Introduction
Dicamba is a synthetic auxin herbicide that has been extensively used to control broadleaf weeds in corn, small grains, and pasture crops since the 1960s (Al-Khatib and Peterson Reference Al-Khatib and Peterson1999). Widespread use of dicamba often resulted in off-target movement and injury to susceptible plants, including soybean and cotton (Auch and Arnold Reference Auch and Arnold1978; Wax et al. Reference Wax, Knuth and Slife1969). Introduction of dicamba-resistant (DR) soybean and cotton technology in 2017 increased the application of dicamba in the vicinity of actively growing, susceptible soybean. Various injury symptoms, including epinasty of petioles and shoots, leaf malformation, terminal bud chlorosis, malformed pods, and delayed maturity are associated with off-target dicamba movement to non-DR soybean (Solomon and Bradley Reference Solomon and Bradley2014; Wax et al. Reference Wax, Knuth and Slife1969). Less injury is often observed in soybean exposed to dicamba during vegetative stages compared with early reproductive stages (Auch and Arnold Reference Auch and Arnold1978; Wax et al. Reference Wax, Knuth and Slife1969). In addition, soybean during vegetative stages exposed to dicamba doses high enough to cause 30% injury were not likely to cause yield reductions greater than 5%; however, exposure during reproductive stages to doses high enough to cause 12% injury were likely to cause greater than 5% yield loss (Kniss Reference Kniss2018).
Soybean injury and yield loss are influenced by the dicamba dose and exposure timing, but there is also speculation that soybean cultivars may respond differently to dicamba exposure (Auch and Arnold Reference Auch and Arnold1978). However, minimal peer reviewed literature exists comparing varieties. Auch and Arnold (Reference Auch and Arnold1978) observed no differences in yield among non-DR soybean varieties exposed to various rates of dicamba. Most available literature has explored the differences among indeterminate and determinate soybean (McCown et al. Reference McCown, Barber and Norsworthy2018; Wax et al. Reference Wax, Knuth and Slife1969; Weidenhamer et al. Reference Weidenhamer, Triplett and Sobotka1989). It is understood that whether a soybean is indeterminate or determinate may affect the way the plant respond to dicamba at different times (Wax et al. Reference Wax, Knuth and Slife1969). Weidenhamer et al. (Reference Weidenhamer, Triplett and Sobotka1989) reported more severe yield loss for indeterminate varieties at more mature growth stages compared to determinate varieties, but McCown et al. (Reference McCown, Barber and Norsworthy2018) reported that indeterminate and determinate varieties responded similarly.
Quantifying soybean yield components provides greater understanding of the impacts of dicamba exposure on soybean yield (Robinson et al. Reference Robinson, Simpson and Johnson2013). McCown et al. (Reference McCown, Barber and Norsworthy2018) and Solomon and Bradley (Reference Solomon and Bradley2014) observed greater reductions in pods per plant following dicamba application during early reproductive stages compared to vegetative stages. Similarly, Solomon and Bradley (Reference Solomon and Bradley2014) reported dicamba exposure at R2 reduced seeds per pod and nodes per square meter more than exposure at vegetative stages.
The effects of off-target dicamba movement on soybean offspring are not well understood; however, understanding the impacts is crucial for soybean seed production (Jones et al. Reference Jones, Norsworthy, Barber, Gbur and Kruger2018). Off-target dicamba movement has been shown to reduce germination, emergence, and vigor, and injure offspring of soybean treated with dicamba during reproductive stages (Auch and Arnold Reference Auch and Arnold1978; Jones et al. Reference Jones, Norsworthy, Barber, Gbur and Kruger2018; Thompson and Egli Reference Thompson and Egli1973; Wax et al. Reference Wax, Knuth and Slife1969. In addition, offspring that emerge may have reduced vigor or foliar injury symptoms following dicamba application to the parent plant during reproductive growth (Jones et al. Reference Jones, Norsworthy, Barber, Gbur and Kruger2018, Reference Jones, Norsworthy and Barber2019). Few peer-reviewed articles have described the relative responses of different soybean varieties. Therefore, the main objective of this research was to determine the response of non-DR soybean varieties with various herbicide-resistant traits when exposed to a reduced rate of dicamba at V3 and R1.
Materials and Methods
Field Experiments
Field studies were conducted at the Kansas State University Ashland Bottoms Research Farm in Manhattan, Kansas (39.12N, 96.63W) in 2018 (MHK18) and 2019 (MHK19) and in 2019 at the Kansas State University East Central Experiment Field in Ottawa, Kansas (OTT19; 38.54N, 95.25W). Seed beds were prepared using a field cultivator on the day of planting at MHK18 and MHK19, and before the early preplant herbicide application at OTT19. Soil series at MHK18, MHK19, and OTT19 are as follows: Wymore silty clay loam, Reading silt loam, and Woodson silt loam. Information regarding planting dates, in-season precipitation, and PRE herbicide treatments for early-season weed control for each site-year is summarized in Table 1. Plots were hand-weeded as needed to keep weed free throughout the growing season across all site-years.
Table 1. Planting date, total in season rainfall, and maintenance herbicide application timing, date, product, and rate used prior to crop emergence in experiments evaluating dicamba drift injury in Manhattan, KS in 2018 and 2019, and in Ottawa, KS in 2019. a
a Abbreviations: EPP, early preplant; MHK18, Manhattan, KS in 2018; MHK19, Manhattan, KS in 2019; OTT19, Ottawa, KS in 2019; PRE, pre-emergent.
b Sourced from Kansas State University (2019).
c Authority Elite (FMC Corporation, Philadelphia, PA) at 0.7 L ha−1.
d Authority Maxx (FMC Corporation, Philadelphia, PA) at 0.49 kg ae ha−1.
e Cinch (Corteva Agriscience, Wilmington, DE) at 1.8 and 1.2 L ha−1.
The same soybean varieties were planted across all site-years. Soybean varieties included Asgrow AG4135RR2Y® (glyphosate-resistant), Credenz 3841LL® (glufosinate-resistant), Credenz 4748LL® (glufosinate-resistant), and Stine 40BA02® (glyphosate and isoxaflutole-resistant). Details regarding maturity group, herbicide traits, and company are presented in Table 2. Soybean were planted at approximately 308,880 seeds ha−1, 3.8 cm deep, and in 76.2-cm rows using a 4-row row crop planter across all site-years.
Table 2. Soybean varieties planted in Manhattan, KS in 2018 and 2019, and in Ottawa, KS in 2019 with corresponding herbicide traits, maturity groups, abbreviations, and companies.
a Bayer Crop Science, St. Louis, Missouri.
b BASF Agriculture, Florham Park, New Jersey.
c Stine Seed Company, Adel, Iowa.
The N,N-Bis-(3-aminopropyl) methylamine (BAPMA) salt of dicamba (Engenia®, BASF Corp., Research Triangle Park, NC) was applied to all varieties at V3 or R1 (Fehr and Caviness Reference Fehr and Caviness1977). Dates of dicamba application and environmental conditions during application for each site-year are presented in Table 3. Dicamba was applied at 5.6 g ae ha−1 (1/100th the field-use rate). A split-plot design with four replications was used. Soybean variety was the main plot, and the application timing was randomly assigned to subplots. Each main plot included a nontreated check. Individual plots were 3 m by 9 m in size.
Table 3. Application date and meteorological conditions during all application timings in experiments evaluating dicamba injury in 2018 and 2019.
a Abbreviations: MHK18, Manhattan, KS in 2018; MHK19, Manhattan, KS in 2019; OTT19, Ottawa, KS in 2019.
Spray solution was applied directly to plots with 140 L ha−1 spray volume using a CO2 powered backpack sprayer and a 4-tip, 1.9-m hand-held boom equipped with TTI110015 nozzles (TeeJet Technologies, 1801 Business Park Dr, Springfield, IL 62703) at 220 kPa. The center two rows of each plot received the full rate, whereas the two outer rows acted as a buffer between treatments. Evaluations were conducted from 2 wk after treatment (WAT) through progeny seed analysis.
Soybean injury was visually assessed 2 and 4 WAT and at the onset of senescence (R7). Soybean plants were evaluated on a 0% to 100% crop injury scale with 0% indicating no injury and 100% indicating plant death. Symptomology at lower injury levels included leaf cupping, leaf crinkling, and chlorosis of terminal buds. Symptomology at greater injury levels included the aforementioned symptomology and necrosis of terminal buds, pod malformation, and stunting (Behrens and Lueschen Reference Behrens and Lueschen1979; Sciumbato et al. Reference Sciumbato, Chandler, Senseman, Bovey and Smith2004; Wax et al. Reference Wax, Knuth and Slife1969). To further evaluate soybean response to dicamba, the heights of five randomly selected plants from the center two rows of each plot were recorded at the onset of senescence. Soybean yield component data were collected at harvest from 1 m row−1 from one of the two center rows of each plot. Yield components measured included seed weight, pods per plant, seeds per pod, and main stem nodes per plant. Soybeans were harvested for grain yield from the center two rows of each plot with a small plot combine and grain moisture was adjusted to 13%.
Soybean Offspring Fitness
To test seed germination, 50-g soybean seed samples were taken from samples harvested by the small-plot combine and sent to the Seed Laboratory at the Kansas Crop Improvement for analysis. One-hundred soybean seeds were counted and placed on Kimpack (Anchor Paper Co., St. Paul, MN), which was placed on a food tray and moistened with 500 ml of tap water. Soybean seeds were then covered with 0.6 to 1.3 cm of mason sand. Trays were placed in a germination chamber at 30 C for 8 h and 20 C for 16 h with lights on during the warm cycle. After 8 d of incubation, seedlings were evaluated to determine normal, abnormal, dead, or hard seed (AOSA 2019).
To quantify response of offspring, 10 seeds from each plot sample were planted into 14-cm pots containing Miracle-Gro Moisture Control® potting mix (The Scotts Company LLC, Marysville, OH) and grown in the Kansas State University Weed Science greenhouse until V3. Pots were arranged by plot within site-year. The daytime temperature was 30 C and the nighttime temperature was 22 C. The 15-h photoperiod was supplemented with a metal-halide lighting system. Soybean were subirrigated with municipal water as need to maintain adequate moisture levels. The number of injured offspring and percent soybean injury of emerged offspring were recorded on a 0% to 100% crop injury scale with 0% indicating no injury and 100% indicating plant death when soybean reached V3. No injury was observed in soybean offspring.
Statistical Analyses
Soybean height, yield components, and yield were converted to a relative percent of the nontreated check before statistical analysis to account for differences that may naturally exist among varieties. The relative percent of the nontreated check was calculated by subtracting the plot value from the nontreated check value for the corresponding replication and dividing the difference by the nontreated check value. Raw data were visually assessed for normality and did not violate ANOVA assumptions. Soybean injury ratings, relative soybean height, relative soybean yield components, relative soybean yield, germination, offspring emergence, offspring height, number offspring injured, and soybean injury of offspring were subjected to ANOVA using the GLIMMIX procedure in SAS (SAS v.9.4, SAS Institute Inc., Cary, NC). Replication was considered a random effect and site-year, dicamba application timing, and soybean variety were considered fixed effects. Means were separated using Fisher’s protected LSD (α = 0.05). Soybean injury ratings, soybean height, soybean yield components, soybean yield, germination, offspring emergence, offspring height, number of offspring injured, and soybean injury of offspring were subjected to the CORR procedure in SAS. Pearson coefficients were considered weak if less than 0.3, moderate if greater than 0.3 but less than 0.5, and strong if greater than 0.5 (Mukaka Reference Mukaka2012). Yield and height data were subjected to the REG procedure in SAS using a linear model, which was selected as the best fit based on R 2 and P-values.
Results and Discussion
Soybean Injury
Soybean injury 2 WAT had a significant interaction among site-year, application timing, and variety (Table 4). At MHK18, injury was greater in soybean treated at V3 rather than at R1, regardless of variety. (Table 5). At MHK19, soybean similar injury was similar regardless of timing or variety, except for ST40B treated at R1, which resulted in 40% injury (Table 5). Soybean treated at R1 had greater injury than soybean treated at V3 regardless of the variety at OTT19 (Table 5). At 4 WAT, there was also a significant interaction among site-years, application timing, and variety for soybean injury (Table 4). Soybean injury was greater for soybean treated at R1 than V3 at all locations (Table 5). Reduced injury observed in CR4748 could be attributed to the longer maturity group. Greater injury observed in ST40B may be linked to the presence of two herbicide-resistant events; however, no research has investigated this hypothesis.
Table 4. Analysis of variance of fixed effects and all interactions for soybean injury as a response of different soybean varieties exposed to a reduced rate a of dicamba at varying application timings at Manhattan, KS in 2018 and 2019, and in Ottawa, KS in 2019.
a Rate = 5.6 g ae ha−1 dicamba.
b Abbreviation: WAT, weeks after treatment.
Table 5. Soybean injury at 2WAT and 4WAT as a result of varying soybean varieties exposed to dicamba a at varying application timings at Manhattan, KS in 2018 and 2019 and in Ottawa, KS in 2019. b,c
a 5.6 g ae ha−1 dicamba.
b Abbreviations: AG4135, Asgrow AG4135; CR3841, Credenz 3841LL; CR4748, Credenz 4748LL; MHK18, Manhattan, KS in 2018; MHK19, Manhattan, KS in 2019; OTT19, Ottawa, KS in 2019; ST40B, Stine 40BA02; WAT, weeks after treatment.
c Means separated within site-year and means followed by the same letter within a column are not statistically different according to Fisher’s protected LSD (α = 0.05).
Soybean injury 2 WAT at all site-years was similar to injury previously reported for similar dicamba application rates and timings (Andersen et al. Reference Andersen, Clay, Wrage and Matthees2004; McCown et al. Reference McCown, Barber and Norsworthy2018; Solomon and Bradley Reference Solomon and Bradley2014). In general, soybean treated at R1 had greater injury 4 WAT than those treated at V3. Soybean injury observed following dicamba application at V3 was similar to what Osipitan et al. (Reference Osipitan, Scott and Knezevic2019) observed at lower doses of dicamba, but was greater than injury observed by McCown et al. (Reference McCown, Barber and Norsworthy2018) when soybeans were treated at V4 with 2.2 and 8.8 g ae ha−1 dicamba. Similarly, injury following dicamba application at R1 was greater than injury observed by McCown et al. (Reference McCown, Barber and Norsworthy2018) and Jones et al. (Reference Jones, Norsworthy, Barber, Gbur and Kruger2018) when soybeans were treated at R1 with 2.2 and 8.8 g ae ha−1 dicamba. Injury observed in this trial was similar to injury observed by Soltani et al. (Reference Soltani, Nurse and Sikkema2016) when soybeans were treated at R1 with 30 g ae ha−1 dicamba.
Soybean injury at the onset of senescence had significant interactions for site-year by variety and timing by variety (Table 4). When pooled over application timings, soybean injury ranged from 26% to 42% across all site-years (Table 6). At MHK18 and OTT19, CR4748 had the least injury at senescence, whereas ST40B was among the varieties with the greatest injury at all locations. When pooled over site-years, soybean injury was similar across all varieties as a result of dicamba application at V3 and injury was 7% or less. However, dicamba application at R1 was associated with soybean injury ranging from 50% in CR4748 to 69% in ST40B (Table 6).
Table 6. Soybean injury at the onset of senescence of varying soybean varieties exposed to dicamba a at varying application times at Manhattan, KS in 2018 and 2019, and in Ottawa, KS in 2019. b,c
a 5.6 g ae ha−1 dicamba.
b Abbreviations: AG4135, Asgrow AG4135; CR3841, Credenz 3841LL; CR4748, Credenz 4748LL; MHK18, Manhattan, KS in 2018; MHK19, Manhattan, KS in 2019; OTT19, Ottawa, KS in 2019; ST40B, Stine 40BA02.
c Means followed by the same letter within a column are not statistically different according to Fisher’s protected LSD (α = 0.05).
d Means within site year pooled across application timings and means within application timing pooled across locations.
Limited injury at senescence following dicamba application at V3 is consistent with reports by Al-Khatib and Peterson (Reference Al-Khatib and Peterson1999), Osipitan et al. (Reference Osipitan, Scott and Knezevic2019), and Soltani et al. (Reference Soltani, Nurse and Sikkema2016) who observed end-of-the-season recovery in soybean treated at V3 with low rates of dicamba. The varieties in this study responded differently to dicamba at R1, with injury similar to that reported by France et al. (Reference France, Norsworthy, Ross, Castner and Barber2019) after soybeans were treated with similar dicamba rates at R1.
Height
There was a significant effect of dicamba application timing on soybean height across varieties (Table 7). Dicamba application at V3 resulted in 5% height reduction, whereas application at R1 resulted in 36% height reduction when pooled over site-years and varieties (data not shown). Height reductions were similar to those previously reported in the literature as a result of application at both V3 (Foster and Griffin Reference Foster and Griffin2018) and R1 (Kelley et al. Reference Kelley, Wax, Hager and Riechers2005).
Table 7. Analysis of significance of fixed effects and all interactions for soybean trait response to a reduced rate a of dicamba at multiple timings.
a 5.6 g ae ha−1 dicamba.
Yield Components and Yield
There were significant interactions between site-year and variety, site-year and application timing, and application timing and variety for relative main stem nodes per plant (Table 7). Main stem nodes per plant ranged from 59% to 98% of that of the nontreated checks when pooled across varieties (Table 8). Dicamba application at R1 resulted in a greater reduction of main stem nodes per plant than application at V3 at OTT19, but the main stem nodes per plant were similar for both application timings at MHK18 and MHK19. When pooled across site-years, main stem nodes per plant ranged from 57% to 98% of that of the nontreated checks. Dicamba applications at V3 had less impact on main stem nodes per plant compared to applications at R1 for all varieties except CR3841. Reductions in main stem nodes per plant were similar for all varieties when pooled across site-years (data not shown). These observations are in agreement with those of Robinson et al. (Reference Robinson, Simpson and Johnson2013) who observed 5% to 20% reduction of reproductive nodes per square meter when soybeans were treated at V3, V5, and R2 with dicamba rates ranging from 0.073 to 2.72 g ae ha−1. There were no significant differences observed among relative seed weight, pods per plant, and relative seeds per pod (data not shown).
Table 8. Soybean main stem nodes per plant relative to the plants in the nontreated control as a result of varying soybean varieties exposed to dicamba a at multiple application timings at Manhattan, KS in 2018 and 2019, and in Ottawa, KS in 2019. b,c
a 5.6 g ae ha−1 dicamba.
b Abbreviations: AG4135, Asgrow AG4135; CR3841, Credenz 3841LL; CR4748, Credenz 4748LL; MHK18, Manhattan, KS in 2018; MHK19, Manhattan, KS in 2019; OTT19, Ottawa, KS in 2019; ST40B, Stine 40BA02.
c Means separated within site-year and means followed by the same letter within a column are not statistically different according to Fisher’s protected LSD (α = 0.05).
There were significant interactions between site-year and variety and application timing and variety for relative yield (Table 7); however, means within site-years were similar for all varieties (data not shown). When pooled across site-years, dicamba application at V3 resulted in 95% or greater relative yield, regardless of the variety, whereas dicamba at R1 resulted 81% to 66% relative yield, with significant yield losses occurring in all varieties except for CR4748 (Figure 1). Yields in the nontreated checks for each variety ranged from 3,635 to 3,904 kg ha−1. Yield loss observed in this study is similar to yield loss reported by Foster and Griffin (Reference Foster and Griffin2018) and McCown et al. (Reference McCown, Barber and Norsworthy2018).
Figure 1. Soybean yield relative to nontreated plots for each variety following dicamba application at V3 and R1 at Manhattan, KS in 2018 and 2019, and at Ottawa, KS in 2019.Means followed by the same letter are not statistically different according to Fisher’s protected LSD (α = 0.05). AG4135, Asgrow AG4135, nontreated plot yield = 3,837 kg ha−1; CR3841, Credenz 3841LL, nontreated plot yield = 3,769 kg ha−1; CR4748, Credenz 4748LL, nontreated plot yield = 3,904 kg ha−1; ST40B, Stine 40BA02, nontreated plot yield = 3,635 kg ha−1.
Correlations with Yield
There was a strong, negative correlation between yield and soybean injury 4WAT and at senescence (Table 9). Robinson et al. (Reference Robinson, Simpson and Johnson2013) indicated that height reduction may be a quick way to estimate potential yield loss as a result of dicamba exposure. This was supported by these data, which show that in addition to a strong correlation, a linear relationship existed between soybean height and yield loss. As height increased by 31 cm, yield increased by 1 kg ha−1 (Figure 2).
Table 9. Pearson correlation coefficients and corresponding P-values for soybean trait response to a reduced rate a of dicamba at multiple timings at Manhattan, KS in 2018 and 2019, and in Ottawa, KS in 2019. b
a 5.6 g ae ha−1 dicamba.
b Abbreviation: WAT = weeks after treatment.
c Bolded text indicates strong correlations with levels of significance ≤0.05, Correlation coefficients were nonsignificant or significant at *P ≤ 0.05, **P ≤ 0.01 or ***P ≤ 0.001.
d Pearson coefficient: <0.3 = weak correlation, >0.3 but <0.5 = moderate correlation, >0.5 = strong correlation.
Figure 2. Linear regression of yield and height as a result of varying soybean varieties exposed to dicamba at varying application timings at Manhattan, KS in 2018 and 2019, and at Ottawa, KS in 2019. AG4135, Asgrow AG4135; CR3841, Credenz 3841LL; CR4748, Credenz 4748LL; ST40B, Stine 40BA02.
Height reduction was also correlated with main stem nodes per plant, pods per plant, and seed weight, similar to that reported by Robinson et al. (Reference Robinson, Simpson and Johnson2013). There were also strong correlations between yield components and relative yield. Relative pods per plant, seed weight, and seeds per pod were positively correlated to relative yield. Robinson et al. (Reference Robinson, Simpson and Johnson2013) noted that seeds per square meter, pods per square meter, and nodes per square meter need to be characterized in order to understand the total effect of dicamba exposures on non-DR soybean.
Effects on Offspring
There were no significant differences among site-years, variety, and timing of application for reduction in germination, reduction in offspring emergence, number of offspring injured, offspring soybean injury, and reduction in offspring height (data not shown). No injury was observed in offspring and there were no differences from the nontreated check. Previous studies have shown reduced germination, emergence, and vigor and increased injury to offspring of soybean treated with dicamba and that the response became more severe as dicamba rate increased (Auch and Arnold Reference Auch and Arnold1978; Jones et al. Reference Jones, Norsworthy, Barber, Gbur and Kruger2018; Thompson and Egli Reference Thompson and Egli1973; Wax et al. Reference Wax, Knuth and Slife1969). Differences in experimental parameters, specifically dicamba application rates and the environment in which offspring were tested, may explain the contrasting results obtained in this study.
In conclusion, yield loss due to dicamba was influenced by application timing, but not variety. Application at V3 resulted in minimal to no yield loss, but application at R1 resulted in 19% to 34% yield loss. Of the yield components impacted, the greatest effects were observed in main stem nodes per plant, where V3 applications resulted in less severe reduction than R1 applications. Pods per plant, seeds per pod, and seed weight did not result in significant reductions regardless of variety or timing of application. These data support additional label restrictions for use of dicamba in dicamba-resistant soybean to reduce the risk of injury during reproductive growth stages.
Acknowledgments
We thank the Kansas Soybean Commission for funding this project. We also thank Cathy Minihan for her assistance. No conflicts of interest have been declared. This publication is contribution no. 21-047-J from the Kansas Agricultural Experiment Station, Manhattan, KS.
Open access
