|
|
||||||||
a Dep. of Soil, Water, and Climate, Univ. of Minnesota, St. Paul, MN 55108
b Southern Research and Outreach Center, Univ. of Minnesota, Waseca, MN 56093
c Dep. of Agronomy and Plant Genetics, Univ. of Minnesota, St. Paul, MN 55108
* Corresponding author (mschmitt{at}soils.umn.edu)
Received for publication July 27, 2000.
| ABSTRACT |
|---|
|
|
|---|
Abbreviations: DM, dry matter
| INTRODUCTION |
|---|
|
|
|---|
Preplant N application for soybean has given limited or inconsistent effects on seed yield. Sorensen and Penas (1978) reported seed yield increases with preplant N at 9 of 13 sites in a 3-yr Nebraska study. Yield increases were associated with more favorable yield environments as well as increased seed size. These researchers concluded that increased lodging and disease potential with preplant N application times limited positive yield increases. A primary factor reported for predicting a yield response is soil inorganic N concentrations early in the growing season. Positive increases in soybean yields from supplemental N applications are a function of soil N concentrations (Stone et al., 1985). They concluded that fertilizer N increased soybean yields only if preplant soil NO3N concentrations were <190 kg ha-1 in the top 180 cm. Al-Ithawi et al. (1980) measured significant seed yield increases at all three sites in one year and only at a moisture-limiting site in a second year. They attributed the yield response to inadequate residual soil NO3N at planting for the responding sites in Nebraska. In addition to soil N status, Lyons and Earley (1952) suggested that soybean response to either plowdown or topdressed N is dependent on rainfall and temperature during the growing season because these environmental factors influence sufficiency of symbiotically fixed N.
Seed yield increases were not obtained with N rates up to 448 kg ha-1 in Illinois (Johnson et al., 1975). In Iowa, preplant N applications up to 270 kg ha-1 did not affect seed yields (Bharati et al., 1986). Beard and Hoover (1971) also found no soybean seed yield response to a series of N rates under irrigation in California. No fertilizer N was recommended for dryland soybean on Coastal Plain soils near planting or at flowering, based on research in which only 1 of 10 sites responded to N applications (Reese and Buss, 1992). Deibert et al. (1979) also reported no statistical soybean seed yield increases from N management scenarios that included three N rates applied either near planting or at full bloom. Although seed yield increases due to N fertilization practices were measured at five of seven siteyears in Alabama, the researchers concluded fertilizer N application to soybean is risky because yield responses were very inconsistent with regard to N rate and timing treatments (Wood et al., 1993).
Despite the lack of consistent positive yield responses, the concept of applying N to soybean late in the growing season has scientific merit. Shibles (1998) states that N2-fixing capacity begins to decline rapidly after growth stage R5, which is approximately the same time as peak N demand for protein synthesis. Thus, adding fertilizer N to soybean later in the growing season can extend NO3 reductase activity beyond R5 (Shibles, 1998). Afza et al. (1987) concluded that inadequate N supply during pod-filling limited soybean seed yields and that soil-applied fertilizer N judiciously applied during the late stages of growth can enhance seed yields. In a study evaluating time and methods of fertilizer N applications in Austria, Afza et al. (1987) reported that N applied during pod-fill resulted in a 37% increase in seed yield compared with the control, which had 20 kg N ha-1 applied as a starter. Late-season N applications to soybean resulted in significantly increased yields at six of eight siteyears in Kansas (Wesley et al., 1998). They concluded that yield response is positively correlated with higher yielding sites and that supplemental N should be considered at the R3 growth stage for irrigated soybean. Lawn and Brun (1974) measured positive responses to supplemental N applied at late-flowering, a time that nodule activity was previously found to decline. While only one of the two cultivars they grew produced a seed yield increase, seed protein content was measured in both.
Late season N application appears to negate some of the liabilities and uncertainties that early-season applications present. Wallace et al. (1990) concluded that late-season N applications are not easily achieved, yet potential seed yield increases may be more likely with determinate soybean because the N would not directly contribute to lodging and disease issues. The combination of lodging and disease issues are directly influenced by the stimulation of vegetative growth early in the growing season. By delaying N application until the latter stages of soybean growth, it is hypothesized that more of the N will contribute to pod-fill rather than vegetative growth (Afza et al., 1987).
The acceptance of refined, custom fertilizer application management strategies and the demonstrated potential for increasing soybean yield in certain environments encourages producers to question the decades-old debate on N fertilization of soybean. The objective of our research was to evaluate several in-season N fertilization strategies on the seed yield response of soybean. A secondary objective was to determine if fertilizer N additions significantly affect plant N accumulation, seed N removal, plant N concentration, soil NO3N concentrations, and seed protein and oil composition.
| MATERIALS AND METHODS |
|---|
|
|
|---|
|
|
|
|
All soybean plots were seeded with high-yielding cultivars appropriate for the location and managed with optimum cultural practices (i.e., plant densities and planting dates). Row width at all locations was 76 cm. Weeds were controlled with labeled rates of herbicides at all sites. Seed yields were taken at physiological maturity by harvesting the center two rows, which were end-trimmed to 7.6 to 12.2 m, with a small-plot combine. Plot size was 3 m wide by 9.1 to 15.2 m long. Seed moisture content was determined, and yields were expressed on a dry matter basis. Seed samples from the harvest plots were analyzed for protein and oil content as whole seeds using a NIRS Systems (Silver Springs, MD) Model 6500. Values for protein and oil content are expressed on a 13% moisture basis.
Total aboveground plant samples were collected at the R6 stage during the first week of September (before leaf drop began) to estimate maximum N and DM accumulation by the plants. Within each four-row plot, two 60-cm lengths of row were collected from the middle rows. These samples were weighed, dried, ground, digested, and then colorimetrically analyzed for total N (Bremner and Mulvaney, 1982). Soil samples were collected from the top 30 cm of soil at the same time as the plant samples. Eight cores were collected and composited from the area where the plant samples were removed in each plot. A systematic sampling scheme was used that consisted of selecting two random areas within the plot and collecting four vertical soil cores taken at 19-cm intervals on a transect perpendicular to the direction of the plant rows and possible fertilizer application bands. All soil samples were dried, ground to pass a 2-mm sieve, and analyzed for NO3N using colorimetric methods with Cd reduction.
Analysis of variance was conducted for experiments combined over locations for each of the dependent variables of seed yield, total DM yield at R6, plant N concentration at R6, N accumulation at R6, soil NO3N at R6, estimated seed N removal, seed protein, and seed oil content (SAS Institute, 1996). Single degree of freedom contrasts were used to compare the control with the fertilized treatments and the treatment factors of application time, placement methods, and N source. In addition, correlation methods were used to evaluate the relationship between soil NO3N and seed yield. Individual siteyear statistics are presented for seed yield and soil NO3N at R6, which was the only variable with a statistically significant location x treatment interaction.
| RESULTS AND DISCUSSION |
|---|
|
|
|---|
Seed Yield
Statistical probability for significance of 10% was chosen for the treatment source of variation when using the 12 siteyear combined analysis of variance (Table 2). Treatment means ranged from 3.32 to 3.43 Mg ha-1, a range of 3%. Using single degree contrasts to further compare treatment factors indicated the control treatment was different than the fertilized treatments at the 0.10 level, method of application was insignificant, and urea source and timing were significant at the 0.05 and 0.07 level, respectively. Because the statistical probabilities for seed yield are very close to the commonly accepted 0.10 significance level, individual siteyear data and statistics were conducted for seed yield.
Whereas the treatment source of variation was statistically significant (0.10 probability) in the combined analysis, it was not statistically significant at any of the 12 siteyears (Table 3). Although all siteyear coefficients of variation were less than 10% for seed yield, combining the 12 siteyears increased this level of precision, which decreased the treatment mean differential necessary for statistical significance at any given probability level. In the combined analysis, both the form of urea and the application timing were significant treatment variables at the 0.10 probability level, according to single degree-of-freedom contrasts (Table 2). Nonetheless, the combined mean for the standard urea was 3.38 Mg ha-1 and the mean for the polymer-coated urea was 3.41 Mg ha-1, an increase of less than 1%. Three of the 12 siteyears had significant differences between the two urea sources, with poly-coated urea having greater yields at Sites A and L; whereas yields were greater with standard urea at Site B (Table 3). Application time differences were significantly significant at 2 of 12 siteyears, with the overall treatment means being the same for the July and August application comparison (3.38 Mg ha-1). Broadcast applications in August resulted in greater yields (0.290.53 Mg ha-1) at Sites C and G compared with July applications.
Regardless of the mixed statistical interpretation with these results, the practical interpretation is that, for the majority of the trials, there is little to no impact of fertilizer N on seed yield. Any seed yield increases, whether from the overall analysis or an individual siteyear, must be tempered by economic and logistical risk management. These results generally concur with those of several others (Deibert et al., 1979; Reese and Buss, 1992), yet do not necessarily contradict other results that involve different conditions, primarily irrigated agriculture (Wesley et al., 1998; Beard and Hoover, 1971).
Forage Yield, Nitrogen Accumulation, and Seed Nitrogen Removal
Total DM yield at R6, the growth stage considered to represent maximum forage dry matter before leaf drop begins, was not affected by any of the N fertilizer treatments. Total N concentration in the forage at R6 ranged from 29.4 g kg-1 (control treatment) to 30.3 g kg-1 but was not affected by any of the treatments (Table 2). Total N accumulation at R6, calculated as the product of forage yield and forage N concentration, ranged from 224 to 229 kg ha-1 and was not statistically significant. These findings directly contrasted those of Hanway and Weber (1971), who reported an increase in N accumulation with the addition of N fertilizer.
Seed N removal was significantly affected by all treatment factors (Table 2). Seed N removal was calculated by multiplying seed DM yield by seed N concentration, which was approximated by converting seed protein to N concentration. While seed N removal was significant in the combined analysis, the statistics of the variables comprising this data (seed yield and protein) should be interpreted independently. Nitrogen removal, which was very consistent among the treatments, ranged from 172 to 178 kg ha-1. This result was similar to those reported by Varvel and Peterson (1992) who reported 150 to 200 kg ha-1 N removed regardless of N fertilizer applications.
Seed Protein and Oil
Seed protein content was significantly affected by all treatment factors in this study (Table 2). Protein content averaged 372 g kg-1 for the control compared with the maximum of 376 g kg-1 for the August treatment (a 1.2% increase). The knifed treatments resulted in a 0.7% increase in protein compared with the broadcast treatment. The poly-coated urea also resulted in an additional 0.7% increase compared with standard urea. Seed oil content ranged from 195.3 to 196.3 g kg-1 and was not affected statistically by the treatments (Table 2). This research corroborates the conclusions of several researchers (Deibert et al., 1979; Wood et al., 1993; Weber, 1966)additional fertilizer N does not affect oil content of normal nodulating soybean. It should be noted that both seed protein and oil data have very low coefficients of variation (1.3 and 2.1%, respectively) in this study.
Soil Nitrate-Nitrogen
Soil NO3N concentrations were initially measured in late May, after soybean emergence and uniform stands were established. Soil NO3N concentrations ranged from 5.0 to 14.8 mg kg-1 (Table 1). Nitrate N concentrations are not normally measured for soybean crops; thus, these early season measurements cannot be categorized into formal sufficiency ranges. However, these concentrations are typical of soil NO3N concentrations measured from corn fields at the same time of year in Minnesota. Comparisons of early-season NO3N concentrations with late-season concentrations will not be made because the N mechanisms that both contribute and remove NO3N during the season were not measured in this study. Relative comparisons of the treated plots with the untreated control provide meaningful information on the impact of fertilizer N applications during late-season podfill.
Soil NO3N concentrations when the soybean plants were at the R6 stage were significantly affected by all treatment variables in this study (Table 2). Compared with the overall control treatment mean (5.7 mg kg-1), the addition of 84 kg ha-1 applied in-season almost doubled the amount of soil NO3N at R6 (10.5 mg kg-1). Knife-injection of the fertilizer N increased soil NO3N by 33% compared with broadcast application. Polymer-coated urea resulted in 15% more NO3N compared with standard urea. This is expected, as the polymer coating is designed to delay hydrolysis of the urea, thus slowing N release and providing greater NO3N concentrations later in the season. Delaying fertilizer application from July (7.3 mg kg-1) to August (14.0 mg kg-1) significantly increased soil NO3N at R6. These combined data indicate that differences in soil NO3N concentrations were present during the pod-filling period as a result of the treatments selected.
The high coefficient of variation for soil NO3N (31.7%) illustrates the traditional variability in soil N measurements even though a systematic sampling scheme was used to account for band applications of fertilizer N. Also, there is much variation in the data, as indicated by the coefficients of variation at each location (Table 4). Evaluation of individual siteyear data shows a range in absolute amounts of soil NO3N (Table 4). Averaged across all treatments, SiteYears A and H averaged 14 mg kg-1 NO3N, whereas SiteYear J averaged only 6.6 mg kg-1. SiteYears E and I only had a range of 3.5 mg kg-1 NO3N between the control and the greatest treatment; whereas SiteYear H had a 16.5 mg kg-1 difference.
At three of the individual siteyears (E, I, and L), there were no statistical differences in soil NO3N concentrations (Table 4). At two of these siteyears (E and I), very low soil NO3N concentrations were measured in all plots, including the control, most likely indicating loss of inorganic N was great. Except for these three nonsignificant siteyears, application of fertilizer N consistently increased soil NO3N. At six of nine significant siteyears, knife applications of N resulted in significantly greater soil NO3N. At only three siteyears did the polymer-coated urea result in significant soil N differences, with the greatest effect at SiteYear F where a 44% increase was measured compared with standard urea. The other two sites had soil NO3N increases of 34 and 23%. At seven siteyears, application of fertilizer N in August provided greater NO3N in early September than July application. Because the August applications of standard urea were topdress applied, volatilization may have occurred that would reduce soil NO3N concentrations.
| SUMMARY |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
D. A. Ruiz Diaz, P. Pedersen, and J. E. Sawyer Soybean Response to Inoculation and Nitrogen Application Following Long-Term Grass Pasture Crop Sci., May 11, 2009; 49(3): 1058 - 1062. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. L. Osborne and W. E. Riedell Starter Nitrogen Fertilizer Impact on Soybean Yield and Quality in the Northern Great Plains Agron. J., October 3, 2006; 98(6): 1569 - 1574. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. J. Temperly and R. Borges Tillage and Crop Rotation Impact on Soybean Grain Yield and Composition Agron. J., June 5, 2006; 98(4): 999 - 1004. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. W. Barker and J. E. Sawyer Nitrogen Application to Soybean at Early Reproductive Development Agron. J., March 1, 2005; 97(2): 615 - 619. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. W. Singer, K. A. Kohler, M. Liebman, T. L. Richard, C. A. Cambardella, and D. D. Buhler Tillage and Compost Affect Yield of Corn, Soybean, and Wheat and Soil Fertility Agron. J., March 1, 2004; 96(2): 531 - 537. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||