we suggested that moderate dense planting with less N application might be an environmentally friendly mode for wheat cropping for high yield and N use efficiency in actual agricultural production.