Published online 17 November 2005
Published in Agron J 97:1622-1625 (2005)
DOI: 10.2134/agronj2005.0123
© 2005 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
Production Papers
Rice Yield Components as Affected by Cultivar and Seeding Rate
Brian V. Ottisa,* and
Ronald E. Talbertb
a Div. of Plant Sciences, Univ. of Missouri-Columbia, P.O. Box 160, Portageville, MO 63873
b Dep. of Crop, Soil, and Environmental Sciences, Univ. of Arkansas, Fayetteville, AR 72704
* Corresponding author (ottisbv{at}missouri.edu)
Received for publication April 26, 2005.
 |
ABSTRACT
|
|---|
New rice (Oryza sativa L.) cultivars have been released that possess yield potential >10000 kg ha–1. Some of these new cultivars have increased costs associated with them due to patented input traits and/or hybrid technology. Research analyzing the effects of seeding rate on yield components with modern cultivars is not well-documented. Therefore, field trials were conducted near Stuttgart, AR, in 2002 through 2004 with the objective of determining the effect of rice seeding rate on yield components of three modern, long-grain rice cultivars. This research was done in an effort to determine if lower-than-recommended seeding rates would produce yields similar to currently recommended rates. Rice seeding rates from 57 to 500 seeds m–2 resulting in a rice density range of 73 to 373 plants m–2 did not effect rice aboveground biomass production, panicle density, harvest index (HI), or rice yield, regardless of cultivar. ‘Wells’ produced higher panicle weights and had a higher HI than ‘CL161’ across the range of rice densities. Cultivars XL8 and Wells produced similar yields, and these yields were higher than CL161. Cultivar, rice density, and thermal time were significant factors affecting rice canopy coverage. Cultivar XL8 achieved canopy coverage sooner than CL161 or Wells. As rice density increased, canopy coverage increased by 3% for every additional 100 plants m–2. As degree days (DD50) accumulated, canopy coverage increased 0.4% °Cd–1. Results from this study indicate that recommended seeding rates for CL161, Wells, and XL8 can be reduced while maintaining similar yields.
Abbreviations: DD50, degree day 50 • HI, harvest index • PAR, photosynthetically active radiation
|
INTRODUCTION
|
|---|
OVER THE YEARS, rice yields have increased due to advances in breeding and crop management. Several new rice cultivars have recently become available for production, and many of these cultivars have exceptional yield potential (Moldenhauer et al., 2001). Main-crop plus ratoon-crop yields with hybrid rice have exceeded 12990 kg ha–1 in Texas; however, hybrid rice milling quality has been inferior to that of conventional and semidwarf cultivars in the past (G.N. McCauley, personal communication, 2002) More recent hybrid rice introductions have shown milling quality equal to that of modern conventional cultivars (Wilson et al., 2004).
Researchers recommend a planting rate of 431 seeds m–2 for conventional cultivars and suggest the optimum rice density at emergence is 161 to 215 plants m–2 (Saichuk et al., 2002; Klosterboer and Turner, 2001; Wilson et al., 2005) for conventional cultivars and 86 to 108 plants m–2 for hybrids (Anonymous, 2005). Essentially, growers are recommended to plant twice as many seeds as needed for optimum rice density, and an even greater rate if planting in a clay soil or under imperfect seedbed conditions (Wilson et al., 2005).
Research using older cultivars has illustrated the effects of seeding rate on rice yields. Wells and Faw (1978) found no differences in rice yields among seeding rates ranging from 67 to 303 kg ha–1 at low N rates, but the 67 kg ha–1 seeding rate produced higher yields at high N rates. Furthermore, Gravois and Helms (1996) reported that ‘Millie’ and ‘Adair’ rough rice yields decreased as rice density increased in a quadratic fashion above rice densities of 331 and 337 plants m–2, respectively; however, ‘Kaybonnet’ rough rice yields were not affected by seeding rates ranging from 161 to 867 seeds m–2. They also determined that ‘Katy’ and Kaybonnet head rice yields increased linearly as seeding rate decreased. These researchers and others have reported that as rice seeding rates increased, panicle density increased and filled grains per panicle decreased with no changes in yield (Gravois and Helms, 1992; Jones and Snyder, 1987). These findings indicated that certain rice cultivars will compensate for lower rice densities by producing more reproductive tillers to fill voids in the canopy.
Recently, digital imagery has been used to estimate canopy coverage in agricultural crops. Purcell (2000) used digital imagery to measure canopy coverage differences among soybean [Glycine max (L.) Merr.] maturity groups at different planting rates. Wheat (Triticum aestivum L.) senescence rates have also been measured using digital imagery (Adamsen et al., 1999). Richardson et al. (2001) used digital imagery to estimate cover and rate of spread for various turfgrass species. Taking advantage of this technology to estimate canopy coverage for rice cultivars at various densities should provide another estimate of the compensatory ability of rice not previously found in the literature.
Unlike other major agronomic crops grown in the USA, rice cultivar development has traditionally been accomplished through public breeding programs. With the recent development and commercialization of herbicide-resistant (CLEARFIELD*) conventional and hybrid rice, the emergence of private rice breeding programs and patented input traits has led to increases in seed costs of some cultivars. These developments have led rice researchers to revisit traditional seeding rate recommendations and evaluate lower seeding rates in an effort to save input costs for rice producers. Therefore, our objective for this research was to examine the effect of rice seeding rate on yield components of three modern rice cultivars in an effort to determine if optimum yield could be attained at lower-than-recommended seeding rates.
|
MATERIALS AND METHODS
|
|---|
A field study was established during the summers of 2002 through 2004 at the Rice Research and Extension Center near Stuttgart, AR. The experiment was designed as a randomized complete block with four replications. A factorial treatment arrangement was used with factors consisting of three rice cultivars (CL161, Wells, and XL8) and four seeding rates in an effort to establish rice densities of 52, 104, 208, and 416 plants m–2. Plot size was 1.8 by 4.9 m. The soil type was a DeWitt silt loam (Typic Albaqualf) with a pH of 5.8 and a soil organic matter content of 0.94%.
The three rice cultivars used in the experiment represented the three major types of long-grain rice grown in the USA. Cultivar CL161 represented semidwarf rice, Wells represented conventional-height rice, and XL8 represented hybrid rice. Rice was sown on 14 May in 2002, 13 June in 2003, and 22 May in 2004. The late planting date in 2003 was the result of poor seed germination from the original planting on 20 May; therefore, the study was abandoned and replanted 13 June. Seeding rates were established by counting and weighing 1000 seeds, then adding the amount of seed to the planter to achieve the aforementioned seeding rate based on seed weight and the minimum germination percentage for each cultivar. Minimum germination for CL161 and Wells was 80% whereas minimum germination for XL8 was 90% as indicated on the seed containers. Therefore, 20% was added to the CL161 and Wells seeding rates, and 10% was added to the XL8 seeding rate. Seeding rates on a kg ha–1 basis that were utilized are shown in Table 1. Rice was drill-seeded with a nine-row cone planter set on 19-cm row spacings. Weed control was maintained with a preemergence application of clomazone {2-[(2-chlorophenyl)methyl]-4,4-dimethyl-3-isoxazolidinone} at 0.33 kg ha–1 + quinclorac [3,7-dichloro-8-quinolinecarboxylic acid] at 0.25 kg ha–1 followed by (fb) quinclorac at 0.25 kg ha–1 to four- to five-leaf rice fb an application of triclopyr {[(3,5,6-trichloro-2-pyridinyl)oxy]acetic acid} at 0.28 kg ha–1 + halosulfuron {3-chloro-5-[[[[(4,6-dimethoxy-2-pyrimidinyl)amino]carbonyl]amino]sulfonyl]-1-methyl-1H-pyrazole-4-carboxylic acid} at 0.05 kg ha–1 + crop oil concentrate at 1% V/V 1 wk following flood establishment.
Nitrogen fertilizer was applied as urea preflood at a rate of 100 kg N ha–1 followed by an application of urea at 67 kg N ha–1 to CL161 and Wells at panicle differentiation. Cultivar XL8 did not receive a panicle differentiation N application, but was fertilized instead with 67 kg N ha–1 at panicle emergence.
Digital images taken 1 m above the canopy of plots were collected on a weekly basis in 2003 and 2004 beginning at emergence and continued for 8 wk. Images were analyzed using software (SigmaScan Pro 5.0; SigmaScan Pro, 2003) to determine the canopy coverage at each interval. The software counted the number of green pixels in a frame in relation to nongreen pixels to estimate canopy coverage. Canopy coverage for the cultivars was evaluated as a function of rice density and heat unit accumulation. Heat units, or degree days (DD50), were calculated with the equation:
where °Cd was equal to degree days, °Cmax was equal to the daily maximum temperature, and °Cmin was equal to the daily minimum temperature. If °Cmax exceeded 34.4°C, then °Cmax was set at 34.4, whereas if °Cmin was less than 21.1°C, then °Cmin was set at 21.1.
Rice stand counts were taken 2 wk after emergence by counting the number of plants in 1 m of row in the center of each plot. Stand counts were recorded as rice density and calculated on a plant m–2 basis. Flags were placed at the front and back of the counted area. Once rice matured, plants from the 1-m section of row were excised at the soil surface, placed in paper bags, and dried in an oven for 60 h at a temperature of 60°C. Then, the plants from the sample area were weighed and total aboveground biomass was recorded. Panicles were counted to determine panicle density. The panicle density was divided by stand counts recorded early in the season to determine panicles plant–1. Panicles were threshed to separate seed. The separated seed were weighed, and the resulting seed weight was divided by the panicle density to determine panicle weight. Harvest index (HI) was then calculated using the following equation:
Grain yield was collected by harvesting the center four rows with a small plot combine. Yield from the 1-m row samples was added to the plot weights for yield calculations. In 2002, severe lodging occurred due to high wind and rain as a result of tropical storms Isidore and Lili, which impacted the study in late September and early October, respectively. Therefore, yield data in 2002 were calculated from seed collected from the 1-m row samples. Yield samples in all 3 yr were weighed and moisture was adjusted to 120 g kg–1.
Data were analyzed using PROC GLM (SAS Inst., 2001). Data were analyzed as a factorial using ANOVA, and means were separated using Fisher's protected LSD at the 5% level of significance. All data were initially analyzed by year and variance among years was homogenous; therefore, data presented are combined across the 3 yr of the experiment. The fixed effects of cultivar and seeding rate were analyzed for their effects on rice density, yield, and yield components. Regression analysis was conducted on the effects of cultivar and rice density on canopy coverage over DD50 accumulation. Means were separated using Fisher's protected LSD at the 5% level of significance.
|
RESULTS
|
|---|
Analysis of the effects of rice cultivar and seeding rate on yield and yield components indicated no significant cultivar x seeding rate interactions during the 3 yr of the study (Table 2). Actual rice densities resulting from the lowest seeding rate were 40% higher than expected; however, densities resulting from the highest seeding rate were 12% lower than expected from the target density. A late planting date combined with less intraspecific competition at low seeding rates most likely caused higher seed germination than expected based on minimum germination percentages (80% for CL161 and Wells; 90% for XL8) indicated on the seed containers. However, at the highest seeding rate, intraspecific competition likely resulted in lower-than-expected rice densities.
View this table:
[in this window]
[in a new window]
|
Table 2. Main effects of rice cultivar and seeding rate on rice density, aboveground biomass production, panicle density, panicle weight, harvest index, and yield near Stuttgart, AR, 2002 through 2004.
|
|
Aboveground biomass production and panicle density was stable as rice density increased, providing evidence for the compensatory nature of rice to fill voids in the canopy by producing larger amounts of biomass (Table 2). For all cultivars, a decreasing trend in panicles plant–1 was observed as seeding rate increased, indicating that each of these cultivars responded to low densities by producing more reproductive tillers. Similarly, panicle weight decreased as seeding rate increased for all cultivars. Harvest index and yield remained constant across the four seeding rates, with seeding rate having no effect.
Cultivar, rice density, and thermal time were significant factors affecting rice canopy coverage (Table 3). Cultivar XL8 achieved canopy coverage sooner than CL161 or Wells as indicated by the less negative intercept coefficient for XL8; however, the rate of canopy coverage as a function of rice density and thermal time was similar for all cultivars. Canopy coverage was most likely attained in XL8 plots sooner due to its propensity to tiller early in the season. As rice density increased, canopy coverage increased by 3% for every additional 100 plants m–2. As thermal time increased, canopy coverage increased 0.4% °Cd–1 for each cultivar.
View this table:
[in this window]
[in a new window]
|
Table 3. Regression coefficients for the effect of rice density and DD50 accumulation on canopy coverage of CL161, Wells, and XL8 rice near Stuttgart, AR, 2003 and 2004.
|
|
|
DISCUSSION
|
|---|
Similar to previous findings, this research found that yield was not decreased with decreased seeding rates (Gravois and Helms, 1992; Jones and Snyder, 1987). In contrast to work done by Counce (1987) using an older cultivar, results from this study did not indicate significant yield losses at rice densities ranging from 79 to 150 plants m–2. Counce (1987) concluded that reductions in rice yield at excessive rice densities appeared to be associated with population-dependent stresses such as water deficits, disease, and nutrient deficiencies. Furthermore, Gooding et al. (2002) reported that grain yield of wheat crops under suitable growing conditions was largely a function of total aboveground biomass production. Results from this study found that rice aboveground biomass production did not decrease with decreasing seeding rate, providing further evidence that rice compensates for voids in the canopy by producing more reproductive tillers at low densities.
Gravois and Helms (1992) reported that optimum rice yield could not be attained without optimum panicle density of uniform maturity. A high correlation (R2 > 0.85) was found in our experiment between yield and panicle density, which supports their findings (data not shown). Our results were in contrast to the findings of Wells and Faw (1978), who found that initial rice plant population, rather than tillering, accounted for the number of panicles per unit area. We did not detect a rice density effect on panicle density (Table 2), indicating that optimum panicle density was achieved from rice densities ranging from 73 to 373 plants m–2.
Contrary to findings by Gooding et al. (2002) in wheat, CL161 and Wells HI in our experiment were unchanged with decreasing seeding rate. Our results suggest that a lower rice density will allow photosynthetically active radiation (PAR) to be more efficiently intercepted by the canopy due to the reduction of mutual leaf shading, which allows for more efficient photosynthate production. Similarly, Wells and Faw (1978) reported that rice yield was limited under dense populations due to reduced light interception and CO2 accumulation. Further evidence of this phenomenon is supported here by the stability of aboveground biomass with decreasing rice density. Since newer cultivars are bred to partition photosynthate more efficiently to the reproductive portions of the plant, a lower rice density is compensated for by the production of more reproductive tillers rather than vegetative biomass. Hence, the ratio of grain to aboveground biomass is maintained.
Based on canopy coverage results from this study, XL8 achieved canopy coverage sooner than CL161 and Wells; however, all cultivars established canopy coverage at the same rate as a function of rice density and thermal time. A study evaluating cultivar effects on red rice found that XL8 reduced red rice seed production compared with several conventional cultivars (Ottis et al., unpublished data, 2005). These findings may have been related to the ability of XL8 to achieve canopy coverage sooner than other cultivars, probably due to its hybrid vigor and early tillering.
Rice producers continue to use seeding rates exceeding the necessary level for optimum yield based on current recommendations and to compensate for unforeseen biotic and abiotic factors. With newer cultivars, it should be possible to lower seeding rates and maintain high yield levels. Currently, the recommended seeding rate is 431 seeds m–2 for CL161 and Wells (Wilson et al., 2005) and 151 seeds m–2 for XL8 (Anonymous, 2005). Wilson et al. (2005) recommend increasing the seeding rates above 431 seeds m–2 if soil conditions are not optimum. Based on our findings at the highest seeding rate, rice density was below expected levels, indicating that some seeds did not emerge. Poor emergence at this seeding rate may have been partially due to intraspecific competition among neighboring plants, which reduced emergence, possibly indicating that seeding rates at this rate or greater may reduce emergence efficiency. Better-than expected emergence at the lowest seeding rate was probably due to a reduction in intraspecific competition along with warm soil conditions, which allowed for more efficient seedling emergence. Results from this study indicate that seeding rates for CL161, Wells, and XL8 can be reduced while still maintaining yields similar to those at currently recommended seeding rates.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank Andrew Ellis, Danny Boothe, Jamie Branson, and Jon Wright for their assistance with this research. This research was funded by the Arkansas Rice Research and Promotion Board.
|
REFERENCES
|
|---|
- Adamsen, F.J., P.J. Pinter, and E.M. Barnes. 1999. Measuring wheat senescence with a digital camera. Crop Sci. 39:719–724.[Abstract/Free Full Text]
- Anonymous. 2005. Planting recommendations. Available at www.ricetec.com/recomAdobe.pdf (accessed 18 Jan. 2005; verified 22 Aug. 2005). RiceTec, Alvin, TX.
- Counce, P.A. 1987. Asymptotic and parabolic yield and linear nutrient content responses to rice population density. Agron. J. 79:864–869.[Abstract/Free Full Text]
- Gooding, M.J., A. Pinyosinwat, and R.H. Ellis. 2002. Responses of wheat grain yield and quality to seed rate. J. Agric. Sci. 138:317–331.
- Gravois, K.A., and R.S. Helms. 1992. Path analysis of rice yield and yield components as affected by seeding rate. Agron. J. 84:1–4.[Abstract/Free Full Text]
- Gravois, K.A., and R.S. Helms. 1996. Seeding rate effects on rough rice yield, head rice, and total milled rice. Agron. J. 88:82–84.[Abstract/Free Full Text]
- Jones, D.B., and G.H. Snyder. 1987. Seeding rate and row spacing effects on yield and yield components of drill-seeded rice. Agron. J. 79:623–626.[Abstract/Free Full Text]
- Klosterboer, A.D., and F.T. Turner. 2001. Seeding rates. In C. Blair (ed.) 2002 Rice production guidelines. Bull. D-1253:7. Texas Agric. Exp. Stn., College Station, TX.
- Moldenhauer, K.A.K., J.W. Gibbons, F.N. Lee, R.J. Norman, J.L. Bernhardt, M.A. Anders, C.E. Wilson, N.A. Slaton, J.N. Rutger, T. Tai, R. Bryant, M.M. Blocker, and A.C. Tolbert. 2001. Breeding and evaluation for improved rice varieties: The Arkansas rice breeding and development program. In R.J. Norman and J.F. Meullenet (ed.) B.R. Wells Rice Res. Studies 2000. Univ. of Arkansas Agric. Exp. Stn. Res. Ser. 485:20–26.
- Purcell, L.C. 2000. Soybean canopy coverage and light interception measurements using digital imagery. Crop Sci. 40:834–837.[Abstract/Free Full Text]
- Richardson, M.D., D.E. Karcher, and L.C. Purcell. 2001. Quantifying turfgrass cover using digital image analysis. Crop Sci. 41:1884–1888.[Abstract/Free Full Text]
- Saichuk, J., P. Bollich, B. Castro, T. Croughan, R. Dunand, E. Eskew, et al. 2002. Rice varieties and management tips 2003. Publ. 2270. Louisiana State Univ. Coop. Ext. Service, Baton Rouge, LA.
- SAS Institute. 2001. SAS user's guide. Release 8.2. SAS Inst., Cary, NC.
- SigmaScan Pro. 2003. Systat Software. Release 5.0. Point Richmond, CA.
- Wells, B.R., and W.F. Faw. 1978. Short-statured rice response to seeding and N rates. Agron. J. 70:477–480.[Abstract/Free Full Text]
- Wilson, C.E., J.W. Branson, and C.H. Davis, Jr. 2005. RICESEED 2002. Available at www.uaex.edu/Other_Areas/publications/PDF/FSA-2017.pdf (accessed 18 Jan. 2005; verified 22 Aug. 2005). Arkansas Coop. Ext. Service, Little Rock, AR.
- Wilson, C.E., R.J. Norman, N.A. Slaton, M. Duren, D.L. Boothe, K.A.K. Moldenhauer, and J.W. Gibbons. 2004. Grain yield performance of rice cultivars grown in northeast Arkansas as influenced by seeding date. In R.J. Norman et al. (ed.) B.R. Wells Rice Res. Studies 2003. Univ. of Arkansas Agric. Exp. Stn. Res. Ser. 517:327–332.
This article has been cited by other articles:
|
|
|
|
 
T. W. Walker, J. A. Bond, B. V. Ottis, P. D. Gerard, and D. L. Harrell
Hybrid Rice Response to Nitrogen Fertilization for Midsouthern United States Rice Production
Agron. J.,
February 26, 2008;
100(2):
381 - 386.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
