Agronomy Journal Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 3 October 2006
Published in Agron J 98:1532-1543 (2006)
DOI: 10.2134/agronj2006.0038
© 2006 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Maddonni, G. A.
Right arrow Articles by Otegui, M. E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Maddonni, G. A.
Right arrow Articles by Otegui, M. E.
Agricola
Right arrow Articles by Maddonni, G. A.
Right arrow Articles by Otegui, M. E.
Related Collections
Right arrow Crop Physiology & Metabolism
Right arrow Maize
Right arrow Crop Ecology
Right arrow Maize Management

Production Papers

Row Width and Maize Grain Yield

Gustavo A. Maddonnia,*, Alfredo G. Cirilob and M. E. Oteguia

a Dep. de Producción Vegetal, Fac. de Agronomía, Univ. de Buenos Aires, Av. San Martín 4453, Ciudad de Buenos Aires (C1417DSE), Argentina
b Estación Experimental Agropecuaria Pergamino, Instituto Nacional de Tecnología Agropecuaria, Ruta 32 km 4.5, C.C. 31, Pergamino (B2700WAA), Buenos Aires, Argentina

* Corresponding author (maddonni{at}agro.uba.ar)

Received for publication February 9, 2006.

   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Maize (Zea mays L.) grain yield increase in narrow rows (0.35–0.50 m) may be related to the improvement of light interception around silking, but percentages of grain yield increase are generally lower than those of light interception, suggesting a lower efficiency to convert the amount of intercepted photosynthetic active radiation (IPAR) into aboveground phytomass. We analyzed the effects of plant population and row spacing on grain yield and its components (kernel number and kernel weight) and on the underlying processes, the IPAR around silking and during the effective grain filling period, and radiation use efficiency (RUE) during both periods. Field experiments were conducted in Argentina from 1997 to 2001. Five hybrids were cultivated at a wide range of plant population densities (3, 4.5, 9, and 12 plants m–2) and row spacings (0.35, 0.50, 0.70, and 1 m) without water and nutrient limitations. Row spacing reduction increased IPAR around silking at low plant densities ({approx}8 and 4% for 3–4.5 and 9–12 plants m–2, respectively) but did not modify RUE during this period. Morphogenetic limitations in the reproductive organs of plants (number of florets per ear) cultivated at low stand densities, suppressed the slight benefits of enhanced light capture under narrow rows, yielding similar kernel numbers at any row spacing. Contrarily, a postsilking RUE reduction ({approx}13–16%) of crops in narrow rows compared to those in wide rows minimized or counterbalanced any positive effect on IPAR during the grain-filling period. Hence, for the tested growing conditions, no benefits could be expected in terms of grain yield by reducing row spacing from the present 0.7- to 0.8-m inter-row distance.

Abbreviations: CG, crop growth • ENSO, El Nino Southern Oscillation phenomenon • fPAR, fraction of incident photosynthetic active radiation intercepted by crops • IPAR, the amount of intercepted photosynthetic active radiation • LAI, leaf area index • NR, narrow rows • PAR, photosynthetic active radiation • R/FR, red: far-red ratio • RUE, radiation use efficiency • WR, wide rows


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
MAIZE GRAIN YIELD is mainly determined by kernel number per unit land area (Bolaños and Edmeades, 1993, 1996; Cirilo and Andrade, 1994; Otegui, 1995; Otegui et al., 1995). This grain yield component is positively related to crop growth around silking (Andrade et al., 1999; Tollenaar et al., 1992), and biomass allocation to the ears (Echarte et al., 2004). Crop growth depends on the amount of intercepted photosynthetic active radiation (IPAR) and the efficiency to convert IPAR in aboveground phytomass, commonly referred to as radiation use efficiency (RUE). If we focus on IPAR and RUE, the former was related to canopy size, canopy architecture (spatial distribution of shoot organs) and incident photosynthetic active radiation (PAR) (Boote and Loomis, 1991; Flénet et al., 1996; Jones and Kiniry, 1986; Maddonni and Otegui, 1996; Maddonni et al., 2001a; Otegui et al., 1995), and the latter was species specific (Kiniry et al., 1989) and was modified by nutrient stress, water stress (Muchow, 1989; Muchow and Davis, 1988; Uhart and Andrade, 1995), and low temperatures (Andrade et al., 1992, 1993a). Hence, most cultural practices involved in maize husbandry such as sowing date, plant population density, row spacing, fertilization, irrigation, and pest management affect grain yield by modifying IPAR, RUE, or both variables.

Under temperate environments without water and nutrient restrictions, plant population density increases had the greatest positive impact on IPAR and crop growth rate around silking, which was reflected in kernel number and grain yield (Andrade et al., 1993b, 1999, 2000; Maddonni et al., 2001a; Maddonni and Otegui, 2004; Westgate et al., 1997). Conversely, RUE around silking was not affected by plant population density increases (Westgate et al., 1997).

Contradictory row spacing effects on light interception and grain yield have been documented (Farnham, 2001; Flénet et al., 1996; Hodges and Evans, 1990; Ottman and Welch, 1989; Widdicombe and Thelen, 2002) when the resulting plant spatial arrangement (square or rectangular planting geometry, commonly referred to as narrow and wide rows, respectively) was not considered. In a previous work (Maddonni et al., 2001a), it was demonstrated that a more uniform plant spatial distribution increased light interception at silking only when canopy size was below the critical leaf area index (LAI) value to maximize light interception, which was close to four in maize (Maddonni and Otegui, 1996). Consequently, grain yield increase in narrow rows depends on the actual improvement in light interception at silking (Andrade et al., 2002; Barbieri et al., 2000). The rate of increase in grain yield per unit increase of light interception at silking under narrow rows was less than one (Andrade et al., 2002), which suggested reduced RUE in narrow rows as compared to wide rows. Unfortunately, the response of RUE was not reported in the mentioned work. Thus, benefits of narrow rows in terms of light interception around silking could be minimized or counterbalanced by reduced RUE.

Evidences summarized above describe plant population density and row spacing effects on grain yield mainly by observed changes in IPAR around silking and its relationship with kernel number. For some maize hybrids, generally those cultivars with a large kernel size, kernel weight may contribute to grain yield variations (Echarte et al., 2000; Otegui, 1995) and is related to plant weight gain per kernel during the grain-filling period (Borrás and Otegui, 2001; Cirilo and Andrade, 1996; Maddonni et al., 1998). The postsilking source of assimilates depends on LAI duration, incident PAR values and RUE. In a recent work, Borrás et al. (2003) found that a reduced LAI duration could be expected in response to increased plant population density due to a higher leaf senescence rate during grain filling. Similarly, an impoverished light environment with an enriched far-red radiation at the lowermost leaf stratum of crops at high plant population densities and narrow rows (Borrás et al., 2003) increased the vertical profile of leaf N content (Chelle and Andrieu, 1999; Drouet and Bonhomme, 1999) and reduced the crop RUE (Sinclair and Horrie, 1989). Consequently, plant population density and row spacing effects on IPAR and RUE during the grain-filling period could be reflected on kernel weight and grain yield in large kernel size hybrids.

We hypothesize that at low plant population densities higher grain yields in narrow rows compared to wide rows are the result of an increased IPAR during silking and a similar RUE. In contrast, at high plant population densities, the lower grain yields in narrow rows compared to wide rows are the result of a similar IPAR around silking, but a lower RUE and/or IPAR during the effective grain-filling period. The plant population density and row spacing interaction effect on grain yield would be determined by an increased kernel number of crops cultivated at low plant population densities in narrow rows and an enhanced kernel size of crops cultivated at high plant population densities in wide rows. For testing these hypotheses, we grew five hybrids with contrasting kernel weight (<300 mg kernel–1 and >300 mg kernel–1) at a wide range of plant population densities (3, 4.5, 9, and 12 plants m–2) and row spacings (0.35, 0.50, 0.70, and 1 m). We analyzed the effects of treatments on grain yield and grain yield components, and the underlying processes: the IPAR around silking and during the effective grain-filling period, and the RUE during the mentioned periods.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Experimental Design and Growing Conditions
Field experiments were conducted in Argentina during the 1997–1998 and 1998–1999 growing seasons at Salto (34°33' S, 60°33' W), and the 2000–2001 and 2001–2002 growing seasons at Pergamino (33°56' S, 60°34' W) on silty clay loam soils (fine, illitic, thermic Typic Argiudoll; USDA Soil Surv. Syst.). At Salto, sowing took place on 8 Oct. 1997 (E1), and 12 Nov. 1998 (E2). Treatments at this site were a factorial combination of three plant populations (3, 9, and 12 plants m–2), two row spacings (0.35 and 0.70 m) and two (DK696 and DK757 during 1997–1998) or three (DK696, DK757 and an experimental hybrid, Exp980, during 1998–1999) hybrids. At Pergamino, sowing took place on 12 Oct. 2000 (E3), and 31 Oct. 2001 (E4) with a factorial combination of two plant populations (4.5 and 9 plants m–2), four row spacings (0.35, 0.50, 0.70 and 1 m) and two hybrids (AX882 and AX889).

The combination of plant population densities and row spacings determined a wide range of plant spatial arrangements, which were described by plant rectangularity (Willey and Heath, 1969; Eq. [1]):

Formula 1[1]
where X1 is the longest distance between two plants (i.e., row spacing or the distance between plants within the row) and X2 is the shortest distance. The plant rectangularity values varied from 1.1 to 9.9.

At 4.5, 9, and 12 plants m–2, plant spatial arrangement of crops in rows 0.35 and 0.50 m apart was referred to as narrow rows (plant rectangularity approximately 1.55 ± 0.19). In contrast, plant spatial arrangement of crops in rows 0.70 and 1 m apart was referred to as wide rows (plant rectangularity approximately 5.37 ± 1.14). At 3 plants m–2, crops in narrow rows (plant rectangularity = 1.70) were those in rows 0.70 m and crops in wide rows (plant rectangularity = 2.74) were those in rows 0.35 m apart.

Tested cultivars were classified as small (DK696, DK757 and Exp980) and large (AX882 and AX889) kernel weight hybrids, based on actual kernel weight (<300 m per kernel and >300 mg kernel–1 for small and large kernel weight, respectively) at commercial plant densities (Gambín et al., 2006; Maddonni and Otegui, 2006). The small kernel weight hybrids have a similar thermal time requirement with a base temperature of 8°C (Ritchie and NeSmith, 1991) from sowing to silking ({approx}900°Cd) and to physiological maturity ({approx}1800°Cd). The large kernel weight hybrids have more thermal time requirement than the small kernel weight hybrids, either from sowing to silking ({approx}975°Cd) or to physiological maturity ({approx}1955°Cd).

In all experiments, treatments were arranged in a split-split-plot design with three replicates. At Salto, row spacing was the main-plot, plant population density the sub-plot and hybrids the sub-subplot. Each plot was 5 (1997–1998) or 12 (1998–1999) rows wide by 20 m long. At Pergamino, plant population density was the main-plot, row spacing the subplot and hybrids the sub subplot. At this site, each plot was five rows for row spacings of 0.7 and 1 m, seven rows for row spacing of 0.5 m, and eight rows for row spacing of 0.35 m. All plots were 10 m long. In all experiments, plots were hand-planted at three seeds per hill, and thinned to one plant per hill at the V3 stage of development (Ritchie et al., 1993). Rows always had an east–west orientation to have a similar row orientation effect on light environment beneath the canopy (Sinoquet and Bonhomme, 1992).

To minimize N restrictions, urea fertilizer (N–P–K. 46–0–0, 200 kg N ha–1) was side-dressed and incorporated at V4. Plots were kept free of weeds, insects, and diseases. Water stress was prevented by means of furrow (at Salto) or sprinkler (at Pergamino) irrigation, with the soil near field capacity throughout the growing season. Mean air temperature and solar radiation were recorded daily at each experimental site.

Meteorological conditions differed among years. The 1997–1998 growing season was characterized by intermediate air temperatures ({approx}20.6°C), and the lowest daily irradiance values ({approx}21.4 MJ m–2 d–1) registered during the last 30 yr due to a strong El Niño phase of the El Niño Southern Oscillation phenomenon (ENSO). Contrarily, maize crops cultivated during 1998–1999 experienced higher ({approx}25.8 MJ m–2 d–1) irradiance levels and lower air temperatures ({approx}20.3°C) than the previous season. During 2000–2001 and 2001–2002, irradiance values were intermediate ({approx}24.4 MJ m–2 d–1) between the other growing seasons, but air temperatures were the highest ({approx}22.7°C).

Leaf Area and Light Environment
At V3 five successive plants were tagged in the central row of each plot. Tags were placed between Leaves 3 and 4, which allowed the identification of individual leaves. Tags were moved upward (between Leaves 8 and 9) at V10, and at silking (R1; Ritchie et al., 1993) were placed at the ear leaf. Phenology was recorded weekly on tagged plants. Green leaf area per plant at R1 was measured on tagged plants using a nondestructive method based on lamina length and maximum lamina width (Montgomery, 1911). A leaf was considered senesced when half or more of its area had yellowed. The LAI was calculated as the product of green leaf area per plant and plant population density (Muchow and Carberry, 1989). Ear-leaf number from the base of the plant was recoded at R1 for each tagged plant and LAI above the ear was calculated. Area of senesced leaves was discounted from LAI at R1 to obtain weekly values of LAI until physiological maturity (R6; Ritchie et al., 1993). Leaf area index duration was the integral of the evolution of postsilking LAI values on a thermal time basis using a base temperature of 8°C (Ritchie and NeSmith, 1991).

The fraction of incident PAR intercepted by crops (fPAR) was measured weekly or fortnightly between the onset of the critical period for kernel set (15 d before silking; R1 – 15 d) and R6. The fPAR was calculated from PAR measurements obtained above the canopies and incident light registered between the green and the senesced leaves strata. Measurements were made with a line quantum-sensor of 1 m long (LI-191SA, LI-COR, Lincoln, NE) in E1 and E2, and of 0.8 m long (AccuPAR radiometer; Decagon Devices, Pullman, WA) in E3 and E4. Five independent records were taken within each plot, between 1100 and 1400 h on clear days. The sensor bar was placed diagonally across the rows, to fit it between three (crops at 0.35 m between rows), and two (crops at 0.50 m and 0.70 m between rows) inter-row spaces. Thus, the width under measurement in these plots included two rows of plants in crops grown at 0.35 m, and one row of plants in crops grown at 0.50 m and 0.70 m (Flénet et al., 1996). In crops at 1 m between rows, the fPAR value for the plot was obtained as the average of 10 measurements: five with the sensor bar placed at equidistant positions along the inter-row space and perpendicular to the row, and five with the sensor bar centered and perpendicular to the row. Global solar radiation was converted into PAR by multiplying by 0.45 (Monteith, 1965). Daily fPAR was obtained by interpolation and applied to the corresponding values of PAR to estimate IPAR during the critical period for kernel set (a 30-d period centered at R1, R1 – 15 d to R1 + 15d), and during the effective grain-filling period which starts 15 d after R1(R1 + 15d) and ends at R6.

The light environment beneath maize canopies was characterized 7 to 10 d after R1. Light attenuation within the canopy was estimated from incident PAR and transmitted PAR at two canopy layers: immediately below the ear leaf, and below the lowermost leaf stratum but above senesced leaves, as was previously described. In E1 and E2, measurements of the red/far-red ratio (R/FR) within the canopy were performed at R1 + 15d using a two-channel radiometer (Model SKR 110, Skye Instruments, Powys, UK) with a cosine-corrected head and narrow band filters centered at 660 (red) and 730 nm (far-red). The sensor was connected to a hand-held meter for direct instantaneous readout. The R/FR was measured at canopy layers mentioned above. The sensor was placed at the mid-distance between plants in the row (west side) and in the middle of the inter-row (south side), with its sensing surface facing upward. Six measurements per plot and position were performed between 1200 and 1300 h on clear days.

Crop Growth, Grain Yield, and Grain Yield Components
Crop growth was recorded for the period around silking at E2 and the effective grain-filling period at E2, E3, and E4. For this purpose, shoot biomass samples were taken at R1 – 15d, R1 + 15d, and R6. At each sampling date, 9 plants plot–1 were harvested and oven dried at 70°C. Shoot biomass was expressed per square meter considering the corresponding plant population density. Crop growth was estimated as the difference between shoot biomass per square meter at R1 + 15d and at R1 – 15d for the period around silking, and at R6 and R1 + 15d for the effective grain filling period. Radiation use efficiency was estimated as the ratio between crop growth and the IPAR during mentioned periods.

In all experiments grain yield and grain yield components were estimated from plants sampled at R6. Kernel number was counted and kernel weight was calculated as the quotient between grain yield and kernel number.

Statistical Analyses
In each experiment, differences among treatments were tested by ANOVA and mean comparison were made using least square means. Data were pooled over years in the absence of interactions.

Several correlation analyses among variables were tested for the whole database. The fitting of the models was performed by an optimization technique (Jandel Scientific, 1992). A linear regression was fitted to the relationship between LAI duration and LAI, crop growth and IPAR during the effective grain-filling period, crop growth and RUE during the effective grain-filling period, IPAR per kernel during the effective grain-filling period and IPAR per kernel during the period around silking, grain yield and kernel number, kernel number and IPAR during the period around silking, and kernel weight and postsilking source-sink ratio. The postsilking source-sink ratio was calculated as the quotient between IPAR during the effective grain-filling period and kernel number at E1, E2, E3, and E4, and crop growth during the effective grain-filling period and kernel number at E2, E3, and E4.

To analyze the response of LAI duration to plant population density increases (i.e., at larger LAIs), LAI duration and LAI were represented in relative units (Eq. [2] and [3]).

Formula 2[2]

Formula 3[3]
where subindexes n and 3 pl correspond to individual values and to values at 3 plants m–2, respectively. Data of both groups of hybrids were pooled together and linear regressions were fitted to the relationship between relative LAI duration and LAI. Using this approach, a slope lower than 1 indicates an enhanced postsilking leaf senescence progress in response to increased LAIs. Differences between groups of hybrids for the intercept and the slope of the linear regressions were tested by ANOVA and a t test (Steel and Torrie, 1960).

Comparisons between wide rows and narrow rows for grain yield and grain yield components, the IPAR during the period around silking and the effective grain-filling period and the R/FR were also performed, and linear regressions were fitted.

An inverse function (Andrade et al., 1999) was fitted to the relationship between kernel number and crop growth rate during the period around silking (Eq. [4]).

Formula 4[4]
where CG rate is crop growth rate.

A power function was fitted to the comparison between maximum fPAR values obtained in wide rows and those in narrow rows (Eq. [5])

Formula 5[5]
where fPARWR is maximum fraction of incident PAR intercepted by crops in wide rows and fPARNR is maximum fraction of incident PAR intercepted in narrow rows.

Main effects, interactions and regression equations were considered significant when probability of greater F values were less than or equal to 0.05.


   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Leaf Area Index at Silking and Leaf Area Index Duration
Plant population density increased maximum LAI (Table 1). At silking, the highest LAI values (4.84–6.67) were recorded at plant population densities ≥9 plants m–2. Maize cultivated at plant population densities ≤4.5 plants m–2 never attained a LAI > 4. In contrast, row spacing did not modify maximum LAI values. Plant population density and row spacing interaction effect on maximum LAI was only recorded in E2 (P < 0.05), where crops at 12 plants m–2 in narrow rows exhibited larger LAI values ({approx}6.12) than in wide rows ({approx}5.55). Among tested hybrids, Exp980 (small kernel weight hybrid) and AX889 (large kernel weight hybrid) exhibited the maximum LAI values. Green LAIs above the ear leaf stratum were always <4. Plant population density and hybrid effects on LAI above the ear were similar to those described for maximum LAI. Row spacing increased LAI above the ear 7 to 12% in E1 and E2.


View this table:
[in this window]
[in a new window]
  		Table 1. Plant population density, row spacing and hybrid effects on leaf area index (LAI), leaf area index above the ear leaf stratum, the fraction of incident photosynthetic active radiation (PAR) transmitted at the lowermost leaf stratum and at the ear leaf stratum 10 d after silking, and the postsilking LAI duration of maize hybrids in four experiments (E).

 
Maize cultivated at plant population densities ≥9 plants m–2 exhibited the largest LAI durations (Table 1). In contrast, row spacing did not affect LAI duration. Hence, LAI duration values were positively related to maximum LAIs for both small (LAI duration = 326.4 LAI + 233.2; r2 = 0.95, n = 30) and large (LAI duration = 544.8 LAI + 1230.6; r2 = 0.92, n = 32) kernel weight hybrids. The ordinate and the slope of the linear model fitted to large kernel weight hybrids were larger (P < 0.001) than those obtained for the small kernel weight group. These differences in linear model parameters were related to the larger LAI durations of the former.

Relative LAI durations were also related (r2 = 0.92–0.94) to relative LAIs for both groups of hybrids. The ordinates of the linear model did not differ from zero, and the slopes were significantly (P < 0.05) lower than 1 (0.80 ± 0.04 and 0.58 ± 0.03 for the small and the large kernel weight hybrids, respectively). Hence, for both groups of hybrids postsilking leaf senescence was increased in response to treatments that maximized LAI (i.e., at high plant population densities), but this response was of higher magnitude in the large than in the small kernel weight hybrids.

Light Environment, Crop Growth, and Radiation Use Efficiency
When maximum LAI was attained, fPAR values increased from 0.60 to 0.95 in response to plant population density increase from 3 to 4.5 plants m–2 to 9 to 12 plants m–2 (Fig. 1A ). At plant population densities ≥9 plants m–2, the fraction of incident PAR transmitted at the ear and lowermost leaf strata were less than 27 and 11%, respectively (Table 1). In contrast, at low plant population densities, the fractions of incident PAR transmitted at the ear and lower most leaves were 50 and 37%, respectively. Hybrids with the greatest LAIs exhibited the lowest proportion of incident PAR transmitted at any leaf stratum. Row spacing reduction increased fPAR from 0.90 to 0.93 in E3. A curvilinear function adequately described the relationship between maximum fPAR in wide rows vs. narrow rows for the whole data set (Fig. 1A). Most data points were positioned below the first bisecting line (21 out of 31) suggesting an improvement of fPAR in narrow rows, especially for fPARs ≤ 0.90 in wide rows. Thus, row spacing reduction determined a greater (P < 0.05) enhancement of fPARs at 3 to 4.5 plants m–2 ({approx}8.6% ± 1.6) than at 9 to 12 plants m–2 ({approx}1.8% ± 1.2).


Figure 1
View larger version (20K):
[in this window]
[in a new window]
  		Fig. 1. (A) Maximum fraction of incident photosynthetic active radiation intercepted (fPAR) by crops cultivated in wide rows (WR) vs. those in narrow rows (NR); and (B) red:far-red ratio (R/FR) reaching the ear leaf layer and the lowermost green leaf layer of crops in WR vs. those in NR. Data points represent the mean values of each treatment (plant population x row spacing x hybrid) in all experiments (Fig. 1A), and in E1 and E2 (Fig. 1B). Black symbols, hybrids of small kernel weight; gray symbols, hybrids of large kernel weight. The solid lines indicate the model fitted to the data set. The dashed line indicates the 1:1 relationship between variables.

 
Plant population and row spacing also affected the R/FR perceived by plants at the ear leaf stratum and below the lowermost green leaf stratum (Fig. 1B). There was a consistent reduction in the R/FR whenever plant population was increased and row spacing was reduced. A linear function adequately described the relationship between R/FR in wide rows vs. narrow rows. The ordinate of the linear regression differed from 0, and the slope was not different from 1. Hence, after silking leaf strata of crops in wide rows perceived higher R/FRs ({approx}7, 14.3, and 45% for 3, 9, and 12 plants m–2, respectively) than those in narrow rows.

Daily IPAR values accumulated over the period around silking were greatest at plant population densities ≥9 plants m–2 (Table 2). Differences among experiments were related to differences in irradiance values among years, and differences among hybrids were related to phenotypic differences in canopy size. Crops under narrow rows in E3 and E4 exhibited greater IPARs around silking than those under wide rows. The IPARs during the period around silking of crops in wide rows were positively related to those at the same plant population densities but in narrow rows (Fig. 2A ). A linear regression satisfactorily depicted the relationship between variables. The slope value was not different from 1, and the origin did not differ from zero, but almost all data points (27 out of 31 points) fell below the first bisecting line. Hence, for these treatments IPAR around silking of crops in wide rows was lower ({approx}8 and 4% for 3–4.5 and 9–12 plants m–2, respectively) than for those in narrow rows.


View this table:
[in this window]
[in a new window]
  		Table 2. Plant population density, row spacing and hybrid effects on accumulated intercepted photosynthetic active radiation (IPAR) and IPAR per kernel during the period around silking, and IPAR, crop growth (CG) and radiation use efficiency (RUE) during the effective grain-filling period of maize hybrids in four experiments (E).

 

Figure 2
View larger version (32K):
[in this window]
[in a new window]
  		Fig. 2. (A) Amount of photosynthetic active radiation intercepted (IPAR) during the period around silking, and (B) during the effective grain filling period by crops cultivated in wide rows (WR) vs. those in narrow rows (NR). Data points represent the mean values of each treatment (plant population x row spacing x hybrid) in all experiments. Black symbols, hybrids of small kernel weight; gray symbols, hybrids of large kernel weight. The solid lines indicate the model fitted to the data set. The dashed line indicates the 1:1 relationship between variables.

 
Crop growth during the period around silking of the small kernel weight hybrids was measured in E2. Crop growth response to treatments was similar to that described for IPAR around silking. Crop growth was maximized (P < 0.01) at population densities ≥9 plants m–2 with a crop growth rate of 17.9, 26.2 and 32.8 g m–2 d–1 for 3, 9, and 12 plants m–2, respectively. The RUE around silking was affected (P < 0.05) by crowding, and the highest RUE values corresponded to crops at 12 plants m–2 (2.46, 2.50, and 3.20 g MJ–1 at 3, 9, and 12 plants m–2, respectively). Crops cultivated in narrow and wide rows exhibited similar crop growth rate and RUE during this period. The hybrid with the highest IPAR (Exp980) did not exhibit the highest crop growth rate (data not showed), because it had the lowest (P < 0.01) RUE value ({approx}2.94, 2.98, and 2.24 g MJ–1 for DK696, DK757, and Exp980, respectively).

The IPAR during the effective grain-filling period was affected by plant population density and row spacing (Table 2). In all experiments except E4, maximum IPAR values during this period were recorded in crops cultivated at plant population densities ≥9 plants m–2. In contrast, in E4 the highest IPAR was measured at the lowest plant density. In E3 and E4, crops in narrow rows presented higher IPARs than those in wide rows. In E2, however, the largest IPAR value was measured in wide rows. When IPARs in wide rows were plotted as a function of those in narrow rows (Fig. 2B), a positive relationship between variables was observed. The origin of the linear regression differed from zero and the slope value was lower than 1. Thus, light capture improvement in narrow rows ({approx}4%) was mainly registered for cropping conditions that maximized IPAR during the effective grain-filling period, such as hybrids with longer LAI duration (the large kernel weight hybrids).

For both groups of hybrids, IPAR during the effective grain-filling period explained 70% of the variation in crop growth during this stage (Fig. 3A ). Crop growth, however, did not exhibit the same response to treatments as IPAR (Table 2). At low plant population densities in E2 and E3, crops had a higher postsilking RUE than at high densities (Table 2). Radiation use efficiency was also depressed in response to row spacing reduction (E3 and E4). Thus, postsilking RUE variations explained 88% (small kernel weight hybrids) and 94% (large kernel weight hybrids) of crop growth variability across stand densities and row spacings (Fig. 3B). The origin of fitted functions did not differ between groups of hybrids but the slope of the large kernel weight hybrids was greater than that of the small kernel weight group. These results indicate that differences between groups of hybrids in crop growth during the effective grain-filling period were mainly related to IPAR, but differences in crop growth among plant population densities and row spacings within each group were mainly related to postsilking RUE.


Figure 3
View larger version (34K):
[in this window]
[in a new window]
  		Fig. 3. Crop growth (CG) during the effective grain-filling period as a function of (A) the amount of intercepted photosynthetic active radiation (IPAR), and (B) radiation use efficiency (RUE) during this stage. Data points represent the mean values of each treatment (plant population x row spacing x hybrid) in E2, E3, and E4. Black symbols, hybrids of small kernel weight (full symbols crops in narrow rows, NR; empty symbols crops in wide rows, WR); gray symbols, hybrids of large kernel weight (dark gray crops in NR, light gray crops in WR). The lines indicate the model fitted to the data set.

 
Grain Yield and Grain Yield Components
An increase in plant population density increased grain yield and kernel numbers but reduced kernel weight (Table 3). The highest grain yields and kernel numbers were recorded at plant population densities ≥9 plants m–2 while the largest kernel weights were obtained at low plant densities (3–4.5 plants m–2). Maize hybrids exhibited a similar kernel number in wide and narrow rows. Similarly, grain yield was not affected by row spacing except in E1, where higher grain yield in wide rows than in narrow rows was determined by the larger kernel size. Plant population density and row spacing interactions on grain yield and grain yield components were not detected, and positive relationships between row spacings (wide rows vs. narrow rows) were determined for grain yield (r = 0.82, P < 0.001), kernel number (r = 0.93, P < 0.001) and kernel weight (r = 0.98, P < 0.001) (data not presented). For grain yield and kernel number, these relationships did not differ from the 1:1 ratio. In contrast, for kernel weight the ordinate (29 ± 9.7) of the linear regression differed from 0, and the slope (0.91 ± 0.04) was lower than 1, indicating that kernel size of crops in wide rows was less affected by light stress (i.e., high plant population density) than kernel size of crops in narrow rows.


View this table:
[in this window]
[in a new window]
  		Table 3. Plant population density, row spacing and hybrid effects on grain yield (in dry weight basis), grain yield components, and source–sink ratio during the effective grain-filling period. The source was quantified as: (i) the amount of PAR intercepted (IPAR) per square meter, and crop growth (CG) during the effective grain-filling period, and the sink size as kernel number per square meter. Maize crops were cultivated in four experiments (E).

 
In each experiment, hybrids cultivated at the same plant density generally had a similar grain yield (except in E4) due to compensations between grain yield components. For each group of hybrids, grain yield was related to kernel number (r2 = 0.80 and 0.50 for small and large kernel weight hybrids, respectively) and the relationship between the mentioned traits was adequately described by a linear regression. The ordinate of the large kernel weight hybrids (571.6 ± 95.4) was greater than that of the small kernel weight group (321.3 ± 67.1), but the slopes yielded similar values ({approx}0.15). These results suggest that for the range of kernel number of both groups of hybrids (2569–7460 and 2235–4457 kernels m–2 for the small and large kernel weight hybrids, respectively) the large kernel weight hybrids attained higher grain yields than the small kernel weight group.

Differences in relative kernel number (i.e., excluding the genotypic effect) were related to relative IPAR around silking (i.e., excluding year effects) (Fig. 4A ). When crop growth around silking was measured, variations of this trait explained 58% of kernel number variability of the small kernel weight hybrids (Fig. 4B).


Figure 4
View larger version (31K):
[in this window]
[in a new window]
  		Fig. 4. (A) Relative kernel number per unit land area as a function of the relative amount of intercepted photosynthetic active radiation (IPAR) during the period around silking. Kernel number of each treatment was expressed as a proportion of the maximum kernel number recorded in each experiment (7460, 6053, 4181, and 4557 kernels m–2 for E1, E2, E3, and E4, respectively). The IPAR was also expressed as a proportion of the maximum IPAR registered in each experiment (218.6, 317.2, 343.7, and 357.6 MJ m–2 for E1, E2, E3, and E4, respectively). (B) Kernel number per square meter as a function of crop growth (CG) rate during the period around silking in E2. Data points represent the mean values of each treatment (plant population x row spacing x hybrid). Symbols as in Fig. 3. The solid lines indicate the model fitted to the data set.

 
Kernel weight variability of the whole data set was accounted for by IPAR per kernel during the effective grain-filling period (Fig. 5A ). Interestingly, IPAR per kernel variability during grain filling was related to IPAR per kernel during the period around silking (IPAR kernel–1 during grain filling period = 1.7 IPAR kernel–1 around silking – 0.05; r2 = 0.90; Tables 2 and 3). These results suggest that postsilking source activity, i.e., LAI duration, was regulated by sink capacity determined around silking. Hence, kernel weight variability could also be explained by IPAR per kernel during the period around silking (Kernel weight = 2046.2 IPAR kernel–1 around silking + 116; r2 = 0.75). Nevertheless, there were differences in kernel weight that could not be attributed to IPAR per kernel, neither during the period around silking nor during the effective grain-filling period (Tables 2 and 3). For these treatments, differences in RUE were recorded during the mentioned periods in spite of similar IPAR per kernel values. For these situations, therefore, variations in crop growth related to differences in RUE (Fig. 3B) should be attributed primarily to source activity per se and not to light capture (Fig. 3A). Consequently, kernel weight variability of all experiments (except E1, where crop growth was not measured) was better accounted for by crop growth per kernel (Fig. 5B) than by IPAR per kernel during the effective grain-filling period (Fig. 5A).


Figure 5
View larger version (34K):
[in this window]
[in a new window]
  		Fig. 5. Kernel weight as a function of the amount of photosynthetic active radiation intercepted (IPAR) per kernel (A) and crop growth (CG) per kernel (B), during the effective grain-filling period. Data points represent the mean values of each treatment (plant population x row spacing x hybrid) in all experiments (Fig. 5A), and in E2, E3, and E4 (Fig. 5B). Symbols as in Fig. 3. The solid lines indicate the models fitted to the data set. The dashed line indicates the 1:1 relationship between variables.

 

   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Maize grain yield was stable in response to changes in plant spatial arrangement at all plant population densities. This result rejects the hypothesis of our work, which predicted a higher grain yield in narrow rows than in wide rows at low plant densities, but a lower grain yield in narrow rows than in wide rows at high plant densities. This hypothesis was based on previous evidences of an improved light interception by crops in narrow rows when canopy size did not attain the critical LAI value (Andrade et al., 2002; Barbieri et al., 2000; Maddonni et al., 2001a) and speculations of a reduced postsilking IPAR and/or RUE at high plant densities in narrow rows. Row spacing effects on IPAR around silking and postsilking RUE were verified, but grain yield components did not vary as was hypothesized.

Row spacing did not modify maximum LAI (Maddonni et al., 2001a; Westgate et al., 1997) but for most cropping conditions a more square planting pattern improved light interception at silking, especially at low plant densities. For a few data points a similar fPAR was recorded at any row spacing. These dissimilar fPAR responses to row spacing can be attributed to the contrasting canopy sensitivity to plant spatial distribution detected in maize hybrids (Maddonni et al., 2001a), and its effect on light interception at silking (Maddonni et al., 2001b). In the mentioned works, maize hybrids were classified as "plastic" or "rigid" based on the ability of their leaves to undergo spatial reorientation. In "plastic genotypes," plant reaction to empty spaces was a reorientation of expanding leaves in the horizontal plane (i.e., leaf azimuth distribution), to fill the gaps (e.g., inter-row or intra-row space). This reaction allowed for a similar light interception value in different inter-row distances. In contrast, "rigid genotypes" did not exhibit this reaction capacity, and presented a random leaf azimuth distribution independently of planting pattern. For this type of hybrids there was a light interception reduction when grown under a rectangular planting pattern. In the present work the different canopy sensitivity to plant spatial distribution of hybrids was not quantified.

Despite possible differences in canopy behavior among tested hybrids, crops in this study at low plant densities exhibited similar kernel numbers. At 3 plants m–2, crop growth rate around silking was 18 g m–2 d–1 with a mean plant growth rate of 6 g plants d–1. This plant growth rate was greater than threshold values of 3 to 4 g plant–1 d–1 as reported by Echarte et al. (2004) above which plants yield their potential kernel number. Hence, plant spatial arrangement effects on IPAR around silking were not reflected in variations of kernel number per square meter at low plant densities, due to morphogenetic limitations in the reproductive organs of plants (almost all flowers ear–1 yield a kernel). Grain yield per square meter followed a similar trend as kernel number per square meter, because kernel weight generally did not vary among row spacings. Thus, the interesting approach proposed by Andrade et al. (2002) for the analysis of grain yield per square meter responses to narrow rows based on the improvement of light interception at silking could not be extended to cropping conditions where grain yield per plant is maximized but crop grain yield is sink limited, such as crops cultivated at low stand densities without resource restrictions. This crop husbandry system is unusual in present maize production systems of humid regions, but may represent a constraint in water-limited environments (Du Toit and Prinsloo, 2000).

Optimal crop management should aim to maximize light interception and RUE during the critical stages of grain yield definition (Andrade, 1995). At high plant population densities, crops in wide rows attained fPAR values larger than 0.95. Hence, row spacing reduction in this condition did not contribute to increase IPAR around silking or during the postsilking period. Interestingly, postsilking RUE was reduced in response to increased crowding (18–52% reduction) and to narrow rows (5–16% reduction). Consequently, the effect of plant population density on postsilking RUE was opposite to that recorded for the period around silking (25% lower at 3 plants m–2 than at 9–12 plants m–2). A reduced RUE around silking (10–15%) or during the presilking period (20–46%) in response to low stand densities has been previously documented (Andrade et al., 1993b; Kiniry et al., 1989), but to our knowledge there is no report of reductions in postsilking RUE in response to crowding and to narrow rows. Discrepancies about plant population density effects on RUE seem to arise from the period used for computing this effect. During the presilking period, plants at low stand densities were likely sink limited, and high carbohydrate levels in shoots at anthesis may have depressed photosynthesis rate and RUE (Nafziger and Koller, 1976; Prioul, 1995). In contrast, during the postsilking period, photosynthesis production of plants at high stand densities may have declined due to both the enhanced light attenuation within the canopy and the fast leaf senescence progress. Low leaf strata of crops in narrow rows perceived reduced irradiance values and R/FRs than those in wide rows, which could have depressed photosynthesis activity and RUE, without modifying LAI duration. Hence, the higher postsilking RUE of crops cultivated at low stand densities overcompensated their lower IPAR during this period. Similarly, the higher postsilking RUE of crops in wide rows was reflected in a higher crop growth. This apparent strong relationship between postsilking crop growth and RUE rather than IPAR has never been previously described, but is substantiated from results of a previous study (Otegui et al., 1995) where less than 30% of the postsilking shoot growth variability of maize hybrids cultivated at commercial plant populations was explained by IPAR variations during that period.

Finally, present knowledge on kernel weight determination indicates that kernel growth during the effective grain-filling period of most maize crops is not source-limited (Borrás et al., 2004), final kernel weight is determined by variations in grain filling rate during the effective grain-filling period rather than by variations in grain filling duration (Borrás and Otegui, 2001; Cirilo and Andrade, 1996), and differences in this rate are determined by the source-sink ratio established during the critical period for kernel set and not by variations in this ratio during the effective grain-filling period (Gambín et al., 2006). In this conceptual framework, source activity during grain filling seems to be controlled by sink activity during this stage, and this sink capacity depends on the source-sink ratio set around silking. Lack of source limitations during active grain filling is evident in the almost nil response of kernel weight to improved source conditions (e.g., kernel number reductions) during this stage (Borrás et al., 2004). On the other hand, maize kernel weight will be severely affected by reductions in source availability that take place during this period (e.g., by abiotic stress, hail storms, defoliating insects, leaf rust). These concepts are supported by evidences from studies where leaf senescence is promoted by sink manipulations (Ceppi et al., 1987). In agreement with these evidences, we determined that final kernel weight (sink capacity) was strongly related to variations in crop growth per kernel during the effective grain-filling period (source activity), and that changes in crop growth responded to accumulated IPAR during this stage, a relationship mainly driven by differences in LAI duration of hybrids. Variations in these indicators of source activity during active grain filling were related to the amount of IPAR per kernel accumulated around silking, a good surrogate of source availability per kernel during this period. However, an important part of the variation in kernel weight did not fit within this analysis. We determined significant differences in kernel weight at constant stand densities but contrasting row spacings, and among different stand densities with similar row spacings, which could not be attributed to variations in IPAR per kernel at any stage. Differences among these treatments in crop growth during active grain filling were related to variations in postsilking RUE, attributable to the source activity per se rather than to light capture. For these treatments, therefore, source activity during grain filling was not entirely under sink control, and seemed to depend on variations in R/FR related to plant spatial arrangement.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Maize was traditionally cultivated in wide spaced rows, but a trend to reduce inter-row distance was common during the last decade. Our results revealed that under temperate environments without water and nutrient limitations, maize grain yield did not respond to row spacing reduction at 3 to 12 plants m–2. At low plant population (3–4.5 plants m–2), morphogenetic limitations to final kernel number per plant (i.e., number of florets per ear) suppressed any possible effect of enhanced light interception around silking promoted by a more square planting pattern. At commercial and high plant populations (9–12 plants m–2), canopy sizes were always above the critical LAI value, and maximum levels of light interception at silking were recorded at all row spacings. Hence, under temperate environments without water and nutrient restrictions, no benefits could be expected in terms of grain yield by reducing row spacing from the present 0.7- to 0.8-m inter-row distance. On the other hand, present breeding systems are shifting toward selection environments under narrow rows (0.52 m) and high plant densities (e.g., see web site of Nidera S.A. www\nidera.com.ar/), which in the near future may deliver hybrids with improved performance under these cropping conditions that promote an unfavorable environment for source activity during grain filling.


   ACKNOWLEDGMENTS
 
Authors wish to thank Dekalb-Monsanto S.A. and Nidera Semillas Argentina S.A. for providing the seeds and E. Whitechurch for the revision of English style. G.A. Maddonni and M.E. Otegui are members of the National Council for Research (CONICET).


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
G.A. Maddonni and M.E. Otegui are members of CONICET.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Agron. J.Home page
P. A. Barbieri, H. E. Echeverria, H. R. Sainz Rozas, and F. H. Andrade
Nitrogen Use Efficiency in Maize as Affected by Nitrogen Availability and Row Spacing
Agron. J., June 23, 2008; 100(4): 1094 - 1100.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager