| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Plant Agriculture, Univ. of Guelph, Guelph, ON, Canada, N1G 2W1
jspecht1{at}unl.edu
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
in the USA have risen steadily during the last 50 yr. Still, yield improvement will have to continue well into the next century to meet the dietary needs of the 1 x 1010 people expected to occupy earth by the year 2050 (Waggoner, 1994). Projection of crop yield to the middle of the 21st century is not an easy task. In most instances, a linear model often provides the best fit when crop yield is regressed on production year. Although a linear equation can be used to project crop yield a few years into the future (see Table 3-3 of Specht and Williams, 1984), extrapolation of crop yield to the far distant future requires a different equation. Why? Simply put, crop yield has a biological maximum. A finite supply of solar energy cannot be translated into an infinite supply of reduced carbon. Thirty years ago, de Wit (1967), using some reasonable assumptions about photosynthetic efficiency and respiratory losses, deduced that the biological maximum for biomass yield was about 45 Mg ha-1 (at 40° latitude). Assuming a 50% harvest index, the seed yield limit is inferred to be 22.5 Mg ha-1.
Given a limit to crop yield improvement, Waggoner (1994) observed that a logistic model was more suitable for crop yield vs. time data
 |
The logistic response curve is sigmoid because the exponential change in yield (Y) over time (T) is modulated by the degree to which present yield differs from its limit (K). The inflection point (Tm) separates the positive and negative exponential phases of the sigmoid. Waggoner (1994) used average U.S. corn (Zea mays L.) yields from 1940 to 1992, and set K to 21 Mg ha-1 (the highest grain yield recorded in U.S. corn production), to show that U.S. corn yield was rising at a logistic rate of 3.6% yr-1.
Attaining the theoretical yield maximum on each U.S. farm each year for each crop would unquestionably require optimization of every yield-affecting biotic or abiotic factor in every production environment. The degree to which such multiple-factor optimization is possible, or economically feasible, will determine what fraction of the crop's yield potential that can actually be realized on the average U.S. farm on a year-to-year basis. In any event, mitigation of the yield-limiting factors will require in the future, as it has in the past, continuous injections of technological innovation into the agricultural enterprise (Specht and Williams, 1984; Waggoner, 1994).
Genetic progress in five major U.S. crop species was last documented and reviewed in a special publication edited by Fehr (1984). With the 20th century nearing its end, it seems appropriate to again review the historic improvements achieved in crop yield and other traits, particularly those improvements of the last quarter century. Soybean, a C-3 photosynthetic species, is the focus of this report, with corn, a C-4 species, discussed only for comparative purposes. The objective of this paper is to provide the reader with our perspective (genetic and physiological) on past soybean yield improvement and what it might portend for yield improvement in the future.
|
Materials and methods |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
Each group of yield vs. time data was subjected to regression. The best-fitting model in each regression analysis was a linear one, except that the linear and exponential models fit the U.S. data (1924–1988) at essentially the same R2 values. All linear regression coefficients computed in this study were significantly different from zero (
= 0.05), so standard errors were computed to ascertain significant differences between those coefficients. Trend lines shown in the figures were derived from the linear equation, Y = bT + a, or the exponential equation, Y = a x e-r x T, where Y = yield, T = year, a = constant, b = linear coefficient, and r = exponential coefficient. The equations and their R2 values are displayed in boxed legends next to the trend lines in the figures. A logistic trend line was derived from the equation, Y = K/[1+e-r(T-Tm)], where K = soybean yield limit, r = logistic coefficient, and Tm = year of the inflection point on the sigmoid curve. Arguments for using 8000 kg ha-1 as a soybean yield asymptote are presented in this paper. The values of two other logistic parameters, r and Tm, were ascertained by determining the best fit of the accelerating phase of the logistic curve to the exponential curvature present in the soybean yield vs. time data (cf. Waggoner, 1994).
Yield contest data were obtained from three state soybean associations (NE, IA, MO) that have sponsored this activity in recent years. The data consisted of the top two winning yields each year within irrigated and rainfed production classes, but if not so categorized, then the top two yields overall. The top two winning yields were used here to discern cases in which the first-place yield was substantially larger than the second-place yield. The top five winning yields in the U.S. national contest conducted for 3 yr in the mid-1960s, and the highest yields attained in yield maximization research conducted by Richard Flannery (Rutgers Univ.) and Richard Cooper (USDA-ARS, Ohio State Univ, Wooster, Ohio), were provided by Dr. Cooper (personal communication). Contest-winning yields were plotted vs. the year of occurrence to better visualize possible soybean yield trends over time in farm environments of exceptional productivity.
Physiological Studies
Detailed studies were conducted on 14 soybean cultivars representing six decades of releases. The experiments were planted from 1993 to 1996 at Ottawa, ON. Specific procedures were described by Morrison et al. (2000).
Dry Matter and N Accumulation and Depodding Effects. Two old cultivars (Mandarin, released in 1934; and Pagoda, released in 1939) and two new (Maple Glen, 1987 and OAC Bayfield, 1994) cultivars were planted in field trials in 1996 and 1997. Pagoda and Maple Glen were selected to have similar maturities (113–115 d) as were Mandarin and OAC Bayfield (122–124 d). Plantings were at the Elora Res. Stn. in both years. The soil is a tile-drained London loam (Typic hapludalfs). Planting dates were 27 May 1996 and 28 May 1997. Main plots were 15 m long and 16 rows wide, with 36 cm between rows. Plots were over planted and thinned to 50 plants m-2. Weeds were controlled with herbicides and hand-weeding. The treatment design was a split-plot, with cultivars as main plots and 10 bordered subplots, each 76 by 76 cm. The experimental design was a RCB with four replicates. Seven subplots were allocated at random for harvest at stages V5, R1, R4, R5, R6, R6.5, and R7 (Fehr and Caviness, 1980). Biomass samples were dug at each harvest. A 10-plant subsample from each subplot was divided into roots, stems and branches, leaves; and pods and seeds, if present. Samples were dried to constant moisture and weighed. Component part percentages were used to calculate biomass in individual plant parts per unit area. Two subplots were allocated to a 50% depodded treatment and an untreated control. Depodding involved removing the pods from every second node as they formed. These subplots were harvested at maturity (R8). A 1.44 x 6.0 m bordered strip was harvested with a plot combine for final yield. Analyses of variance appropriate for a split-plot RCB were conducted for each response variable.
Effects of Year of Release on N2 Fixation and Soil N Uptake. Six cultivars, representing releases from 1957 to 1993, were planted near Elora, ON. in 1995 and 1997. Cultivars were Crest (released in 1957), Beechwood (a selection out of Evans, released in 1974), Maple Arrow (1976), Bicentennial (1983), PS42 (1990), and OAC Bayfield (1993). Plots were either heavily inoculated with Bradyrhizobium japonicum inoculant or not inoculated. Planting was into soil free of B. japonicum. The treatment design was a split-plot with inoculant treatments as main plots and cultivars as sub-plots in a RCB with four replications. Plots were 6 m long and seven rows wide with 38 cm between rows. Other cultural details were as described in the previous section. Total plant N was measured with a LECO N analyzer (LECO Corp, St. Joseph, MI). N2 fixation was calculated from the difference between inoculated and uninoculated plots of the same cultivar. Soil N uptake was measured as the total N content of the uninoculated treatments.
Effects of Year of Release on Response to Planting Density. The same six cultivars, representing two each for the pre-1976, post-1976, and modern eras, were planted in 1995 and 1997 at the Elora Research Station. Plots were marked by hand and planted with Planet Junior planters with 19, 38, or 57 cm between rows. Plots were over planted and thinned to the same population per row to give plant populations of 33, 50, or 100 plants m-2. Plots were 6 m long and eight rows wide. The treatment design was a split plot with plant density as the main plot and cultivars as subplots in a RCB with four replicates. Yields were measured with a plot combine.
|
Results and discussion |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
|
|
The degree of yield improvement occurring in soybean versus that in corn is worthy of examination (Fig. 2B). The linear rates of yield improvement of soybean and corn NE irrigated production are 35.1 ± 5.3 and 98.0 ± 18.9 kg ha-1 yr-1, respectively, which ostensibly suggests that yield improvement in corn is 2.8 x faster than that in soybean. In fact, this 2.8 to 1 advantage of corn over soybean also exists in terms of their absolute yields, whether measured in terms of mean crop productivity over the last 25 yr, or in terms of the y-intercepts of the corn-soybean trend line equations. Water stress is presumably not a factor in NE irrigated corn and soybean production, and the two crops are generally rotated over the same fields. That this 2.8 to 1 ratio persists in two very different measures of yield (one absolute, the other its rate of improvement) is an empirical reflection of an intrinsic corn-soybean productivity difference.
There are at least two biological reasons why the productivity of corn is superior to that of soybean. The first is their differing photosynthetic mechanisms. Corn is a C-4 species for which photosynthesis is very efficient, due to a CO2-concentrating mechanism that mitigates or eliminates photorespiratory carbon loss, despite the "cost" of a higher quantum mole requirement per mole of CO2 fixed. Soybean, on the other hand, is a C-3 species in which photorespiratory carbon losses substantially constrain its photosynthetic output, particularly with warmer temperatures and/or greater water stress. The second reason is that the two species deposit substantially differing fractions of carbohydrate, protein, and lipid in their seeds. The respective averages by weight are about 840, 100, and 50 g kg-1 for corn, and about 380, 380, and 200 g kg-1 or soybean (Sinclair and de Wit, 1975). Extending the work of Penning de Vries et al. (1974), McDermitt and Loomis (1981) imputed "production values" of 0.83, 0.40, and 0.33 for carbohydrate, protein, and lipid, to demonstrate how much of each constituent can be produced from a unit mass of (photosynthetically produced) glucose. Comparatively stated, more seed mass can be produced from a given supply of glucose when carbohydrate is the predominant seed constituent (i.e., corn), than when it is not (i.e., soybean). To make soybean yield improvement as fast as that of corn would likely require radical changes in its seed composition, thus destroying the characteristics for which it is valued as a crop species. At any rate, expressing each crop's yield improvement rate as a function of its current absolute yield reveals rates of relative improvement that are effectively identical.
Annual improvement in U.S. soybean yields (as depicted in Fig. 1) is attributable to: (i) rapid producer adoption of repetitive waves of agricultural innovation in the form of genetic, agronomic, or genetic x agronomic technologies that provide producers with ever better means for attenuating their "on-farm" yield constraints (Specht and Williams, 1984), and (ii) the inexorable annual rise in atmospheric CO2 concentration (currently 1.5 µL L-1 yr-1), with each annual increment in "carbon fertilization" enhancing soybean photosynthesis, biomass yield, and water-use-efficiency (Kimball, 1983; Waggoner, 1984; Allen et al., 1987, 1998). Let us consider each of the above items in turn.
Genetic Improvement
How much of the annual improvement in soybean yield is attributable to genetic technology? Several research groups addressed this question in the USA during the late 1970s and early 1980s, and a Canadian group has published more recent data on the regression of cultivar yield on year-of-cultivar release (Table 1)
. Linear estimates of the annual genetic improvement in soybean yield in these studies were in the range of about 10 to 30 kg ha-1 yr-1. Specht and Williams (1984) observed that their initial estimate of 18.8 kg ha-1 yr-1 was about 80% of the 23.7 kg ha-1 yr-1 rate of (on-farm) yield improvement. However, they also noted a major discontinuity in the 75-yr yield trend line. Regression analyses conducted with cultivars of hybridization origin (mostly post-1943 releases) vs. those conducted with cultivars of plant introduction origin (mostly pre-1943 releases) revealed substantial different y-intercept values. The intersection of these two regression lines in the year 1943 revealed that the former group of cultivars had an absolute yield advantage of 25%. This "quantum jump" in the genetic yield improvement—the effective equivalent of what occurred in corn when hybrids replaced open-pollinated varieties—resulted when breeders shifted from plant introduction to hybridization as their means for developing and releasing new cultivars, thereby expanding (via recombination) the genetic variance available to them for selection. Specht and Williams (1984) noted that the rate of genetic yield gain since 1943 was only 12.5 kg ha-1 yr-1 for the public cultivar releases. Because improved public cultivars were being rapidly adopted by U.S. soybean producers, they suggested that only 50% (not the 80% mentioned earlier) of 23.7 kg ha-1 yr-1 gain in northcentral U.S. soybean yield was attributable to genetic improvement.
|
A substantial portion of the "new" cultivar market is now serviced by the proprietary seed trade. We recently examined proprietary information that two entities in the seed trade (Asgrow Seed Company and Pioneer Hi-Bred International) graciously provided to us, which consisted of yield vs. year of release data for cultivars in maturity group II and III. The genetic yield gains we calculated for proprietary cultivar releases (i.e., 25–30 kg ha-1 yr-1) were the approximate equivalents of the Voldeng et al. (1997) estimate for public cultivar releases (30 kg ha-1 yr-1). These data suggest that genetic improvement in soybean yield is occurring at nearly triple the rate of 12.5 kg ha-1 yr-1 rate that Specht and Williams (1984) estimated for pre-1977 cultivar releases.
It should be kept in mind that all estimates of genetic improvement are relative, not absolute. The yield difference between new and old cultivars tends to be narrower when the evaluation is conducted in test environments of severely constrained yield potential, and inversely wider when evaluated in substantially more productive environments. Simply put, an interaction between the degree of genetic yield gain and the yield of the evaluating environments is an intuitive expectation, and this has been clearly documented in wheat (Triticum aestivum L.) (Slafer et al., 1994). What is the case for soybean? Wilcox et al. (1979) computed the stability-regression coefficients for old and new cultivars and, except for the cultivar Amsoy 71, was unable to detect significant differences between these two cultivar classes, despite a range in coefficient value from 0.88 to 1.17. Voldeng et al. (1997), using a larger number of cultivars, was also unable to discern a difference in stability-regression coefficients between old and new cultivars (range of 0.39–2.23). The high-yield Argentina environments used by Salado-Navarro et al. (1993) might have provided relevant data for this question, but estimates of the genetic yield gain there were (surprisingly) zero. This zero yield gain observation leads us to a crucial issue for plant breeders. Assuming no pests or lodging, does the yield difference between old and new cultivars (i) widen, (ii) stay the same, or (iii) actually narrow as the productivity of the evaluation environment approaches the biological yield limit of the crop? That is, if genetic yield gain were to be measured in a yield-contest winning environment, would it be zero? Experiments with soybean, similar to an ongoing one in corn (Duvick, 1997), will be required to answer this question. Such experiments will need to be designed and executed with precision, since statistical power is ordinarily too low (i.e., Type II error probability is too high) to be able to detect meaningful differences among the estimates of genetic yield gain generated from environments of intentionally different productivities.
Agronomic Improvement
Unlike genetic technology, which is adopted by producers as soon as it becomes available, agronomic technologies often entail lengthy learning curves, or else their implementation requires substantial capital expenditures for equipment or products (e.g., combine yield monitors, global positioning system software, to name a few recent examples). Consequently, a significant lag time invariably occurs between the development of an agronomic technology and its adoption by a majority of producers.
Of the many agronomic and management practices that have contributed to U.S. soybean yield improvement, earlier planting, narrower rows, better weed control, and lower harvest losses have probably had a major impact. Greater producer recognition of the cash value of soybean vis-a-vis corn, plus the phase-out of governmental subsidies for corn, have favorably impacted state soybean yield averages. For example, the upward trend in Nebraska irrigated soybean yield probably reflects to some degree the soybean yield benefit that arises as more producers shift from monoculture corn on their best rainfed or irrigated lands to a corn-soybean rotation. Closing the large gap that exists between the yield possible in agricultural research plots, and that actually realized on the "average" U.S. farm might, as Specht and Williams (1984) observed, require more emphasis on agronomics, since the "average" U.S. farmer commonly grows the same (new) cultivars that yield-contest-winning farmers do. Yield contest data will be discussed in the last section of this paper.
Impact of Rising CO2 on Soybean Yield
Waggoner (1984) projected a 0.1 to 0.2% annual rise in U.S. soybean yield if, as he arbitrarily proposed, atmospheric CO2 were to rise from 340 µL L-1 In 1983 to 400 µL L-1 in 2000. He then contrasted those annual percentages with the 1.1% annual increase in U.S. soybean yield that occurred from 1963 to 1976. Implicit in his contrast was the suggestion that a 60 µL L-1 increase in CO2 over 17 yr (i.e., 3.5 µL L-1 yr-1) might account for about 1.7 (i.e., 10%) to 3.4 (20%) kg ha-1 yr-1 of the 17.3 kg ha-1 yr-1 yield improvement from 1963 to 1976.
Allen et al. (1987) used a rectangular hyperbola model for soybean response to increasing CO2 and predicted a cumulative 32% increase in soybean yield if the CO2 concentration of 315 µL L-1 in 1958 were to double to 630 µL L-1 by the year 2058. Given the U.S. soybean yield average of 1540 kg ha-1 in 1958, a 32% yield increase would translate into 493 kg ha-1 of future soybean yield. With a 100-yr time frame for CO2 doubling, the 32% increase translates into a 5 kg ha-1 yield increase per annum, which conceivably would have accounted for about 15% of the 31.4 kg ha-1 yr-1 rate of on-farm U.S. soybean yield improvement from 1972 to 1997 (Fig. 2).
Continued increases in atmospheric CO2 could, of course, affect global climate parameters in ways that might mitigate any CO2-induced increase in soybean productivity. However, collateral improvement in the photosynthesis/transpiration ratio (i.e., water use efficiency) would certainly offset some of the negative effects of global warming, particularly for a C-3 species like soybean that is often exposed to water stress. As Farquhar (1997) put it: "... doubling the CO2 concentration is almost like doubling the rainfall ...".
Impact of Past Yield Improvement on Physiological Traits
Voldeng et al. (1997) observed that the post-1976 acceleration of genetic improvement in Canadian cultivars was likely attributable to the introduction of early-maturing, cold-tolerant germplasm from Hokkaido via Sweden. A number of physiological studies have now been conducted on the old and new MG 00 and 0 cultivars. Published and unpublished data relevant to yield improvement are reported in the next sections.
Dry Matter Accumulation
In arriving at higher grain yields, there are basically two major pathways: increased harvest index or increased dry matter accumulation. In a review of genetic improvement of soybean, Frederick and Hesketh (1994) noted that there appeared to be little change in soybean harvest index. This is consistent with the results of a physiological evaluation of old versus new corn hybrids, for which genetic improvement was found to be associated with assimilate supply during seed filling period (SFP), but not with changes in harvest index (Tollenaar and Aguilera, 1992). In a soybean study comparing old and new soybean cultivars, a similar association was found between assimilate supply during the SFP and yield improvement (Kumudini and Hume, 1996-1997, unpublished data). Two old and two new soybean cultivars, evaluated over a 2-yr period, had similar patterns of dry matter accumulation until the beginning of the SFP (Fig. 3)
. After the onset of the SFP, the newer cultivars accumulated dry matter at a greater rate than old cultivars. These findings were consistent with those of Shiraiwa and Hashikawa (1995), in which the two modern Japanese cultivars had more than double the dry matter increase during the SFP, when compared to two old cultivars.
|
|
|
Partial depodding increased root N content in older cultivars (Fig. 5) , whereas the root N content of more recent releases was unaffected (Kumudini and Hume, 1996-1997, unpublished data), suggesting that more recent releases are more effective in moving N from the roots. Voldeng et al. (1997) reported that, during 58 yr of Canadian cultivar releases, seed protein content declined by 4 g kg-1 whereas oil rose by 4 g kg-1, directional changes that Wilcox et al. (1979) had observed earlier (Table 1). On average, a 1 kg-1 increase in oil content will usually lead to a 2 kg-1 decrease in protein content (reflecting a negative genetic correlation between the two). Canadian soybean breeders have apparently been able to increase seed yield while still ensuring that the combined percentage of seed protein and oil content did not appreciably change (Voldeng et al., 1997). Although seed protein concentration declined over the 6 decades by a total of 24 g kg-1, that decline was more than offset by a 36 kg increase in seed protein "yield on a per ha basis" resulting from the genetic improvement in yield. Omielan and Hume (unpublished data) conducted field studies with six old-to-new cultivars from 1995 to 1997. The objective was to determine whether the extra seed N per ha in the new cultivars resulted from greater soil N uptake, greater N2 fixation, or both. For hybridization-derived cultivars released from 1957 and 1993, the experimental data clearly indicate that the improvement in N content was due to greater N2 fixation (Fig. 6) . This would support the concept that new cultivars are better able to supply assimilates during the SFP than old cultivars, thus allowing more assimilate supply to roots and nodules.
|
|
|
The Biological Limit to Soybean Yield Improvement
We conclude this paper with a discussion of the biological limit to soybean yield and what it might portend for future soybean improvement. Although some might legitimately argue its magnitude, none can logically argue against its existence, although one must certainly allow for its gradual increase so long as atmospheric CO2 continues its gradual rise.
Yield contest data can provide useful information about crop yield potential (as defined by Evans and Fischer, 1999, this issue), so long as one recognizes that these "cherry-picked" examples of high yield tend to arise from favorable confluences of genotype, management, soil type, rainfall, weather, etc. Soybean producers in the USA who have won national and state yield contests (NE irrigated and rainfed, plus IA and MO), and U.S. scientists who have conducted maximum (field-based) yield research (Fig. 8) have not (yet) exceeded the 8000 kg ha-1 yield level (R.L. Cooper, 1998, personal communication). Although the data are obviously sparse, the winning yields of the national yield contests conducted in the mid-1960s, which were accomplished with then available cultivars and CO2 levels, were only recently approached as shown by a 1997 winning yield of 6660 kg ha-1 in Nebraska's irrigated contest category. The upward yield trend apparent in recent years for all contests (except the NE rainfed class) may just be an artifact of the weather "sampled" in recent years.
|
Returning to Fig. 1, it can be seen that either a linear or an exponential model provides a good fit to the long-term (1924–1998) U.S. yield improvement data. The two models have virtually identical R2 values (0.929 and 0.924, respectively). The linear model provides a quite conservative projection of future yield improvement, given the clear evidence that genetic yield improvement has substantially accelerated during the past 20 yr (Voldeng et al., 1997 vs. Specht and Williams, 1984). Conversely, the exponential model provides a very liberal estimate of future yield improvement, since an indefinite compounding of the annual yield improvement rate would ultimately produce (biologically illogical) annual improvements of infinite magnitude. By specifying a yield limit of 8000 kg ha-1 for soybean, the upward movement of yield over time takes on a sigmoid response pattern as yield moves towards a (biologically logical) horizontal asymptote. The sigmoid inflection point, arbitrarily chosen here as the year when Y = K/2 (i.e., 4000 kg ha-1), separates the logistic rise in soybean yield into an initial positive phase (acceleration) and a final negative phase (deceleration). In Fig. 9 , we display the soybean yield improvement scenarios that the exponential, logistic, and linear models project for soybean yield improvement in the distant future. The logistic model projects that the national U.S. soybean yield "average" will reach the sigmoid inflection point (i.e., 4000 kg ha-1) by the year 2043. This represents a 60% increase over the 1998 U.S. yield of about 2500 kg ha-1. One needs to keep this 60% estimate in mind given the casual talk about needing to double (100%) or triple (200%) soybean yields in the next 50 yr. The 4000 kg ha-1 yield milestone conceivably could be achieved in an earlier year (2029), but only if the compounding inherent in the exponential model were to continue unabated. Conversely, if the linear model holds, the 4000 kg ha1 milestone would occur in a much later year (2064).
|
Notwithstanding the arguable errors of omission and commission that are inherent in making long-term soybean yield projections, sustaining or increasing annual rates of yield improvement will require technological developments and innovations, and the latter must occur at annualized rates sufficient to support the yield improvement projected by whichever model in Fig. 9 that the reader might find most appropriate. Such technologies will arise from basic and applied, public and private, research at rates proportional to funding and effort. Policy-makers would be wise not to underestimate the importance of this relationship. Their current and future policies will determine which path (in Fig. 9) crop yield improvement will take over the course of the first 100 yr of the next millennium.Slafer 1993
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. B. Egli Comparison of Corn and Soybean Yields in the United States: Historical Trends and Future Prospects Agron. J., May 7, 2008; 100(Supplement_3): S-79 - S-88. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Beebe, I. M. Rao, C. Cajiao, and M. Grajales Selection for Drought Resistance in Common Bean Also Improves Yield in Phosphorus Limited and Favorable Environments Crop Sci., March 19, 2008; 48(2): 582 - 592. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. M. Bastidas, T. D. Setiyono, A. Dobermann, K. G. Cassman, R. W. Elmore, G. L. Graef, and J. E. Specht Soybean Sowing Date: The Vegetative, Reproductive, and Agronomic Impacts Crop Sci., March 19, 2008; 48(2): 727 - 740. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Kumudini, J. Omielan, and D. J. Hume Soybean Genetic Improvement in Yield and the Effect of Late-Season Shading and Nitrogen Source and Supply Agron. J., February 29, 2008; 100(2): 400 - 405. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M.-F. Johnson, R. R. Allmaras, and D. C. Reicosky Estimating Source Carbon from Crop Residues, Roots and Rhizodeposits Using the National Grain-Yield Database Agron. J., April 11, 2006; 98(3): 622 - 636. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. R. Cober, M. J. Morrison, B. Ma, and G. Butler Genetic Improvement Rates of Short-Season Soybean Increase with Plant Population Crop Sci., May 6, 2005; 45(3): 1029 - 1034. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. C. Simbahan, A. Dobermann, and J. L. Ping Screening Yield Monitor Data Improves Grain Yield Maps Agron. J., July 1, 2004; 96(4): 1091 - 1102. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Smalley, W. R. Fehr, S. R. Cianzio, F. Han, S. A. Sebastian, and L. G. Streit Quantitative Trait Loci for Soybean Seed Yield in Elite and Plant Introduction Germplasm Crop Sci., March 1, 2004; 44(2): 436 - 442. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. J. Boote, J. W. Jones, W. D. Batchelor, E. D. Nafziger, and O. Myers Genetic Coefficients in the CROPGRO-Soybean Model: Links to Field Performance and Genomics Agron. J., January 1, 2003; 95(1): 32 - 51. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Zhou, T. E. Carter Jr., Z. Cui, S. Miyazaki, and J. W. Burton Genetic Diversity Patterns in Japanese Soybean Cultivars Based on Coefficient of Parentage Crop Sci., July 1, 2002; 42(4): 1331 - 1342. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Yuan, V. N. Njiti, K. Meksem, M. J. Iqbal, K. Triwitayakorn, My. A. Kassem, G. T. Davis, M. E. Schmidt, and D. A. Lightfoot Quantitative trait loci in Two Soybean Recombinant Inbred Line Populations Segregating for Yield and Disease Resistance Crop Sci., January 1, 2002; 42(1): 271 - 277. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. R. Wilcox Sixty Years of Improvement in Publicly Developed Elite Soybean Lines Crop Sci., November 1, 2001; 41(6): 1711 - 1716. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. Cui, T. E. Carter Jr., J. W. Burton, and R. Wells Phenotypic Diversity of Modern Chinese and North American Soybean Cultivars Crop Sci., November 1, 2001; 41(6): 1954 - 1967. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.E. Specht, K. Chase, M. Macrander, G.L. Graef, J. Chung, J.P. Markwell, M. Germann, J.H. Orf, and K.G. Lark Soybean Response to Water: A QTL Analysis of Drought Tolerance Crop Sci., March 1, 2001; 41(2): 493 - 509. [Abstract] [Full Text] [PDF]
|
|
| ||||||
ween the two treatment means
