| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Agronomy, Purdue Univ., West Lafayette, IN 47907-1150
* Corresponding author (tvyn{at}purdue.edu)
Received for publication October 17, 2001.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Although vertical soil K stratification with no-till management generally does not affect soil-test K levels at the 0- to 15-cm depth, it causes plant K uptake to be more dependent on soil K and root system characteristics in the surface layer. This may reduce plant K uptake, and thus increase the likelihood of K deficiency in crop tissues as well as yield loss in growing seasons when drought occurs, since soil K availability and root growth and activity in the surface layer are more vulnerable to drought stress than those in deeper layers. In addition, presence of crop residue at the soil surface in no-till usually results in higher soil moisture and lower soil temperature in the surface layer, which may reduce soil K availability and restrict root growth early in the season (Barber, 1971; Fortin, 1993). The risks of reduction in plant K uptake by drought or low temperature in no-till fields become severe when soil K concentrations in deeper layers are too low to optimize plant K uptake. Subsurface placement of K fertilizer, therefore, may improve applied K availability and reduce soil K stratification in no-till systems.
Since the area of no-till soybean in North America has increased markedly since the late 1980s, new questions have been raised concerning the effectiveness of applying K management systems, which were originally designed for conventional-till (i.e., moldboard plow) soybean, to no-till soybean production.
Information is limited concerning K fertility management for soybean on no-till fields. Recent investigations in Iowa (Borges and Mallarino, 2000; Buah et al., 2000) showed that no-till soybean yield response to K fertilization was rarely significant on optimum- to very high–testing soils. Research in Illinois (Vasilas et al., 1988) also reported that no-till management did not affect K fertilizer recommendations for soybean compared with a traditional tillage system based on a moldboard plow.
Although subsurface placement of K fertilizer seems most likely to be superior to surface broadcasting for no-till soybean on low-testing soils, soybean response to subsurface K placement has been inconsistent in the limited studies reported to date. Hairston et al. (1990) showed that deep banding (15-cm depth) of K fertilizer resulted in significantly higher yield of no-till soybean than surface broadcasting of K on some Mississippi soils with low K levels. Research in Ohio (Hudak et al., 1989) demonstrated that K fertilizer placement did not affect no-till soybean yield on medium K soils. Borges and Mallarino (2000) reported that both deep-banded and planter-banded (5 cm from the row and 5 cm below the seed) K fertilizer in no-till produced slightly higher soybean yield than surface application on optimum- to very high-testing soils in Iowa, and that positive yield response to banding occurred relatively independent of soil-test K levels or degree of soil K stratification within the ranges encountered in their studies. However, other research in Iowa (Buah et al., 2000) as well as Illinois (Ebelhar and Varsa, 2000) showed no advantage of starter banded K, relative to surface broadcast K, for no-till soybean yield on medium- to high-testing soils. No information is available about the effects of starter K fertilizer alone on no-till soybean in low-testing soils.
Since no-till soybean in North America is predominantly grown in row widths <40 cm, yield responses to K fertilizer bands might be affected by the proximity of soybean rows to the fertilizer bands. Soybean yield generally increases as row width decreases in northern production regions (Ablett et al., 1991; Bullock et al., 1998; Ethredge et al., 1989), but the responses of narrow-row soybean to wider fertilizer bands are relatively unknown. There have been no investigations on no-till soybean responses to K placement when both fertilizer-band width and soybean row spacing have been altered in the same experiment.
Potassium management effects on soybean seed composition are largely unknown. Previous soil fertility research has mainly focused on the optimum levels or placement of nutrients for high profit or yield with little regard to soybean seed composition. Lutz et al. (1973) reported that deep placement of K fertilizer did not affect oil or protein content of soybean on a field with high soil K levels. However, since plant K deficiency reduces soybean leaf photosynthesis and phloem translocation of carbohydrates (Wallingford, 1980), the resulting reduction in photosynthate transport to soybean seed may decrease oil concentrations in seed (Gaydou and Arrivets, 1983). In addition, K deficiency has been reported to adversely affect protein synthesis in seed (Koch and Mengel, 1974, 1977).
In no-till production systems, the most effective K placement strategies for soybean yield and seed composition may be dependent on the accompanying soybean row width. The objectives of this study were to (i) evaluate the effects of K placement and soybean row width on leaf K concentrations, seed yield, and the concentrations of oil, protein, and K in no-till soybean seed on continuous no-till fields with evident soil K stratification; (ii) determine the relationships among the concentrations of oil, protein, and K in seed, seed yield, and leaf K concentrations; and (iii) examine any differential responses to K banding of soybean rows positioned above K fertilizer bands vs. those between K fertilizer bands.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
A randomized complete block split-plot design with four replicates was used in this study. Four K placement methods were randomly assigned to the whole plots, and three soybean row widths were assigned to the split-plots. Each split-plot was 21 m long and 3 m wide. Only spring-applied K was evaluated in this study. The four K placement methods were surface broadcast, 76-cm band, 38-cm band, and zero K. Surface broadcast K fertilizer was evenly broadcast applied to the soil surface. The band K fertilizer was placed 10 cm deep. When K was applied, the rate was 100 kg K ha-1 as muriate of potash (KCl). This rate was chosen based on Ontario recommendations for soil-test K ranges between 0 and 60 mg L-1 (OMAFRA, 1997). Due to resource limitations, only a single fertilizer K rate was used in this study. Soybean row widths were 76, 38, and 19 cm. Soybean "OAC Bayfield" was planted no-till in each season. The planting dates were 19 May 1998, 14 May 1999, and 26 May 2000. The target seeding rates were 400 000 seeds ha-1 for 76- and 38-cm rows, and 600 000 seeds ha-1 for 19-cm rows. The final soybean population for each treatment was >310 000 plants ha-1 for all three widths. Soybean was grown using the recommended no-till production practices in Ontario and weed control was excellent.
Composite soil samples (10 cores per sample, 2.5-cm in diameter) were collected at four depth intervals (0–5, 5–10, 10–20, 20–30 cm) randomly from each split-plot in spring before treatment application. After soil samples were air-dried, ground to pass through a 2-mm sieve, and thoroughly mixed, 1.0 mL soil was placed into a 50-mL flask, and 10 mL of 1 M ammonium acetate (NH4OAc) buffered at pH of 7.0 was added into soil; the resulting solution was shaken for 15 min and filtered (Bates and Richards, 1993). Potassium in the filtrates was determined by atomic absorption spectroscopy. Boundaries of soil-test K at low, medium, high, very high, and excessive categories for soybean are <61, 61–120, 121–150, 151–250, and >250 mg L-1, respectively (OMAFRA, 1997).
A soybean biomass sample was collected at initial flowering stage (R1) in mid- to late July from each plot in 1999 and 2000 (Fehr, 1971). No such sampling was conducted in 1998 because of resource limitations. Each sample consisted of all aboveground plant matter from a total of 2.0-m length of soybean row per split-plot. Biomass samples were oven-dried at 65°C for at least 3 d before weighing.
A leaf sample consisting of 20 most recently fully developed trifoliate leaves (petiole included) of soybean was collected at the initial flowering stage in mid- to late July from each split-plot in all three seasons for the determination of tissue K concentrations. Leaf samples were dried in a forced-air oven at 65°C for at least 3 d and then ground in a Wiley mill (Arthur K. Thomas Co., Philadelphia, PA) to pass a 1-mm sieve. Potassium in leaves was analyzed using a dry ash method (Miller, 1998).
After soybean reached physiological maturity, yield was determined by using a plot combine to harvest a central strip 1.0 m wide for the entire plot length from each split-plot. Yield was adjusted to a moisture content of 130 g kg-1. Seed samples were taken at harvest for the determination of oil, protein, and K concentrations. Seed samples were analyzed for K using the same procedures as those for leaf samples. Seed oil concentrations were determined using GrainSpec (Foss Electric, Great Britain), a near infrared reflectance spectroscopy calibrated with gravimetric method. Seed protein concentrations were measured using the same equipment as for seed oil calibrated by Kjeldahl (N x 6.25). Daily rainfall and air temperature were recorded during the entire growing season at the experimental site each year.
Data were analyzed using an analysis of variance appropriate for a randomized complete block split-plot design. Separate analyses of variance were performed for each measurement. Mean separations were accomplished using Fisher's protected LSD test. For means of the main treatments averaged over subtreatments, a LSD value was calculated for each measurement using the mean square of error associated with the main plot factor; the LSD value for comparing means of subtreatments across main treatments in each measurement was estimated using the corresponding mean square of error related to the subplot factor. When the K placement x row width was significant, the LSD value for comparisons of main treatment means within each subtreatment level was calculated based on the formula of Kuehl (1994), in which the standard error is a weighted combination of the two mean squares for error. Pearson product-moment correlation coefficients were calculated to describe the relationships between leaf K concentrations, seed yield, and concentrations of oil, protein, and K in seed on a split-plot basis each season.
Visual K deficiency symptoms in soybean rows positioned between K fertilizer bands were observed in July and August 1998, whereas soybean rows positioned over the K fertilizer bands did not show any K deficiency symptoms. This phenomenon prompted investigations on the effects of soybean row position, relative to K fertilizer bands, on soybean responses and soil K fertility in 1999 and 2000. Three treatment combinations were selected to conduct this investigation: 19 cm row-width soybean following 76-cm banded K, 38-cm row-width soybean following 76-cm banded K, and 19-cm row-width soybean following 38-cm banded K. In each case, plant and soil measurements were taken from the row seeded above the K fertilizer band as well as from the row positioned midway between the K bands. Soil samples were taken after soybean physiological maturity (10 Sept. 1999 and 11 Oct. 2000), but before soybean harvest. Grain samples were obtained by hand harvesting 8.0-m lengths of soybean rows just before the bulk combine harvest for the entire experiment. Seed yield was adjusted to a moisture content of 130 g kg-1. Soil K, leaf K, and seed K concentrations were determined using the same procedures described before.
The data of soil K, biomass, leaf K, yield, and seed K concentrations arising from the respective row positions were analyzed using an analysis of variance appropriate for a randomized complete block split-plot design. The three treatment combinations and soybean row positions relative to K fertilizer bands were assigned to the whole plots and split-plots, respectively. If interactions between the latter whole plots and split plots were insignificant for a specific parameter, only row position (i.e., split plot) means across the three treatment combinations were presented. The Bartlett tests of experiment error homogeneity were conducted across crop years (1999 and 2000) for each measurement to determine whether the results needed to be discussed separately for each year, or combined across years. Data for a specific measurement were combined over years if the experimental errors associated with main treatments and subtreatments were both homogeneous across years.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Initial Soil Potassium Fertility
Initial soil-test K concentrations (0–15 cm depth) were low (<61 mg L–1) for all three seasons according to the Ontario soil-test K interpretations (Table 1). Vertical soil K stratification was evident in each season; soil K concentrations in the surface 0- to 5-cm layer was approximately 1.8 to 2.5 times higher than K levels present at the 10- to 20-cm depth (Table 1). Soil-test K concentrations decreased sharply with soil depth; this phenomenon is common in no-till management (Ketcheson, 1980; Vyn and Janovicek, 2001; Yin and Vyn, 2002a).
Leaf Potassium Concentrations
Leaf K concentrations in 1998 were influenced by the K placement x row width interaction (Table 2). In 76-cm row width, K applications increased leaf K concentrations by 5.1 to 8.0 g kg-1, relative to zero K, regardless of placement methods. In 38-cm rows, significant gains in leaf K only occurred after banded K treatments. In 19-cm rows, only surface broadcasting of K resulted in significant increases in leaf K compared with zero K. The 1998 results showed that banded K at 76- or 38-cm intervals was more effective in increasing soybean leaf K in 76- and 38-cm rows than those in 19-cm rows, and that 38-cm banded K was more effective than surface broadcast K in increasing leaf K concentrations in 76- and 38-cm row widths. Banding was not beneficial relative to broadcasting for soybean in 19-cm rows since at least half the rows were positioned 15 cm or more from the fertilizer bands.
|
In 2000, the interaction between K placement and row width was not significant (Table 2). Leaf K concentrations were significantly increased by 2.7 to 3.6 g kg-1 with K applications compared with zero K. Furthermore, no significant differences were observed between banded placement (38 or 76 cm apart) and surface broadcasting.
Increases in leaf K concentrations by K applications were much greater in 1998 and 1999 than in 2000; the latter suggested that leaf K concentrations may be more related to soil-test K levels and weather conditions than to the extent of K stratification. Both growing-season rainfall (data not presented) and soil-test K concentrations in 2000 were greater than those in 1998 and 1999. The rainfall in June was 218 mm in 2000, approximately twice that received in 1998 and 1999. However, we did not anticipate leaf K concentrations from most zero K plots exceeding 25.0 g kg-1 in 2000 when the soil-test K was only 54 mg L-1 before soybean planting. High soil moisture levels in June of 2000 may have greatly increased soil K and applied K availability, and thus enhanced leaf K concentrations.
Small and Ohlrogge (1973) suggested that the range of adequate K concentrations for soybean is 17.1 to 25.0 g kg-1 in trifoliate leaves during the initial flowering stage. Thus, soybean from the zero K plots had leaf K levels below the adequate range in 1998, but leaf K concentrations in K-fertilized plots were adequate. In 1999, leaf K concentrations from zero K and surface broadcast treatments were far below the sufficiency range, while soybean in both banded K treatments had adequate leaf K levels—unless they were planted in 19-cm rows and received K fertilizer with the 76-cm banding method. In 2000, leaf K concentrations from all treatments (including zero K) were much greater than 17.1 g kg-1. If these mid-season leaf K nutrition values were indicative, significant seed yield responses to K fertilizer applications were expected in 1998 and 1999, but not in 2000.
Seed Yield
When the 3-yr yield data were combined, a significant interaction of K placement x row width was observed (Table 3). In the 76-cm row width, only 76-cm banded K resulted in a significant yield increase (10%) compared with zero K. In the 38-cm row width, both banded K treatments increased yield 11 to 15% relative to zero K. However, yield responses to banded K were not significant for soybean in 19-cm rows. The latter phenomenon was probably associated with the low proportion of 19-cm soybean rows positioned close to the respective K fertilizer bands. Surface broadcasting of K fertilizer never increased yield significantly in any row width. Yield differences between banded K treatments were insignificant. Trends of seed yield responses to K placement and row width were very similar to those of midseason biomass yield responses (data not presented).
|
Seed Potassium Concentrations and Accumulation
Seed K concentration responses to K placement were not affected by row width in either 1998 or 1999 (Table 4). For both seasons, seed K concentrations were increased by 1.3 to 2.7 g kg-1 with K applications. However, both banded K treatments resulted in significantly higher seed K concentrations than surface broadcast K. In the 2000 season, K placement effects differed with row width (Table 4), but K applications still increased seed K concentrations by 0.4 to 1.4 g kg-1 in each row width. Although seed K concentrations have previously been observed to increase by K fertilization (Coale and Grove, 1991; Terman, 1977; Yin and Vyn 2002b), this study provided the first evidence that banded K was superior to surface broadcast K in increasing seed K concentrations in certain growing seasons. Higher seed K concentrations may have significant influence on other seed attributes (Vyn et al., 2002) since K is widely involved in plant metabolic activities as an enzyme activator.
|
Oil and Protein Concentrations
Both banded K treatments significantly increased oil concentrations, while surface broadcast did not increase oil levels relative to zero K (Table 5). Soybean in 38-cm row width had significantly higher oil concentrations than those in the 76- or 19-cm rows, but the K placement x row width interaction was not significant. Potassium applications resulted in lower protein concentrations, regardless of K placement methods when the results were averaged over 3 yr (Table 5). Banded K (38 or 76 cm apart) decreased protein concentrations more than surface broadcast K. However, protein yield was not reduced by K fertilization (data not presented) since the increases in seed yield were much greater than the decreases in protein concentrations.
|
Oil concentrations had highly significant and positive correlations with leaf K, seed K concentrations, and seed yield in 1998 and 1999 (Table 6). However, in the 2000 season, oil levels were negatively associated with seed yield. Consistently negative correlations of protein concentrations with leaf K and seed K concentrations were observed in all three seasons (Table 6). Protein concentrations were also negatively correlated with seed yield and seed oil concentrations in 1998 and 1999, but not in 2000.
|
Effects of Soybean Row Position on Response Parameters
Soybean biomass, leaf K, and yield responses to soybean row position relative to K fertilizer bands were all significant when the data were combined across years (Table 7). Midseason growth, K nutrition, and seed yield of soybean positioned over K fertilizer bands were significantly improved compared with those located between K fertilizer bands on this low-testing soil. Seed K concentrations also averaged 1.0 g kg-1 higher for soybean seeded above the K fertilizer bands (data not presented). Maximum soybean benefit from deep banding of K fertilizer, therefore, would result from systems where the width of K fertilizer bands equals the intended soybean row width so that all soybean rows were in close proximity to K fertilizer bands.
|
Utilization of a lower K fertilizer rate may have revealed more interactions between K placement and row width on soybean response parameters such as leaf K, seed yield, and seed composition. In addition, our conclusions about merits of banding vs. broadcasting might have been somewhat different at lower rates of K fertilizer [such as those utilized recently in Iowa by Borges and Mallarino (2000)]. Nevertheless, the K rate chosen for this study was not unreasonable given the low soil-test K.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  F. G. Fernandez, S. M. Brouder, C. A. Beyrouty, J. J. Volenec, and R. Hoyum Assessment of Plant-Available Potassium for No-Till, Rainfed Soybean Soil Sci. Soc. Am. J., June 18, 2008; 72(4): 1085 - 1095. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X. Yin and T. J. Vyn Relationships of Isoflavone, Oil, and Protein in Seed with Yield of Soybean Agron. J., August 17, 2005; 97(5): 1314 - 1321. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
X. Yin and T. J. Vyn Critical Leaf Potassium Concentrations for Yield and Seed Quality of Conservation-Till Soybean Soil Sci. Soc. Am. J., September 1, 2004; 68(5): 1626 - 1634. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
