Evaluation and analysis of intraspecific competition in maize: A case study on plant density experiment

ZHAl Li-chao , XlE Rui-zhi MlNG Bo Ll Shao-kun MA Da-ling

1 Institute of Crop Sciences, Chinese Academy of Agricultural Sciences/Key Laboratory of Crop Physiology and Ecology, Ministry of Agriculture, Beijing 100081, P.R.China

2 Center for Agricultural Resources Research, Institute of Genetics and Development Biology, Chinese Academy of Sciences,Shijiazhuang 050021, P.R.China

3 Key Laboratory of Cash Crop Stress Biology for Ulanqab, Jining Normal University, Jining 012000, P.R.China

Abstract Intraspecific competition is a common phenomenon in agricultural production, and maize is one of the most sensitive grass species to intraspecific competition due to its low tillering ability. This study evaluated and analyzed intraspecific competition in maize, and screened competitive indices that could be used to evaluate intraspecific competition in a maize population.A 2-year field experiment was conducted using the maize hybrid Zhongdan 2 at 12 plant densities ranging from 1.5 to 18.0 plants (pl) m–2. The results showed that the response of single-plant grain yield and dry matter at harvest to increased plant density decreased exponentially and that the harvest index decreased linearly. The response of population-level grain yield to plant density was curvilinear, producing a maximum value at the optimum population density. However, the yielddensity equation agreed well with the Steinhart-Hart equation curves, but not with the quadratic equation curves reported by most previous studies. Competitive indices are used to evaluate competition in a plant population or plant species. The present results show that competitive intensity (CI) and absolute severity of competition (ASC) increased with increasing plant density; however, relative yield (RY) and relative reproductive efficiency (RReff) decreased. The different responses of these indices reflect different aspects of competition. According to the analysis of CI, ASC, RY, and RReff, higher CI and ASC values indicate higher intraspecific competition, whereas higher RY and RReff values indirectly reflect lower intraspecific competition. These competitive indices evaluate not only the intraspecific competitive intensity under different plant densities of the same cultivar but also those of different cultivars under the same plant density. However, some overlap exists in the calculations of ASC, CI, and RY, so one could simply select any one of these indices to evaluate intraspecific competition in a maize population. In conclusion, the present study provides a method to evaluate intraspecific competition in maize populations, which may be beneficial for breeding high-yield maize varieties in the future.

Keywords: maize, intraspecific competition, plant density, competitive indices, grain yield

1. lntroduction

Competition is a universal phenomenon, not only in nature but also in agricultural ecosystems. Competition includes interspecific competition and intraspecific competition.Most studies on interspecific competition in agricultural ecosystems have focused on crop-weed competition and intercropping systems (Lemerle et al. 1996; Vandeleur and Gill 2004; Zhang et al. 2011; Radicetti et al. 2012),and competitive indices used to evaluate interspecific competition have also been reported (Snaydon 1991;Willams and McCarthy 2001; Weigelt and Jolliffe 2003;Zhang et al. 2011; Zhai et al. 2015, 2016). However, little information is available on the evaluation of intraspecific competition in maize populations.

Intraspecific competition is closely related to increased plant density (Maddonni and Otegui 2006) for belowground resources (e.g., water and nutrients), above-ground resources (e.g., light), or both. Maize is one of the mostsensitive grass species to intraspecific competition due to its low tillering ability. When plant density increases, the relative resource availability for individual plants decreases,and population competitive intensity increases, resulting in decreases in plant biomass and grain yield (GY) per plant(Edmeades and Daynard 1979; Tetio-Kagho and Gardner 1988). Population-level GY shows a curvilinear response to plant population density, producing a maximum value at the optimum population density (Sangoi et al. 2002; Assefa et al.2016; Qian et al. 2016). However, maize hybrids differ in the yield-density response due to different density tolerances.

During the past 30 years, maize GY has been greatly improved, mainly due to the increased density tolerance of maize hybrids (Duvick 2005a, b; Li and Wang 2009).Increasing plant density is one of the most important culture practices in maize production; however, it also promotes high intraspecific competition pressure, which increases plantto-plant variability within the stand. Donald (1981) argued that individual plants making up a high-yield crop should be weak competitors; thus, a high-yield crop population should have lower intraspecific competitive ability. To date, no study has measured the intensity of intraspecific competition in a maize population. In the present study, we conducted a density experiment ranging from 1.5 to 18 plants (pl) m–2and referred to competitive indices used in plant competition.This study (1) evaluated intraspecific competition in maize through the yield-density response, (2) analyzed dynamic changes in different competitive indices, and (3) screened competitive indices to evaluate intraspecific competition in a maize population.

2. Materials and methods

2.1. Experimental site and weather conditions

A field experiment was conducted during the 2013 and 2014 growing seasons at the Gongzhuling experimental station of the Chinese Academy of Agricultural Science (43°53′N,124°81′E), which is located in a humid, continental monsoon climate in Gongzhuling County of Jilin Province, China.The maize was grown from late April to late September under rain-fed conditions and with ridge planting. The mean annual air temperature at the experimental station is 5.6°C, annual rainfall is 594.8 mm, and the annual frost-free period is approximately 144 d. The primary soil in the area is Cher-nozem, with 2.63% organic matter, 0.15% total N,124.90 mg kg-1available N, 28.52 mg kg-1available P, and 184.47 mg kg–1available K in the upper 0–30 cm of the soil profile. The mean daily solar radiation, rainfall, and the mean temperatures during the two growing seasons are shown in Fig. 1. A low mean temperature was observed in May 2014,which contributed to delayed emergence. In addition, rainfall during the 2014 growing season was significantly lower than that in 2013, particularly during August.

2.2. Experimental design and field management

The experiment was carried out using a single-factor randomized block design with three replicates per treatment.The maize hybrid Zhongdan 2 (ZD2) (Mo17♀×Zi330♂) was used in the experiment. Twelve plant densities ranging from 1.5 to 18.0 pl m–2were tested, with 1.5 pl m–2as the plant density interval. Each plot comprised nine 8-m-long rows planted 0.65 m apart. There was a buffer line between treatments. All measurements were taken from the three central rows of each plot, leaving 0.5-m borders at each row end. Seeds were manually sown on 7 May 2013 and 25 April 2014 at a rate of two seeds per hill, and plots were thinned to one plant per hill at the V3 (three fully expanded leaves) stage (Ritchie et al. 1993). Weeds were controlled manually to eliminate confounding effects. N, P, and K chemical fertilizers were applied at 150, 45, and 45 kg ha-1,respectively. N was top-dressed at 75 kg ha-1during the V12 (12 fully expanded leaves) stage. All treatments were applied under rain-fed conditions.

2.3. Plant sampling and measurements

Shoot dry matter was measured at early anthesis stage and at physiological maturity stage. Three plants within 3 m2of the middle rows of each plot in each treatment were selected randomly and manually cut at the ground level. The plants were separated into stalks, leaves, sheaths, tassels, and ears and oven-dried to measure dry matter accumulation(DMA). Harvesting was manually performed at physiological maturity stage in the middle rows of each plot. GY was assessed at a standard moisture of 14%. The barren plant rate was assessed the day before each hybrid was harvested. Plants were considered barren when they did not present a visible ear or produced a rudimentary female inflorescence with fewer than 20 kernels. Uniformity was calculated based on kernels per ear, and was the reciprocal of the coefficient of variation of kernels per ear.

Fig. 1 Mean daily solar radiation, total rainfall, and air temperature during the two growing seasons at the Gongzhuling experimentalstation, Jilin Province, China.

2.4. Definition and calculation of indices

Relative yield (RY)In the present study, we revised this competitive index used in a mixture or intercrop to quantify the competitive effect of intraspecific competition, and calculated it as:

Where, BYhcand BYlcare measures of plant biomass yield under low (or no) competition (i.e., very wide plant spacing)and high competition, respectively. We assumed that plant density of 1.5 pl m–2was low or no competition, and that the other plant densities represented higher competition. A RY value higher than 1 indicated that intraspecific competition increased DMA, RY less than 1 indicated that intraspecific competition reduced DMA, and RY equal to 1 indicated no intraspecific competition.

Competitive intensity (Cl)CI was calculated as follows(Bonser 2013):

CI=(Sizelc-Sizehc)/Sizelc

Where, Sizelcand Sizehcare measures of plant size under low (or no) competition (i.e., very wide plant spacing)and under high competition, respectively. We used shoot biomass yield as a measure of plant size; plant density of 1.5 pl m–2was taken as low or no competition, and other plant densities represented greater competition.

Absolute severity of competition (ASC)ASC in a purestand at a given density was calculated as (Snaydon and Satorre 1989):

ASCii=log10(Wi0/Wii)

Where, Wi0is the yield per plant of hybrid i under no competition (i.e., at very low density), and Wiiis the yield per plant of hybrid i in a pure stand at a given density. In this study, the population density 1.5 pl m–2was assumed to be the no-competition condition.

Relative reproductive efficiency (RReff)RReffwas calculated as (Bonser 2013):

RReff=log10(Reff-hc/Reff-lc)

Where Reff-hcand Reff-lcare the values for reproductive efficiency under high and low (or no) competition,respectively. RReffvalues were log10transformed due to some extremely high values for reproductive efficiency when reproductive efficiency was estimated as seed number.As these are log10values, a positive RReffvalue indicates instances where reproductive efficiency is greater under competition than it is under no competition.

2.5. Statistical analysis

Data were subjected to analysis of variance using SPSS ver.19.0 Software (SPSS Inc., Chicago, IL, USA). Regressions between maize yield, dry matter yield, uniformity, barren plant rate, competitive indices, and plant density were analyzed. Graphs were plotted in either Sigmaplot 12.0(Systat Software, Inc., San Jose, CA, USA) or Excel 2010(Microsoft, Inc., Redmond, WA, USA) Software. Treatment means were compared by computing least significant differences to identify significant differences at the 0.05 probability level.

3. Results

3.1. Responses of individual-plant GY, DMA, and the harvest index (Hl) to plant density

Individual-plant GY, above-ground DMA, and HI decreased gradually with increasing population density (Fig. 2). The response of GY to increased plant density was similar to that of DMA, as both decreased exponentially. However, HI decreased linearly, the trend was similar in 2013 and 2014.Across plant densities, GY, DMA, and HI decreased by 89.2,90.5, and 80.1%, and by 84.1, 44.9, and 41.1% in 2013 and 2014, respectively. The effects of year and plant density on DAM per plant were significant. However, only significant plant density effect was observed for GY per plant (Table 1).

Fig. 2 Responses of individual grain yield, dry matter accumulation, and harvest index to plant density. pl, plants.＊＊, significant at P＜0.01.

3.2. Responses of population-level grain yield and dry matter accumulation to plant density

The responses of population-level GY and biomass yield to plant density for maize are shown in Fig. 3. Across plant densities, the dynamics of GY agreed well with the curves of the Steinhart-Hart equation, which produced a maximum value at the optimum population density. The optimum plant densities for the tested maize hybrids were 4.5 and 6 pl m–2in 2013 and 2014, respectively. However, the response of population-level biomass yield to plant density fit a cubic curve equation, with optimum plant densities for the maximum biomass yield of 7.5 and 9 pl m–2in 2013 and 2014, respectively. Through the analysis of variance on grain yield and biomass, the effects of year and plant density were significant in both growing seasons (Table 1).

3.3. The response of uniformity to plant density

The uniformity of the maize population decreased linearly with increasing plant density (Fig. 4). The uniformity of the maize hybrid varied from 11.5 to 3.7 across densities in 2013 and from 8.1 to 2.0 across plant densities in 2014.Different performances were observed between years. In most cases, uniformity was better in 2013 than that in 2014.

3.4. Relative yield

RY decreased exponentially with increasing plant density(Fig. 5). The RY values of other treatments were all lower than 1 compared with the 1.5 pl m–2plant density,indicating that increased plant density increase intraspecific competition and thus reduced biomass yield. Although the RY of each treatment under higher plant densities (≥3 pl m–2) was higher in 2013 than that in 2014, the relative trends were consistent across the two years.

3.5. Competitive intensity and absolute severity of competition

Both the CI and ASC of the maize hybrid increased as plant density increased (Fig. 6). However, CI and ASC differed in their response to plant density.

The response of CI to plant density fit a cubic curveequation. CI increased by an average of 0.073 with each density increment of 1.5 pl m–2, but it increased by 0.029,from 13.5 to 18.0 pl m–2, with each increment of plant density from 3.0 to 13.5 pl m–2in 2013. In 2014, CI increased by 0.082 with each density increment from 3.0 to 12.0 pl m–2, and it increased by 0.021 per density increment from 12.0 pl m–2.

Table 1 Analysis of variance of single plant and population level dry matter accumulation (DMA) and grain yield (GY)

Fig. 3 Response of population grain yield and biomass to plant population density. pl, plants. ＊＊, significant at P＜0.01.

ASC increased linearly in response plant density, and thestraight-line slope was nearly the same for the two growing seasons. Average ASC increased by 0.038 and 0.042 with each increment of 1.5 pl m–2in 2013 and 2014, respectively.

Fig. 4 Response of uniformity to plant density. pl, plants. ＊＊ ,significant at P＜0.01.

Fig. 5 Response of relative yield to increased plant density. pl, plants. Different small letters between plant density treatments in the same year indicate significant difference (P＜0.05). ＊＊＊, significant at P＜0.001. Bars mean SE.

3.6. Relative reproductive efficiency and barren plant rate

Fig. 7 shows the response of RReffto increased plant density. Compared with plants under low or no competition(i.e., plant density of 1.5 pl m–2), the single-plant RReffvalues were all negative, and they decreased linearly with increasing intraspecific competition. However, based on the population-level measurement, the RReffvalues were all positive, except under plant density of 18.0 pl m–2in 2014.RRefffirst increased and then decreased as intraspecific competition increased; the response fit a normal quadratic equation.

Fig. 8 shows that the barren plant rate occurred at a plant density of 4.5 pl m–2during both years and linearly increased as plant density increased. In most cases, no significant difference was observed between years under the same plant density.

4. Discussion

Fig. 6 Response of competition intensity and absolute severity of competition to plant density. pl, plants. ＊＊＊, significant at P＜0.001.

Fig. 7 Response of relative reproductive efficiency to plant density. A, relative reproductive efficiency was measured based on single plant. B, relative reproductive efficiency was measured based on population. pl, plants. ＊＊ and ＊＊＊, significant at P＜0.01 and P＜0.001, respectively.

Intraspecific competition during crop production associated with high plant density generates an impoverished environment in terms of available resources. As plant density increases, both plant biomass and GY per plant decline (Edmeades and Daynard 1979; Tetio-Kagho and Gardner 1988) due to decreases in the water, nutrients,and light resources available to individual plants (Echarte et al. 2000; Sangoi et al. 2002; Maddonni and Otegui 2004). Maize is one of the most sensitive grass species to intraspecific competition; however, few studies have evaluated the intensity of intraspecific competition in different maize populations.

4.1. Evaluation of yield-density response of tested maize hybride

Fig. 8 Response of barren plant rate to plant density. ＊＊＊,significant at P＜0.001.

The present results showed that single-plant DMA and GY decreased exponentially in response to increased plant density, similar to results reported by Li et al. (2015).However, our results showed that HI decreased linearly in response to increased plant density, which differs from previous studies (Tollenaar 1992; Echarte and Andrade 2003; Li et al. 2015). Furthermore, the present study showed that the response of population-level GY to plant density was curvilinear, reaching a maximum at the optimum plant density. The optimum plant density was 4.5 and 6 pl m–2in 2016 and 2017, respectively. In addition, the yield-density response curves was similar to previous studies (Sangoi et al. 2002; Assefa et al. 2016; Qian et al. 2016), but the present results show that the yield-density response fit the Steinhart-Hart equation curves, not a quadratic equation as reported in most previous studies (Sangoi et al. 2002; Qian et al. 2016). The difference in the yield-density response curves between the present results and previous reports may be due to the different plant density ranges used; mostprevious studies were conducted under four to six different plant densities (Echarte et al. 2000; Sangoi et al. 2002; Qian et al. 2016), whereas the present research considered 12 plant densities (i.e., 1.5–18 pl m–2). This wide plant density range, employed in a field experiment, better reflected the true yield-density response curves. Through the analysis of the two years’ data, the results were similar across the two growing seasons. However, the yield performance in 2016 was generally higher than that in 2017, and the lower yield in 2017 was mainly caused by less rain at the flowering period (Fig. 1).

4.2. Evaluation and analysis of intra-specific competitive indices

Competition has been evaluated using various indices.According to Weigelt and Jolliffe (2003), competitive indices are grouped into three types: indices to quantify the intensity of competition, the effect of competition, and the outcome of competition. However, most of these indices have been used to evaluate interspecific competition, and the formulae for calculating these indices are based on plant performance in pure stands and mixtures. CI and ASC have been used to quantify the intensity of intraspecific and interspecific competition (Snaydon and Satorre 1989; Weigelt and Jolliffe 2003; Bonser 2013). In the present study, CI and ASC both increased with increasing plant density, indicating that the intensity of intraspecific competition increased, mainly due to a reduction in available resources for individual plants.RY is used to quantify the effect of plant competition.The present results showed that RY decreased as plant density increased. Furthermore, the RY value above a plant density of 1.5 pl m–2was always ＜1, indicating that intraspecific competition had an adverse effect on growth and development of individual plants in a population. RReff,which is used to quantify the outcome of plant competition(Weigelt and Jolliffe 2013), indicates whether reproductive efficiency is greater or less under competition than it is under no competition. According to Bonser (2013), a positive RReffvalue indicates that reproductive efficiency is greater under competition than it is under no competition.In the present study, the RReffvalue based on individual plants was negative under competition, and it decreased with increasing plant density, indicating that intraspecific competition reduced the reproductive efficiency of individual plants. However, population-level RRefffirst increased and then decreased with increasing intraspecific competitive intensity. Within a certain plant density range (i.e., below the optimum plant density), environmental resources meet the growth and developmental requirements of individual plants in the population as intraspecific competition increases.Supra-optimum plant density further increases intraspecific competition; however, the resource cannot evenly meet the needs of individuals in a population. In the present study,intraspecific competition accelerated plant-to-plant variability and the barren plant rate (Figs. 4 and 8), which further led to reductions in population-level RReffand GY. Maddonni and Otegui (2006) also demonstrated that high intraspecific competition pressure generated increased plant-to-plant variability within a stand and the appearance of dominant and dominating plants in a population, which may affect ultimate yield production. Their conclusion agreed with our results. In addition, the responses of competitive indices to plant density differed between indices, mainly due to the fact that different indices reflect different aspects of competition(Weigelt and Jollife 2003). The results were consistent across the two growing seasons. According to the analysis of the competitive indices, intraspecific comeptition was generally lower in 2016 than that in 2017, under the same plant density, this is mainly due to the drought stress cuased by the reduced rainfall after anthesis.

All competitive indices used in this study quantified the intensity of intraspecific competition in a maize population.These indices can be used to evaluate the intraspecific competition intensity not only of different populations of the same cultivar but also of different cultivars under the same plant density. A higher CI or ASC value of a cultivar leads to stronger intraspecific competition at the same plant density. A higher RY indicates that intraspecific competition has less effect on individual plant growth, demonstrating that crop populations have lower intraspecific competition.Similar to RY, the RReffvalue can also be used to indirectly evaluate the intraspecific competitive pressure of different crop populations, as a higher RReffvalue indicates lower intraspecific competition. Taken together, these results indicate that CI, ASC, RY, and RReffcan all be used to evaluate the intensity of intraspecific competition. Moreover,and there is some overlap in the calculations of ASC, CI,and RY. For instance, CI is equal to 1 minus RY, and ASC is the base-10 logarithm of the reciprocal of RY. Therefore,we can simply select any one of these indices (i.e., CI, ASC,RY, or RReff) to evaluate intraspecific competition in a crop population.

Donald (1968, 1981) argued that high-yielding crop populations have lower intraspecific competition. The present study provides a method to evaluate intraspecific competition in a maize population. These indices can be used to evaluate not only intraspecific competition intensity in different populations of the same cultivar but also intraspecific competitive intensity of different cultivars under the same plant density. Therefore, this study may be beneficial for high-yield breeding of maize or other crops. However,maize hybrids differ in plant density tolerance, and the competitive indices based on crop biomass yield may vary depending on the plant density tolerance, so further studies are needed.

5. Conclusion

This study showed that the response of population-level GY to plant density fit the Steinhart-Hart equation curves,producing a maximum value at the optimum population density. Competitive indices were used to evaluate competition within a maize population. The results indicate that CI and ASC increased with increasing plant density,whereas RY and RReffdecreased. According to the analysis of CI, ASC, RY, and RReff, higher CI and ASC values indicate higher intraspecific competition, whereas higher RY and RReffvalues indirectly reflect lower intraspecific competition.All of these indices can be used to evaluate not only the intraspecific competitive intensity of different plant densities of the same cultivar but also that of different cultivars under the same plant density. However, there is some overlap in the calculation of ASC, CI, and RY, so we could simply select any one of these indices to evaluate intraspecific competition in a maize population. The present research provides a method to evaluate intraspecific competition of a maize population, which may be beneficial for maize breeding.

Acknowledgements

Authors wish to thank the National Key Research and Development Program of China (2017YFD0300302), the earmarked fund for China Agriculture Research System(CARS-02-25), and the Science and Technology Innovation Project of Chinese Academy of Agricultural Sciences for their support.