Effect of nitrogen fertilizer on seed yield and quality of Kengyilia melanthera (Triticeae, Poaceae)

View article
Plant Biology

Introduction

The Qinghai-Tibet Plateau (QTP), the so-called third pole of the world, is an important eco-region. The grassland ecosystem in the QTP is undergoing extensive and longtime deterioration and even desertification, due to both human and natural factors (Lu et al., 2017; Wu, Yu & Jin, 2017). Since the beginning of the 21st century, the Chinese government has implemented large-scale ecological restoration projects to conserve grasslands, mitigate degradation, and combat desertification (Ma, Diao & Deng, 2019). However, the return of a desertified grassland to a more natural state by incorporation of phytorestoration technology has been limited chiefly due to the lack of suitable grass species and sufficient seed supply (Yu et al., 2019; Atinkut et al., 2020). Notably, Kengyilia melanthera, a perennial grass species, is widely distributed the alpine sandy grassland in east QTP, lives in habitats at an altitude of 3,300–4,800 m (Yang & Yan, 1992). It is an important ecological barrier against desertification because of its strong adaptation to arid/semiarid climates and barren sandy soil (Xiao et al., 2008). To date, a unique authorized Kengyilia variety in China bred by Sichuan Academy of Grassland Science, namely K. melanthera ‘Aba’, has been approved by the National Herb Variety Approval Committee of China in 2009 (Xiao et al., 2011).

K. melanthera is one of the few grass species that can produce seed on a large scale in east QTP. Nevertheless, the high exigency of seed utilization for ecological recovery of pasture builds demand for the more sufficient and high-quality seeds of K. melanthera. This results mainly from low productivity of native seed fields and the subsequent low efficiency of ecological restoration. Heredity and external conditions determine seed yield, and fertilization is one of the most effective ways to improve seed yield in extrinsic conditions. Usually, nitrogen (N) fertilizers are considered more effective for increasing crop seed yield than phosphate or potassium fertilizers (Amanullah et al., 2016; Li et al., 2018a). In QTP, animal manure was often collected for production of commercial N fertilizer while its N matter were relatively low. Therefore, addition of N fertilizer and screening out optimal dose to ensure the growth and development of plants is imperative (Zhang et al., 2016; Quang et al., 2004). Wan, Wu & Liao (2021)’s study on winter wheat with amounts of N fertilizer application of 0–360 kg·hm−2 demonstrated that the optimal N application rate (240 kg·hm−2) could not only increase the number of tillers, seed number per spike, and seed yield, but also could accelerate their growth process. Xie et al. (2010)’s study that examined the impact of amounts of N fertilizer application (0–330 kg·hm−2) on wild Elymus nutans in Tibet found that the number of tillers, number of fertile tillers, and seed yield reached the peak value under 250 kg·hm−2 and was significantly different from the other treatments. The genetic improvement and breeding of K. melanthera, as yet, mainly relies on mixed selection or recurrent selection of local wild germplasm. This led to a relatively low genetic gain especially on seed traits. As a result, it is an economical and practical option to increase the seed production potential of the of K. melanthera through cultivating measures such as nitrogen application. Moreover, we also found that harvesting seeds of K. melanthera at approximately 30 days after peak anthesis is considered as an optimal time point, which also has the lowest seed shattering occurrence.

The high demand of K. melanthera seeds for the recovery of desertified grassland in the QTP has been a long time, but not yet founded effective localized technology scheme. In order to exploit high-efficiency and inexpensive technical measures for seed production of K. melanthera ‘Aba’, a 2-year investigation was carried out to assess the impact of N fertilization on the seed yield and quality in the northwest Sichuan plateau, located in east of QTP, China.

Materials and Methods

Experimental sites

The experiment was conducted for two consecutive years, between 2015 and 2017. The field is located in Hongyuan county, Aba Autonomous Prefecture, Sichuan province, a semi-humid region of the eastern QTP of China (31°79′35″N, 102°55′47″E, 3,480 m a.s.l.). Before sowing, the grasses and weeds were eliminated in all plots by hand weeding and herbicide spray, then the field was ploughed twice. The soil of experimental land was a subalpine meadow soil in nature, and chemical properties were provided in Table S1. In the 0–30 cm soil layer (pH 6.02), the concentration of organic matter, total nitrogen, total phosphorus, total potassium, available nitrogen, phosphorus, and potassium in the soil was 13.8, 1.32, 0.86, 0.68, 87.47, 62.55, and 227.46 mg·kg−1, respectively.

Experimental management

Seeds of K. melanthera (variety named ‘Aba’) was supplied by the Sichuan Academy of Grassland Sciences, and was harvested in the most recent production year before the field plot experiment. Its 1,000-seed weight and germination rate were 4.97 g and 80%, respectively. Generally, during the sowing year of (2015), the vast majority of K. melanthera plants growth, development, flowering, and seed setting are markedly reduced by the continental plateau’s cold temperate monsoon climate. Hence, the seed yield and quality data in this study were recorded beginning from the second year (2016).

Experimental design

The fertilizer treatments (urea fertilizer, 46% N) were applied to the experimental plots during the elongation stage, in mid June of 2016 and 2017. This trial is adopted as was a randomized complete block design for eight sample plots with different amounts of nitrogen fertilization (N1 – 0 kg·hm−2, N2 – 60 kg·hm−2, N3 – 90 kg·hm−2, N4 – 120 kg·hm−2, N5 – 150 kg·hm−2, N6 – 180 kg·hm−2, N7 – 210 kg·hm−2, N8 – 240 kg·hm−2). Each N fertilizer treatment had four replications’ plots and a total of 32 plots. Sowing was performed by the drill method on May 20, 2015, with a seed density of 22.5 kg·hm−2, sowing depth of 1–2 cm, and 50-cm row spacing. Experimental units consisted of plots with an area of 20 m2 (5 m × 4 m) and 8 rows, and seeds were not sown within 50 cm of the boundary of the plot in order to reduce the marginal effect.

Evaluations

Seed yield components

The number of tillers (NTs), number of fertile tillers (NFTs), number of spikes (NSPs), number of florets (NFLs), number of fertile florets (NFFLs), and 1,000-seed weight (TSW) were measured to evaluate the seed yield components. At full-blooming stage (the proportion of plants that flowered reached 50%), a row was randomly selected from each sample plot for measurement of the NTs and NFTs, and four replicate measurement rows were performed within each plot. Here the length of selected row is 1 m and not on border rows of plot. In addition, the different indices cannot be sampled from the same plant. The NSPs, NFLs, and NFFLs were counted based on ten randomly sampled tillers in full-blooming stage from each plot. The seeds from each plot were harvested from the 30th day after peak anthesis. 1,000 sun-dried seeds were randomly selected to determine the TSW, four replications were carried out.

Seed yield

The seed yield indices could be represented by the harvested seed yield (HSY), potential seed yield (PSY), and presentation seed yield (PRSY). At the 30th day after peak anthesis, for each measurement, four row-segments with the length of 1 m were randomly selected from each plot, and all its seeds were harvested. After these seeds were threshed, cleaned, and naturally sun-dried to a seed moisture content of about 150 g·kg−1, and the HSY of each N treatment could be obtained. The PSY and PRSY per unit area were calculated with the following formulas:

PSY=NFTs/m2×NSPs/NFTs×NFLs/NSPs×Averageseedweight

PRSY=NFTs/m2×NSPs/NFTs×NFFLs/NSPs×Averageseedweight

Seed quality

To reduce or avoid dormancy and dormancy and after-ripening of freshly harvested seeds, seed quality trait analysis was tested approximately two months following harvest. The seed quality indices were based on standard germination rate (SGR), germination energy (GE), accelerated aging germination rate (AAGR), dehydrogenase (DE) and acid phosphoesterase (APH) activity. Because some weed seeds might be mistakenly harvested when K. melanthera seeds were harvested, so these seeds should be removed, the remaining seeds are called pure seeds. We randomly selected 50 pure seeds under different N treatments respectively and placed them into a Petri dish with two layers of filter paper on the bottom, adding a suitable amount of distilled water (preferably submerging seeds), and then moved them to cultivate in a plant growth chamber at a day temperature of 25 °C (16 h) and night temperature of 16 °C (8 h). Finally, we determined the SGR and GE of seeds on the 4th and 7th days of seed germination, respectively. For the seed-aging treatment, 250 pure seeds were randomly selected and embedded into an aging tank kept at 41 °C (72 h). The AAGR was measured after aging treatment, measurements from SGR and this method were identical. The DE and APH activity of seeds were determined by the triphenyl tetrazolium chloride (TTC) (Mao et al., 2001) and the sodium p-nitrophenol phosphate method (Stephen, Daniel & William, 1989), respectively. The assays described above were repeated four times.

Statistics analysis

The statistical analysis was conducted by one-way and two-way analysis of variance (ANOVA) using IBM SPSS Statistics 20.0 software. The two-way analysis of ANOVA with a general linear model at P < 0.05 was applied to determine the combined influence of N fertilizer and trial years on seed yield components, seed yield, and quality. Bonferroni adjustment was applied after multiple hypothesis testing at the P < 0.05 probability level. TBtools software is used to draw heat maps and other graphs were visualized using GraphPad Prism 8 procedure.

The membership function method was used to explore the response of seed yield components, yield, and quality to N fertilizer, and to determine the best fertilizer application rate for K. melanthera. The calculation formula for the forward membership function was y=(xaxmin)/(xmaxxmin), and the negative membership function was y=1(xaxmin)/(xmaxxmin), where the xa is the value of certain index, and the xmax and xmin represent the maximum and minimum values in the same index of seed traits.

Results

The influence of N fertilizer on seed yield components

Because of a typical plateau continental climate, plant growth and flowering of K. melanthera were restrained in the sowing year (2015). Therefore, we collected the seed yield-related and quality-related data from 2016 and 2017. N fertilizer had a significant influence (P < 0.05) on some yield components of seed including number of tillers (NTs), number of fertile tillers (NFTs), number of fertile florets (NFFLs) per spike, and 1,000-seed weight (TSW, Table 1). And non-significant difference was observed in all indices detected in 2016 and 2017.

Table 1:
Analysis of variance of the influence of N fertilizer on K. melanthera seed yield components.
Index Source of variance Df F-ratio P-value
NTs /m2 Year (Y) 1 5.313 0.026
N treatment (N) 7 36.021 0.000
Interaction Y×N 7 4.547 0.001
NFTs /m2 Year (Y) 1 7.790 0.008
N treatment (N) 7 28.417 0.000
Interaction Y×N 7 2.903 0.018
NSPs per fertile tillers Year (Y) 1 0.011 0.916
N treatment (N) 7 1.997 0.059
Interaction Y×N 7 0.293 0.956
NFLs per spike Year (Y) 1 0.799 0.373
N treatment (N) 7 3.802 0.071
Interaction Y×N 7 0.082 0.099
NFFLs per spike Year (Y) 1 0.848 0.361
N treatment (N) 7 104.774 0.000
Interaction Y×N 7 0.259 0.966
TSW/g Year (Y) 1 0.296 0.588
N treatment (N) 7 273.820 0.000
Interaction Y×N 7 5.615 0.000
DOI: 10.7717/peerj.14101/table-1

Note:

NTs, number of tillers; NFTs, number of fertile tillers; NSPs, number of spikes; NFLs, number of florets; NFFLs, number of fertile florets; TSW, 1,000-seed weight; Df, degrees of freedom.

Compared to 0 kg·hm−2 (N1, check), the NTs and NFTs in both trial years were significantly increased when 90–240 kg·hm−2 (N3–N8) of N fertilizer was applied (Fig. 1A). And they reached a maximum value under 180 kg·hm−2 (N6) or 210 kg·hm−2 (N7). The results showed that application of N fertilizer had a non-significant effect on number of spikes (NSPs) per fertile tillers and number of florets (NFLs) per spike in both trial years (P > 0.05, Figs. 1B and 1C). Compared to N1, the NFFLs per spike of N2 to N8 were significantly different, and the peak values were both observed under N6 treatment: 4.1 in 2016 and 4.2 in 2017 (Fig. 1D). The change trend of TSW is similar to that of NTs and NFTs, revealing that application of N fertilizer had a remarkable impact on TSW (Fig. 1E). It is worth noting that we observed a marginal decline (P > 0.05) in NTs, NFTs, NFFLs per spike, and TSW when the application amount of N fertilizer was over N7 in 2016 and 2017, which may be due to the negative effect of excess N fertilizer.

The seed yield components of K. melanthera as influenced by eight different rates of N fertilizer application during the trial years 2016–2017.

Figure 1: The seed yield components of K. melanthera as influenced by eight different rates of N fertilizer application during the trial years 2016–2017.

(A) Number of tillers (mean ± standard error, n = 4), and fertile tillers (n = 4); (B) Number of spikes per fertile tiller (n = 10); (C) Number of florets per spike (n = 10); (D) Number of fertile florets per spike (n = 10); (E) 1,000-seed weight (n = 4). For box plots, top and bottom of box represent 75th and 25th percentiles, respectively. Different lower-case letters above the column and curve indicate significant differences under different N fertilizer treatments in each trial year (Bonferroni, P < 0.05). Vertical bar represents the standard error of mean. Statistical significance was determined as * P < 0.05, ns was P > 0.05. Note: N1 – N fertilizer 0 kg·hm−2 (Control); N2 – N fertilizer 60 kg·hm−2; N3 – N fertilizer 90 kg·hm−2; N4 – N fertilizer 120 kg·hm−2; N5 – N fertilizer 150 kg·hm−2; N6 – N fertilizer 180 kg·hm−2; N7 – N fertilizer 210 kg·hm−2; N8 – N fertilizer 240 kg·hm−2.

The influence of N fertilizer on seed yield and quality

The doses of N fertilizer application had a significant influence on the seed yield (Table 2, P < 0.05), but a non-significant effect in improving seed quality (standard germination rate (SGR), germination energy (GE), accelerated aging germination rate (AAGR), dehydrogenase (DE) and acid phosphoesterase (APH) activity, P > 0.05). With increasing rates of N fertilizer application rate, three seed yield indices (harvested seed yield (HSY), potential seed yield (PSY), and presentation seed yield (PRSY)) showed a tendency of first increasing and then decreasing (Figs. 2A and 2B), and reached their peak values under N6 treatment in both 2016 and 2017. What is noteworthy is that excessive rates of N fertilizer application (N7–N8) had a non-significant suppression or improvement on seed quality, however, was accompanied by a reduction in three seed yield indices (Fig. 2, Table 3). During two-year independent field trials, we found that seed yield in 2017 was higher than that in 2016 under the same N fertilizer treatment, but the difference was non-significant.

Table 2:
Analysis of variance of the influence of N fertilizer on harvested seed yield and seed quality.
Index Source of variance Df F-ratio P-value
HSY/kg·hm−2 Year (Y) 1 0.291 0.113
N treatment (N) 7 145.194 0.000
Interaction Y×N 7 0.283 0.958
SGR/% Year (Y) 1 0.073 0.074
N treatment (N) 7 16.779 0.983
Interaction Y×N 7 0.607 0.747
GE/% Year (Y) 1 0.049 0.850
N treatment (N) 7 18.823 0.920
Interaction Y×N 7 0.404 0.895
AAGR/% Year (Y) 1 0.058 0.792
N treatment (N) 7 17.348 0.772
Interaction Y×N 7 0.265 0.821
DE activity/µg·mL−1 Year (Y) 1 1.654 0.203
N treatment (N) 7 2.168 0.061
Interaction Y×N 7 0.227 0.977
APH activity/nmol·min−1. 50 seeds Year (Y) 1 0.003 0.957
N treatment (N) 7 1.461 0.200
Interaction Y×N 7 0.244 0.972
DOI: 10.7717/peerj.14101/table-2

Note:

HSY, harvested seed yield; SGR, standard germination rate; GE, germination energy; AAGR, accelerated aging germination rate; DE, dehydrogenase; APH, acid phosphoesterase; Df, degrees of freedom.

(A) The seed yield of potential, presentation, and (B) harvested (mean ± standard error, n = 4) of K. melanthera as influenced by N fertilizer during the trial years of 2016–2017.

Figure 2: (A) The seed yield of potential, presentation, and (B) harvested (mean ± standard error, n = 4) of K. melanthera as influenced by N fertilizer during the trial years of 2016–2017.

Different lower-case letters above the column indicate significant differences under different N fertilizer treatments in each trial year (Bonferroni, P < 0.05). Vertical bar represents the standard error of mean. Note: N1 – N fertilizer 0 kg·hm−2 (Control); N2 – N fertilizer 60 kg·hm−2; N3 – N fertilizer 90 kg·hm−2; N4 – N fertilizer 120 kg·hm−2; N5 – N fertilizer 150 kg·hm−2; N6 – N fertilizer 180 kg·hm−2; N7 – N fertilizer 210 kg·hm−2; N8 – N fertilizer 240 kg·hm−2.
Table 3:
Seed quality of K. melanthera as influenced by N fertilizer during trial years 2016–2017.
Year N treatment SGR (%) GE (%) AAGR (%) DE activity (µg/mL) APH activity (nmol/min. 50 seeds)
2016 N1 86.00 ± 3.92a 83.50 ± 4.03a 64.00 ± 2.58a 15.17 ± 0.59a 2.39 ± 0.17a
N2 86.50 ± 3.50a 83.50 ± 4.43a 64.50 ± 4.19a 15.45 ± 0.61a 2.41 ± 0.26a
N3 86.00 ± 3.65a 84.00 ± 2.94a 65.00 ± 2.65a 15.38 ± 0.60a 2.47 ± 0.24a
N4 86.50 ± 3.50a 83.50 ± 3.30a 64.00 ± 3.65a 15.72 ± 0.97a 2.52 ± 0.40a
N5 87.00 ± 3.70a 83.50 ± 4.11a 65.50 ± 3.59a 15.88 ± 0.59a 2.50 ± 0.25a
N6 87.50 ± 4.03a 84.00 ± 4.08a 66.00 ± 2.94a 16.23 ± 1.04a 2.50 ± 0.20a
N7 86.50 ± 4.65a 83.50 ± 2.28a 65.00 ± 3.42a 15.96 ± 0.76a 2.48 ± 0.36a
N8 86.50 ± 3.59a 83.50 ± 3.10a 66.00 ± 4.55a 16.12 ± 0.73a 2.45 ± 0.28a
2017 N1 86.00 ± 9.38a 83.00 ± 2.38a 65.00 ± 4.51a 15.28 ± 0.57a 2.42 ± 0.18a
N2 86.50 ± 8.23a 83.50 ± 4.11a 65.00 ± 3.32a 15.54 ± 0.58a 2.40 ± 0.30a
N3 87.00 ± 3.46a 82.00 ± 3.74a 66.00 ± 2.83a 15.85 ± 0.52a 2.49 ± 0.37a
N4 86.50 ± 7.19a 82.50 ± 4.03a 65.00 ± 3.42a 16.34 ± 0.88a 2.45 ± 0.29a
N5 87.00 ± 11.60a 82.50 ± 2.87a 64.50 ± 3.87a 16.03 ± 0.79a 2.50 ± 0.35a
N6 86.50 ± 10.38a 84.00 ± 3.56a 65.00 ± 3.30a 16.49 ± 0.74a 2.48 ± 0.29a
N7 87.00 ± 8.67a 83.50 ± 2.63a 66.00 ± 1.63a 16.44 ± 0.46a 2.49 ± 0.38a
N8 86.50 ± 7.72a 83.50 ± 3.40a 64.00 ± 2.45a 15.95 ± 0.67a 2.47 ± 0.34a
DOI: 10.7717/peerj.14101/table-3

Note:

Measurement results are means ± standard error (n = 4). In the same column, standard error of mean followed by the same lower-case letters indicate no-significant differences under different N fertilizer treatments in each trial year base on a Bonferroni multiple comparison test at P < 0.05. SGR, standard germination rate; GE, germination energy; AAGR, accelerated aging germination rate; DE, dehydrogenase; APH, acid phosphoesterase; N1 – N fertilizer 0 kg·hm−2 (Control); N2 – N fertilizer 60 kg·hm−2; N3 – N fertilizer 90 kg·hm−2; N4 – N fertilizer 120 kg·hm−2; N5 – N fertilizer 150 kg·hm−2; N6 – N fertilizer 180 kg·hm−2; N7 – N fertilizer 210 kg·hm−2; N8 – N fertilizer 240 kg·hm−2.

The relationship of N application and seed yield or seed components

N application had a significant influence on seed yield of K. melanthera, which could be explained by the increase of seed quantity and weight under the N effect (Table 4). By means of path analysis, we found that seed yield was significantly correlated with number of fertile tillers/m2 (NFTs/m2, r = 0.944, P < 0.05), 1,000-seed weight (TSW, r = 0.912, P < 0.05), and number of fertile florets (NFFLs) per spike (r = 0.899, P < 0.05), but not with number of spikes (NSPs) per fertile tiller (r = 0.806, P > 0.05) and number of florets (NFLs) per spike (r = 0.818, P > 0.05). Also, seed yield (r = 0.909, P < 0.05), NFTs/m2 (r = 0.950, P < 0.05) and TSW (r = 0.921, P < 0.05) were observably associated with N treatment. It was therefore evident that N application was one of the most effective ways to enhance seed yield of K. melanthera.

Table 4:
Path analysis of seed yield to yield components and N application.
Source NFTs/m2 NSPs per fertile tillers NFLs per spike NFFLs per spike TSW/g Seed yield
N treatment 0.950* 0.538 0.683 0.739 0.921* 0.909*
Seed yield 0.944* 0.806 0.818 0.899* 0.912*
DOI: 10.7717/peerj.14101/table-4

Note:

The value in the table is the correlation coefficient “r”. Statistical significance was determined as * P < 0.05. NFTs, number of fertile tillers; NSPs, number of spikes; NFLs, number of florets; NFFLs, number of fertile florets; TSW, 1,000-seed weight.

Comprehensive analysis of membership function and heat map

The changes in the seed trait indices of K. melanthera represented a certain degree of correlation and difference based on the N fertilizer treatments, and a single index cannot fully explain the influence of N fertilizer treatment on plant reproduction and seed production. Consequently, it was necessary to conduct a comprehensive evaluation of multiple indices to determine the optimal amount of N fertilizer application. Based on the values of each seed trait index measured under different N treatments, including indices of six seed yield components, three seed yield, and five seed quality, to calculated the membership function values (Table 5). Higher membership function values represented more distinct improvement of K. melanthera seed yield or quality. The results indicated that the optimal deses of N fertilizer treatment was N6 (followed closely by N7) in both 2016 and 2017 (Table 5 and Table S2).

Table 5:
Membership function analysis based on multiple indicators for the effects of N fertilizer treatments on K. melanthera during trial year 2016.
Parameter N fertilizer treatments
N1 N2 N3 N4 N5 N6 N7 N8
NTs/m2 0.00 0.05 0.21 0.43 0.71 0.89 1.00 0.99
NFTs/m2 0.00 0.02 0.33 0.51 0.85 1.00 0.97 0.95
TSW/g 0.00 0.30 0.36 0.42 0.88 1.00 0.97 0.95
NSPs per fertile tillers 0.00 0.20 0.75 0.70 0.90 1.00 0.60 0.35
NFLs per spike 0.00 0.00 0.23 0.38 0.54 1.00 0.62 0.31
NFFLs per spike 0.00 0.00 0.25 0.50 0.75 1.00 0.75 0.25
PSY 0.00 0.05 0.32 0.44 0.77 1.00 0.83 0.72
PRSY 0.00 0.05 0.31 0.46 0.81 1.00 0.84 0.72
HSY/kg·hm−2 0.00 0.24 0.32 0.43 0.69 1.00 0.95 0.74
SGR/% 0.00 0.33 0.00 0.33 0.67 1.00 0.33 0.33
GE/% 0.00 0.00 1.00 0.00 0.00 1.00 0.00 0.00
AAGR/% 0.00 0.25 0.50 0.00 0.75 1.00 0.50 1.00
DE activity/µg·mL−1 0.00 0.26 0.20 0.52 0.67 1.00 0.75 0.90
APH activity/nmol·min−1. 50 seeds 0.00 0.15 0.62 1.00 0.85 0.85 0.69 0.46
Rank 8 7 6 5 3 1 2 4
DOI: 10.7717/peerj.14101/table-5

Note:

NTs, number of tillers; NFTs, number of fertile tillers; NSPs, number of spikes; NFLs, number of florets; NFFLs, number of fertile florets; TSW, 1,000 seeds weight; PSY, potential seed yield; PRSY, presentation seed yield; HSY, harvested seed yield; SGR, standard germination rate; GE, germination energy; AAGR, accelerated aging germination rate; DE, dehydrogenase; APH, acid phosphoesterase; N1 – N fertilizer 0 kg·hm−2 (Control); N2 – N fertilizer 60 kg·hm−2; N3 – N fertilizer 90 kg·hm−2; N4 – N fertilizer 120 kg·hm−2; N5 – N fertilizer 150 kg·hm−2; N6 – N fertilizer 180 kg·hm−2; N7 – N fertilizer 210 kg·hm−2; N8 – N fertilizer 240 kg·hm−2.

Up-regulation or down-regulation of the indices for seed yield components, seed yield, and seed quality were revealed by the heatmap and hierarchical cluster analysis. In the first trial year, up-regulation of all indices was most obvious for N6, followed by N5 and N7 (Fig. 3A). In addition, N5, N6, N7, and N8 could be classified into one group through hierarchical cluster analysis. Meanwhile, we also noted that germination energy (GE) and acid phosphoesterase (APH) revealed significant up-regulations at N3 and N4, respectively. In the second trial year, up-regulation of most indices were relatively significant for both N6 and N7, and could be classified into one group through cluster analysis (Fig. 3B). Interestingly, accelerated aging germination rate (AAGR) and standard germination rate (SGR) exhibited down-regulation at N6. Therefore, the heat maps obtained the same results as the comprehensive analysis of membership function in both trial years, that is N6 (180 kg·hm−2 urea) was the optimal N fertilizer treatment.

Heat maps. Log2-fold change of K. melanthera seed yield and quality under N fertilizer treatments during trial year 2016 (A) and 2017 (B).

Figure 3: Heat maps. Log2-fold change of K. melanthera seed yield and quality under N fertilizer treatments during trial year 2016 (A) and 2017 (B).

Red is up-regulated, blue is down-regulated. Note: NTs, number of tillers; NFTs, number of fertile tillers; NSPs, number of spikes; NFLs, number of florets; NFFLs, number of fertile florets; TSW, 1,000-seed weight; PSY, potential seed yield; PRSY, presentation seed yield; HSY, harvested seed yield; SGR, standard germination rate; GE, germination energy; AAGR, accelerated aging germination rate; DE, dehydrogenase; APH, acid phosphoesterase; N1 – N fertilizer 0 kg·hm−2 (Control); N2 – N fertilizer 60 kg·hm−2; N3 – N fertilizer 90 kg·hm−2; N4 – N fertilizer 120 kg·hm−2; N5 – N fertilizer 150 kg·hm−2; N6 – N fertilizer 180 kg·hm−2; N7 – N fertilizer 210 kg·hm−2; N8 – N fertilizer 240 kg·hm−2.

Discussion

Effect of N on seed yield

The main features of elite varieties of perennial grass are known to increase the yield of aboveground biomass and vegetation coverage of grassland stands thereby meets the needs of grassland husbandry and ecological restoration (Varadi, Kadar & Duda, 2019). Numerous investigations of increasing seed yield through different agronomic management practices have been carried out for many perennial grasses in the QTP, e.g., Elymus sibiricus, E. nutans, and Phalaris arundinacea, but few have probed into the effects of fertilization on seed yield. It is generally believed that a low fertilization rate results in seed yield not reaching the desired goal, but excess fertilizers led to increasing environmental pollution and reducing of seed yield (Zhang, 2017). Sufficient nitrogen (N) in the soil is the basis for high seed yield. Seed yield is significantly positively correlated with N fertilizer dose within a specific range, even if no other fertilizer is applied (Wang et al., 2005a). Most researchers choose N fertilizer to increase seed yield as it is the most consumable component during plant growth and seed maturation (Li et al., 2018b). In addition, N fertilizer can not only improve the developmental structure of plant roots and absorption of nutrients, but also the formation and size of grains, which can directly determine the seed yield, were promoted during later stages of plant growth (Liu et al., 2020). Therefore, compared with other fertilizers, N fertilizer is one of the most effective measures to improve seed yield (Oliveira et al., 2007). By combining seed yield indices (harvested seed yield (HSY), potential seed yield (PSY) and presentation seed yield (PRSY)) under different N application doses (N1–N6), we propose an optimal rate of N fertilizer application (N6) affected seed yield in K. melanthera (see in Figs. 2A and 2B). Similar results could be observed with an increase in the number of seeds or the weight of a single seed (Martiniello, 1998). The reason may be that the N accumulated in the vegetative organs is transported and redistributed to the grains, and thus enhancing seed yield due to fuller grains (Yi et al., 2020). As N fertilization continues to rise, seed yield indices were significantly decreased from N7 to N8. This may have been caused by excessive growth of competing plants encouraged by excessive N fertilizer application, or by dehydration of plant roots (Aly, 1993). It happens that there is a similar case, a study of a wildrye Elymus nutans in Tibet, which showed a peak yield of 2016 kg·hm−2 at a N fertilizer dose of 250 kg·hm−2, and a trend of falling continuously for HSY when the N fertilizer dose was continued increased (Song et al., 2008). Therefore, we should manage proper N fertilizer rate to optimize N-uptake in seed production of K. melanthera.

Effect of N on seed yield components

For most grass species, seed yield always depends on the multiplicative effect of the yield components (number of tillers (NTs), number of fertile tillers (NFTs), number of spikes (NSPs), number of florets (NFLs), number of fertile florets (NFFLs), and 1,000-seed weight (TSW)) (Dewitt et al., 2021). A large number of studies point out that N fertilizer can significantly improve the seed yield components. Song et al. (2008) found that N fertilizer could significantly raise the NFTs of E. nutans to improve the number of seeds per plant and thus increase seed yield. Xiao (1984) obtained a similar result, namely that the NFTs were higher with abundant N supply. For seed production of Poa pratensis, Zhou & Lu (2008) showed that N fertilizer significantly boosted NFTs and spike weight, thus enhancing seed yield. Investigation of the seed yield of tall fescue accounted for the fact that the amount of N fertilizer application is proportional to NFTs, NFLs and NFFLs per spike within a particular range (Ma et al., 2003). In addition, similar findings have been reported with regard to the seed yield of Lolium perenne, N fertilizer increased NSPs per fertile tiller, NFFLs per spike, and TSW (Hides, Kute & Marshall, 2010; Elgersma, Nijs & Eeuwijk, 1989). In this study, seed yield (r = 0.909, P < 0.05), NFTs/m2 (r = 0.950, P < 0.05) and TSW (r = 0.921, P < 0.05) were dramatically positively correlated with N fertilizer treatment by means of path analysis (Fig. 1 and Table 4). This result indicates that N treatment was one of the most effective approaches to improving K. melanthera production by enhancing NFTs and TSW. Furthermore, we also found that seed yield was non-significantly correlated with NSPs per fertile tiller (r = 0.806, P > 0.05) and NFLs per spike (r = 0.818, P > 0.05). These findings are not entirely consistent with other studies. Wang et al. (2005a) researched the influence of five yield components on the seed yield of Dactylis glomerata, and determined the following order of influence: NFTs/m2 > NFLs/spike > NFFLs/spike > TSW > NFLs/fertile tiller. However, a different order was obtained for K. melanthera: NFTs/m2 > TSW > NFFLs/spike > NFLs/spike > NFLs/fertile tiller. The reasons for the discrepancies may be the difference in sampling regions and studied species.

Effect of N on seed quality

The seed quality can be reflected by the standard germination rate (SGR), germination energy (GE), and accelerated aging germination rate (AAGR). The higher the values are for these, the stronger the seed vigor (Lin et al., 2020). It is worth noting that when AAGR is compared to SGR, the aging-accelerated conditions can explain the resistance to adversity that supplies a more reliable reference for seed growth in specific environments such as the QTP (Li & Mao, 2013; Chen et al., 2017). The number of SGR was inferred under suitable environmental conditions and cannot guarantee actual germination rate in the field and survivability under stress, therefore, it is always in synergy with AAGR to evaluate potential viability of seed (Mao, Hou & Wang, 2016). Dehydrogenase (DE) and acid phosphoesterase (APH) activity are often used as crucial indices to test vigor and utilization value of forage seed, because these two indices more accurately reflect the quality and stress resistance of seeds in different stages of growth than seed germination count (Li & Mao, 2013; Chen et al., 2017). Currently, a large number of researchers have shown that N fertilizer is a key factor in increasing seed yield of perennial herb, but the effect of N on pasture seed quality varied with the different species, growing conditions, harvesting methods, and/or harvesting time (Liu, 2020). The SGR and GE of Medicago sativa (Wang, 2005b) and Festuca elata (Cheng et al., 2003) were significantly increased by N fertilizer application, as well as APH activity of Festuca elata were also significantly upward. The reason may be that N fertilizer can promote the transfer of plant metabolites to reproductive organs and consequently increase plumpness and 1,000-seed weight of plant seed. However, the study of Qiao & Han (2010) demonstrated that N application had no significant effect on seed quality of E. nutans, a native perennial grass in QTP region. The results of this study were in agreement with those results of present study, which the seed quality indices (GR, GE, AAGR, DE and APH activity) of K. melanthera were not significantly affected by N fertilizer in either 2016 or 2017 (Table 2).

Conclusion

K. melanthera is a beneficial multi-purpose perennial grass in the eastern QTP region. It is commonly used in the treatment of degraded and sandy grassland, as well as for artificial planting to supply forage for livestock. However, the limited supply and poor quality of seeds have restricted its popularization and utilization in QTP. In this work, a two-year trial in eastern QTP was implemented to evaluate the effects of seed yield and quality of K. melanthera under different N fertilizer levels. The results showed that the optimal N application rate was 180 kg·hm−2 during trial years 2016–2017, and most of seed yield indicators reached the respective highest values at this N level, including the harvested seed yield (HSY), potential seed yield (PSY), presentation seed yield (PRSY), number of tillers (NTs), number of fertile tillers (NFTs), and 1,000-seed weight (TSW). The present findings provide a practical technical groundwork for improving seed productivity of K. melanthera in eastern QTP.

Supplemental Information

Basic nutrient of experimental soil.

The influence of N fertilizer on seed yield components, seed yield and quality in both trial years.

DOI: 10.7717/peerj.14101/supp-1

Membership function analysis based on multiple indicators for the effects of N fertilizer treatments on K. rigidula during the trial years of 2017.

Note. NTs, number of tillers; NFTs, number of fertile tillers; NSPs, number of spikes; NFLs, number of florets; NFFLs, number of fertile florets; TSW, 1,000-seed weight; PSY, potential seed yield; PRSY, presentation seed yield; HSY, harvested seed yield; SGR, standard germination rate; GE, germination energy; AAGR, accelerated aging germination rate; DE, dehydrogenase; APH, acid phosphoesterase; N1 – N fertilizer 0 kg·hm−2 (Control); N2 – N fertilizer 60 kg·hm−2; N3 – N fertilizer 90 kg·hm−2; N4 – N fertilizer 120 kg·hm−2; N5 – N fertilizer 150 kg·hm−2; N6 – N fertilizer 180 kg·hm−2; N7 – N fertilizer 210 kg·hm−2; N8 – N fertilizer 240 kg·hm−2.

DOI: 10.7717/peerj.14101/supp-2

Raw data of Kengyilia melanthera seed yield components.

DOI: 10.7717/peerj.14101/supp-3

Raw data of Kengyilia melanthera seed yield and quality.

DOI: 10.7717/peerj.14101/supp-4
8 Citations   Views   Downloads