Published online 6 May 2005
Published in Crop Sci 45:1029-1034 (2005)
© 2005 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
CROP BREEDING, GENETICS & CYTOLOGY
Genetic Improvement Rates of Short-Season Soybean Increase with Plant Population
Elroy R. Cober*,
Malcolm J. Morrison,
Baoluo Ma and
Gail Butler
Agric. & Agri-Food Canada, Eastern Cereal & Oilseed Res. Ctr., Bldg. 110, Ottawa, ON, Canada, K1A 0C6
* Corresponding author (coberer{at}agr.gc.ca)
 |
ABSTRACT
|
|---|
Studies with a series of historic soybean [Glycine max (L.) Merr.] cultivars have found genetic improvement rates for seed yield of about 0.5% per year since soybean was first cultivated in Canada. Work with maize (Zea mays L.) has shown seed yield differences between old and new hybrids are more pronounced under higher seeding rates. The objective of our research was to determine the effect of plant populations on the rate of genetic improvement of short-season soybeans. Soybean cultivars released from1934 to 1996 were grown in 40-cm-wide rows at several seeding rates (Exp. 1, 42 cultivars grown at 25, 50, 75 seeds m–2 in 1998 and 1999; and Exp. 2, seven cultivars grown at 25, 50, 100, 150, 200 seeds m–2 in 1999 and 2000) in two fields for 2 yr at Ottawa, ON. Genetic improvement rates were based on cultivar year of release and were compared across plant populations. In Exp. 1, seed yields did not reach a plateau. In Exp. 2, a seed yield plateau was reached and genetic improvement rates generally increased with plant population. Maximum rates of genetic improvement occurred at plant populations three to four times greater than current commercial practice (40 plants m–2). A plateau for seed yield was reached at lower plant populations compared to plateaus for genetic improvement rates which were reached at higher plant populations (>80 plants m–2). Higher plant populations also resulted in earlier maturity, increased plant height, and increased lodging. Seed protein increased and seed oil decreased with higher plant populations. New soybean cultivars appear more tolerant to plant population stress than older cultivars.
Abbreviations: MG, maturity group
|
INTRODUCTION
|
|---|
ACCORDING to Donald (1963) a successful crop plant should be intraspecifically noncompetitive while still exhibiting sufficient plasticity to respond in size and shape to available conditions. Duncan (1969) concluded that the best plant density varied with the individual characteristics of the genotype. He described the soybean as a plant with reprehensible social behavior because it had a bush growth habit and large leaves that intercepted high levels of illumination, overshadowing its neighbors.
Previous research done on soybean plant density has shown that for a given row width, there is a wide range of seeding rates that result in statistically similar yields (Wiggans, 1939; Probst, 1945; Cooper, 1971; Wilcox, 1974; Ablett et al., 1984; Elmore, 1998; Bowen and Schapaugh, 1989; Devlin et al., 1995). Some of these studies have shown that there are cultivar x plant density interactions and propose that adjustments be made to populations on a regional and cultivar basis (Wiggans, 1939; Probst, 1945; Wilcox, 1974; and Ablett et al., 1984). In other studies, there were no cultivar x plant density interactions (Costa et al., 1980; Bowen and Schapaugh, 1989; Oplinger and Philbrook, 1992). Beuerlein (1988) and Ablett et al. (1991) found that semidwarf determinant types had a greater yield response to increased seeding rates than indeterminate or semideterminate types.
Duncan (1986) proposed a three phase theory describing the relationship between soybean plant density and yield. In phase I, yield response to increasing plant density is linear. Phase II begins when the yield response departs from linearity and at a specific density the full canopy will intercept 95% of the incoming solar irradiance before the end of vegetative growth. Increases in plant density may result in 95% interception occurring earlier in the season resulting in greater yield. During phase III, increases in plant density do not result in an increase in yield. Duncan (1986) used data from Wiggans (1939) and Parks et al. (1983) to support this triphasic theory and also concluded that modern cultivars showed the same response to plant density stress as their predecessors.
Yield response to plant density is a function of the leaf area intercepting solar irradiance, the leaf photosynthetic rate, and the degree to which dry matter is partitioned to grain (Egli, 1988). Dwyer et al. (1991) showed that maize hybrids released in 1988 had a greater leaf area index (LAI) and higher photosynthetic rates under full canopy than those released in 1959, and concluded that newer hybrids were more tolerant to plant density stress than their predecessors. Tollenaar and Wu (1999) reviewed yield response of maize hybrids to a range of abiotic and biotic factors such as low night temperature during grain filling, low soil moisture, low soil nitrogen, high population density, and weed interference. New hybrids performed better in these conditions leading these authors to conclude that improvements in stress tolerance were indirect responses to selection for seed yield. Duvick and Cassman (1999) reported work with a series of maize hybrids and plant densities, where yield improvement increased with increasing seeding rates.
In Canada, plant breeding of short-season cultivars has resulted in yield improvements of about 0.5% per year across 58 yr (Voldeng et al., 1997). Newer cultivars had smaller leaf area index accompanied by increased leaf photosynthetic and stomatal conductance rates (Morrison et al., 1999). Yield improvements in newer cultivars have been the result of higher harvest index manifested in a greater number of seeds per plant (Morrison et al., 2000). Our current research was done to determine if newer cultivars were more tolerant to increased plant population stress than older cultivars. The objective of our research was to determine the effect of plant populations on the rate of genetic improvement.
|
MATERIALS AND METHODS
|
|---|
This study was conducted with a series of short-season, maturity group 00 and 0, soybean cultivars released from 1934 to 1996 that was previously used to determine genetic improvement across time (Voldeng et al., 1997). We eliminated seven cultivars from the original list of 41 to have a maximum of two cultivars released per decade (Baron, Maple Isle, KG30, S 00-88, Maple Belle, Nordet, AC Harmony) and we added eight cultivars released since 1992 (Aquilon, PS 32, OAC Bayfield, S00-66, 9063, S00-55, 9042, Alta) to evaluate additional cultivars released since the study of Voldeng et al. (1997). Seeding rate response tests were grown on two sites on the Central Experimental Farm at Ottawa, Canada (45°23' N, 75°43' W) for 3 yr. The two sites were chosen to provide different environments within each year.
Experiment 1 was conducted in 1998 and 1999 with 42 cultivars and three seeding rates (25, 50, and 75 seeds m–2) representing low, recommended and high seeding rates. Experiment 2 was conducted in 1999 to 2000 with a subset of seven cultivars chosen to represent one cultivar released per decade (Mandarin, Capital, Merit, Altona, Maple Arrow, Maple Glen, OAC Bayfield). Five seeding rates were used, 25, 50, 100, 150, 200 seeds m–2 to provide a larger range of plant populations.
In 1998 and 2000, experiments were conducted on a Granby well-drained sandy loam soil (Coarse-loamy, Mixed, Mesic, Endoaquolls) and a North Gower imperfectly drained clay loam (Typic Endoaquolls), both in the Orthic Humic Gleysol subgroup in Canadian soil classification. In 1999, the experimental sites were a poorly drained North Gower clayey loam Orthic Humic Gleysol (fine loamy, Mixed, Mesic, Endoaquolls) and a well drained Uplands sandy loam Orthic Humo-Ferric Podzol (Haplorthods).
In each case, soil samples were taken before seeding and sufficient P, and K were applied during land preparation based on soil test recommendations. Plot size in each of the experiments was 1.6 x 5 m, consisting of four rows spaced 0.400 m apart. The experiments were seeded on 21 and 22 May 1998, 17 and 20 May 1999, and 21 and 28 May 2000. Weed control was achieved by chemical spray. Days to maturity, plant height, and lodging were recorded for each plot. Plant population was determined at maturity by counting plants in the center two rows of the plots. All four rows in each plots were combine-harvested, the seed air-dried, and grain yield reported at 130 g kg–1 moisture. Seed weight was determined for 100 seeds from each plot. Seed protein and oil was determined by near-infra red transmission (Grainspec, Foss North America, Inc., Brampton, ON).
A two factor (seeding rates and cultivars) factorial design, arranged in a randomized complete block design with three replications, was used for each site-year. An ANOVA was performed for each of the two sites in each of the two years using the general linear and mixed model procedures of SAS (SAS Institute, 1999). Linear and nonlinear contrasts were examined for year of cultivar release, seeding rate and their interaction to determine differences in cultivars for different seeding rates. The linear response of seed yield across year of cultivar release (or genetic improvement rate) was estimated for each environment using a reduced model with fixed effects, seeding rate and the interaction of genetic improvement with seeding rate (slope over year of release). These slopes were plotted against plant populations. Combined analyses were done over years for each experiment. In these analyses, interactions with environment (site-year) were assumed to be random effects while the design effects, year of cultivar release and seeding rate, were considered fixed effects. Significance is at the 5% level unless otherwise indicated. Data presented in the figures are observed plant populations rather than seeding rates.
|
RESULTS AND DISCUSSION
|
|---|
In Exp. 1, seeding rates resulted in final plant populations ranging from 17.7 to 63.6 plants m–2 (Table 1). A yield plateau was not evident in 1998 (Fig. 1) so Exp. 2 was initiated in 1999 to evaluate more and higher seeding rates. Final plant populations ranged from 10.6 to 158.1 plants m–2 for Exp. 2. A seed yield plateau was reached in Exp. 2 (Fig. 1) with the mean data in Table 1 indicating that phase II (Duncan, 1986) terminated at approximately 66 plant m–2.
View this table:
[in this window]
[in a new window]
|
Table 1. Mean seed yield and agronomic characteristics of a series of short-season soybean cultivars (42 for Exp. 1; 7 for Exp. 2) released from 1934 to 1996 over a series of seeding rates from trials grown at Ottawa, ON from 1998 to 2000.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 1. Mean seed yield over plant populations derived from a) 42 cultivars (Exp. 1), and b) seven soybean cultivars (Exp. 2) released from 1934 to 1993, from trials grown at Ottawa, ON. LSD 0.05 for Exp. 1 in 1998 on clay loam and sandy loam are 58 and 85 and for 1999 on clay loam and sandy loam are 82 and 111; and for Exp. 2 for 1999 on clay loam and sandy loam are 207 and 214 and for 2000 on clay loam and sandy loam are 225 and 195, respectively.
|
|
In Exp. 2, initial analysis of seed yield indicated variances were homogeneous across site-years. In the analysis across site-years, there was a highly significant seeding rate x site-year interaction (P < 0.01) and a smaller but significant seeding rate x variety x site-year interaction thus the results are presented by site-year (Fig. 1 and 2). In the overall analysis, as well as in each site-year (Fig. 2), there was a significant nonlinear effect for cultivar; both the linear contrast (genetic improvement rate) and quadratic contrast were significant. In the overall analysis, the seeding rate x cultivar interaction was not significant compared; however, the one degree of freedom linear x linear contrast of year of release x seeding rate was significant. In the analysis by site-year, the linear x linear contrast of year of release x seeding rate was significant for all sites except in the 1999 clay soil (P = 0.12). The remaining interaction once this one degree of freedom contrast was accounted for was not significant in any site-year. To investigate this interaction further the genetic improvement rate (linear effect of year of cultivar release) was plotted against plant population (Fig. 3). In Exp. 2, genetic improvement rates were not significantly different from zero at some of the lowest plant populations but as plant population increased, the genetic improvement rate became significantly different from zero (Fig. 2). While the plant populations in Exp. 1 were in a narrower range, genetic improvement rates showed the same trend with plant populations. Specht et al. (1999) reported that the yield difference between two newer soybean cultivars and four older cultivars increased as seeding rate increased from 30 to 100 plants m–2. These results are similar to reports of historic maize hybrids planted over a range of densities (Duvick and Cassman, 1999) where rates of genetic improvement were zero at low populations and increased at higher populations.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 2. Seed yield of seven short-season soybean cultivars and their year of release over a series of seeding rates (Exp. 2) from trials grown on a) clay loam and b) sandy loam in 1999; and c) clay loam and d) sandy loam in 2000 at Ottawa, ON. Least Significant Differences (0.05) between any two cultivar by seeding rate means for 1999 clay loam and sandy loam are 548 and 567 and for 2000 clay loam and sandy loam are 596 and 516, respectively.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 3. Mean genetic improvement rates as influenced by plant population derived from a) 42 cultivars (Exp. 1), and b) seven soybean cultivars (Exp. 2) released from 1934 to 1993, from trials grown at Ottawa, ON. Standard errors of means for Exp. 1 in 1998 on clay loam and sandy loam are 2.5 and 2.4 and for 1999 on clay loam and sandy loam are 3.2 and 2.9; and for Exp. 2 for 1999 on clay loam and sandy loam are 4.3 and 4.1 and for 2000 on clay loam and sandy loam are 4.8 and 5.9, respectively.
|
|
The genetic improvement rates in Exp. 2 ranged from below zero to 20 kg ha–1 yr–1 with a mean of 8.9 kg ha–1 yr–1 (Fig. 3). These estimates are similar to estimates of 11.1 kg ha–1 yr–1 for MG 00 and 0 cultivars reported by Voldeng et al. (1997) and indicate that the subset of seven cultivars in Exp. 2 is representative of the larger set. Our estimates are also within the range of 19.1 kg ha–1 yr–1 for MG 00 and 17.4 kg ha–1 yr–1 for MG 0 reported by Specht and Williams (1984). In our study, maximum rates of genetic improvement were evident at very high plant populations (
100 plant m–2), which are well above commercially recommended populations of 40 plant m–2. Modern soybean cultivars were able to tolerate plant population stress better than their predecessors.
The immediate conclusion would be that new cultivars were more tolerant to increased plant populations because they were selected under modern, high-density conditions while older ones were not. However, the maximum genetic improvement rate occurred at plant populations approximately three to four times greater than currently recommended densities. This suggests that soybean breeders may be able to make more progress if lines are selected at higher plant densities. The resulting lines may be less intraspecifically competitive.
Increased plant populations resulted in slightly but significantly earlier maturity in both experiments and slightly increased plant height, which was significant in Exp. 1 (Table 1). Lodging increased as plant populations increased even when not accompanied by increased plant height. Previous seeding rate research has found that lodging either increases with increasing seeding rate (Cooper, 1971; Wilcox, 1974; Ablett et al., 1984) or there was limited effects (Probst, 1945; Devlin et al., 1995). Luedders (1977) and Morrison et al. (1999) found that older soybean cultivars had higher lodging scores even though the height was not significantly different than newer cultivars.
In our experiment, seed weight was relatively unaffected by plant population, while seed protein increased and seed oil decreased with increasing plant population. Wilcox (1974) found no consistent change in seed weight as affected by seeding rate, while Moore (1991) showed that seed weight was greatest at low plant densities. Surprisingly, plant survival to maturity remained constant over seeding rates as shown by a good fit (r2 = 0.95) to linear regression of final plant stand on seeding rate (b = 0.67). Higher plant populations did not result in more plants dying over the season as a result of inter-plant competition. Devlin et al. (1995) reported that higher seeding rates in 76-cm-wide rows resulted in greater plant mortality than low seeding rates, but there was no effect of seeding rate on plant mortality at 20-cm row widths.
Previous studies with historical soybean cultivars have found that modern cultivars have less total leaf area and higher photosynthetic rates than their predecessors (Morrison et al., 1999). Yield is a function of LAI, photosynthetic rate and harvest index, (Egli, 1988), as well as incident radiation and duration of photosynthetic activity. Since new cultivars have a lower LAI but higher leaf photosynthetic rate they would be able to tolerate greater plant densities before mutual shading reduced photosynthetic rates to a point that would impact on yield. As well, a smaller LAI for new cultivars may increase the optimum population for optimum yield. Dwyer et al. (1991) found that the LAI of recent corn hybrids was larger than older hybrids and while leaf photosynthetic rates declined in all hybrids tested as densities increased, the new hybrids were able to maintain photosynthetic rates at higher densities than the older ones. The interaction of LAI, photosynthetic rate, and plant population in short-season soybean cultivars remains to be studied.
In summary, seed yield differences between old and new soybean cultivars were nonexistent under low plant densities became significant at commercial plant densities, and were maximized at plant populations about three to four times higher than current commercial plant populations. Future yield gains may be achieved by selecting lines under higher plant population stress.
|
NOTES
|
|---|
ECORC Contribution no. 03-384.
Received for publication April 13, 2004.
|
REFERENCES
|
|---|
- Ablett, G.R., W.D. Beaversdorf, and V.A. Dirks. 1991. Row width and seeding rate performance of indeterminate semi-determinate and determinate soybean. J. Prod. Agric. 4:391–395.
- Ablett, G.R., J.C. Schleihauf, and A.D. McLaren. 1984. Effect of row width and population on soybean yield in southwestern Ontario. Can. J. Plant Sci. 64:657–659.
- Beuerlein, J.E. 1988. Yield of indeterminate and determinate semi-dwarf soybean for several planting dates, row spacings, and seeding rates. J. Prod. Agric. 1:300–303.
- Bowen, C.R., and W.T. Schapaugh, Jr. 1989. Relationships among charcoal rot infection, yield and stability estimates in soybean blends. Crop Sci. 29:42–46.[Abstract/Free Full Text]
- Cooper, R.L. 1971. Influence of soybean production practices on lodging and seed yield in highly productive environments. Agron. J. 63:490–493.[Abstract/Free Full Text]
- Costa, J.A., E.S. Oplinger, and J.W. Pendleton. 1980. Response of soybean cultivars to planting patterns. Agron. J. 72:153–156.[Abstract/Free Full Text]
- Devlin, D.L., D.L. Fjell, J.P. Shroyer, W.B. Gordon, B.H. Marsh, L.D. Maddux, V.L. Martin, and S.R. Duncan. 1995. Row spacing and seeding rates for soybean in low and high yielding environments. J. Prod. Agric. 8:215–222.
- Donald, C.M. 1963. Competition among crop and pasture plants. Adv. Agron. 15:1–118.
- Duncan, W.G. 1969. Cultural manipulation for higher yields. In J.D. Eastin et al. (ed.) Physiological aspects of crop yield. ASA, Madison, WI.
- Duncan, W.G. 1986. Planting patterns and soybean yields. Crop Sci. 26:584–588.[Abstract/Free Full Text]
- Duvick, D.N., and K.G. Cassman. 1999. Post-green revolution trends in yield potential of temperate maize in the North-Central United States. Crop Sci. 39:1622–1630.[Abstract/Free Full Text]
- Dwyer, L.M., M. Tollenaar, and D.W. Stewart. 1991. Changes in plant density dependence of leaf photosynthesis of maize (Zea mays L.) hybrids 1959 to 1988. Can. J. Plant Sci. 71:1–11.
- Egli, D.B. 1988. Plant density and soybean yield. Crop Sci. 28:977–981.[Abstract/Free Full Text]
- Elmore, R.W. 1998. Soybean cultivar response to row spacing and seeding rates in rainfed and irrigated environments. J. Prod. Agric. 11:326–331.
- Luedders, V.D. 1977. Genetic improvement in yield of soybeans. Crop Sci. 17:971–972.[Abstract/Free Full Text]
- Morrison, M.J., H.D. Voldeng, and E.R. Cober. 1999. Physiological changes from 58 years of genetic improvement of short-season soybean cultivars in Canada. Agron. J. 91:685–689.[Abstract/Free Full Text]
- Morrison, M.J., H.D. Voldeng, and E.R. Cober. 2000. Agronomic changes from 58 years of genetic improvement of short-season soybean cultivars in Canada. Agron. J. 92:780–784.[Abstract/Free Full Text]
- Moore, S.H. 1991. Uniformity of plant spacing effects on soybean population parameters. Crop Sci. 31:1049–1051.[Abstract/Free Full Text]
- Oplinger, E.S., and B.D. Philbrook. 1992. Soybean planting date, row width and seeding rate response in three tillage systems. J. Prod. Agric. 5:94–99.
- Parks, W.L., J. Davis, R. Evans, M. Smith, T. McCutchen, L. Safley, and W. Sanders. 1983. Soybean yields as affected by row spacing and within row plant density. Tenn. Agric. Exp. Stn. Bull. 615.
- Probst, A.H. 1945. Influence of spacing on yield and other characters in soybeans. J. Am. Soc. Agron. 37:549–554.
- SAS Institute Inc. 1999. SAS/STAT users' guide, Version 8, Cary, NC.
- Specht, J.E., D.J. Hume, and S.V. Kumudini. 1999. Soybean yield potential—A genetic and physiological perspective. Crop Sci. 39:1560–1570.[Abstract/Free Full Text]
- Specht, J.E., and J.H. Williams. 1984. Chapter 3. Contribution of genetic technology to soybean productivity—retrospect and prospect. p. 49–74. In W.R. Fehr (ed.) Genetic contribution to yield gains of five major crop plants. CSSA Special Publ. 7. CSSA and ASA, Madison, WI.
- Tollenaar, M., and J. Wu. 1999. Yield improvement in temperate maize is attributable to greater stress tolerance. Crop Sci. 39:1597–1604.[Abstract/Free Full Text]
- Voldeng, H.D., E.R. Cober, D.J. Hume, C. Gillard, and M.J. Morrison. 1997. Fifty-eight years of genetic improvement of short-season soybean cultivars in Canada. Crop Sci. 37:428–431.[Abstract/Free Full Text]
- Wiggans, R.G. 1939. The influence of space and arrangement on the production of soybean plants. J. Am. Soc. Agron. 31:314–321.
- Wilcox, J.R. 1974. Response of three soybean strains to equidistant spacings. Agron. J. 66:409–412.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
J. V. Wiersma
Iron Acquisition of Three Soybean Varieties Grown at Five Seeding Densities and Five Rates of Fe-EDDHA
Agron. J.,
June 5, 2007;
99(4):
1018 - 1028.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
