Content uploaded by William G Johnson
Author content
All content in this area was uploaded by William G Johnson on Sep 21, 2016
Content may be subject to copyright.
Response of Glyphosate-Tolerant Soybean Yield Components to
Dicamba Exposure
Andrew P. Robinson, David M. Simpson, and William G. Johnson*
Exposure of soybean to dicamba can result in leaf malformation and sometimes yield loss, but it is unclear how yield
components are affected by exposure to low quantities of this herbicide. The objectives were to characterize soybean injury
and quantify changes in seed yield and yield components of soybean plants exposed to dicamba, and determine if seed yield
loss can be estimated from visual injury ratings. Nine dicamba rates (0, 0.06, 0.23, 0.57, 1.1, 2.3, 4.5, 9.1, and
22.7 g ae ha
21
) were applied at three growth stages (V2 – two trifoliates, V5-five trifoliates, or R2-full flowering soybean)
to Beck’s brand ‘342NRR’ soybean planted near Lafayette, IN, in 2009 and 2010 and near Fowler, IN, in 2009. Visually
estimated soybean injury of 20% at the V2, V5, or R2 timing was 0.676 to 0.937 g ha
21
dicamba at 14 d after treatment
(DAT) and 0.359 to 1.37 g ha
21
dicamba at 28 DAT. Seed yield was reduced by 5% from 0.042 to 0.528 g ha
21
dicamba
and a 10% reduction was caused by 0.169 to 1.1 g ha
21
dicamba. The number of seeds m
22
, pods m
22
, reproductive
nodes m
22
, and nodes m
22
were the most sensitive yield components. Path analysis indicated that dicamba reduced seeds
m
22
, pods m
22
, reproductive nodes m
22
, and nodes m
22
which were the main causes of seed yield loss from dicamba
exposure. The correlation of seed yield loss and visual soybean injury was significant (P ,0.0001) for both the V2
treatment timing (R
2
50.92) and the V5 and R2 treatment timings (R
2
50.91). Early-season injury rating of 8% at the
V2 treatment and 2% at the V5 or R2 treatments caused 10% or more yield loss.
Nomenclature: Dicamba; soybean, Glycine max (L.) Merr.
Key words: 3,6-dichloro-o-anisic acid, crop injury, drift, dose response, reproduction.
Soybean plants exposed to dicamba exhibit malformed
leaves, petioles, and reproductive structures. An increase in
occurrences of off-site movement of dicamba onto neighbor-
ing dicamba-sensitive soybean fields is likely to occur with the
commercialization of dicamba-tolerant soybean (Behrens et al.
2007). In addition, sensitive soybean plants can be exposed
inadvertently to dicamba from applications on nearby fields
of maize (Zea mays L.), or sorghum [Sorghum bicolor (L.)
Moench], pastures, or rangeland. Soybean plants are reported
to be injured by dicamba rates greater than or equal to
1.0 g ha
21
(Auch and Arnold 1978) and yield reduced at rates
as low as 0.04 g ha
21
(Weidenhamer et al. 1989). Reduction
in yield has been correlated with seed number and plant
height (Weidenhamer et al. 1989); however, a complete
review of the effect of dicamba on yield components has not
been reported.
Particle drift, volatilization, and contamination of spraying
equipment are the principal ways in which dicamba comes in
contact with sensitive-soybean plants. Up to 16% of the spray
solution was reported to move off-site when wind speeds were
30 km h
21
(Maybank et al. 1978). Drift can be reduced by
nozzle selection, pressure, boom height, travel speed, and not
spraying when wind speeds are high (Maybank et al. 1978;
Wolf et al. 1993). Maybank et al. (1978) reported that
doubling nozzle size to 7.6 L min
21
and spray volume to
100 L ha
21
reduced drift from 16 to 4%. In addition to drift,
dicamba can move off-site through volatilization. Dicamba
can volatilize during a prolonged period of time (2 to 12 h)
when temperatures are high (20 to 30 C) and air movement is
slow (Behrens and Lueschen 1979). The third way in which
sensitive vegetation is injured is by contact with contaminated
sprayer solutions. Spray equipment cleaned with an ammo-
nium–water solution after dicamba application had up to
0.63% dicamba exiting the sprayer on the next application
(Boerboom 2004). In addition to the previously mention-
ed scenarios, misapplication of dicamba may increase in
frequency if some fields are planted with dicamba-tolerant
soybean while nearby fields have soybean that are not tolerant
to dicamba.
As a result of exposure to dicamba, an auxin mimic
herbicide, soybean plants often become malformed and have
altered morphology and reduced growth. Applications during
vegetative growth affected the new leaf development but did
not affect pod and seed production because dicamba was
likely detoxified before reproduction began (Auch and Arnold
1978). Severity of leaf injury was influenced by application
rate, not by growth stage (Auch and Arnold 1978). Early-
season exposure stimulated lateral growth and increased
branching, especially when the apical meristem died (Ander-
sen et al. 2004; Wax et al. 1969). Dicamba applied during
flowering or early-pod production caused abnormal pod
formation (Auch and Arnold 1978; Kelley et al. 2005). These
injuries reduced yield by restricting the production of yield
components including pods, seed mass, and seed number
(Kelley et al. 2005; Wax et al. 1969).
Yield loss from dicamba drift can be hard to describe with
mathematical models. Much of the unpredictable yield loss
has been attributed to application timing and environment
(Andersen et al. 2004; Auch and Arnold, 1978; Weidenhamer
et al. 1989), but the plasticity of soybean make the study of
stresses difficult to explain. In one study, an application
of 56 g ha
21
on V1 to V2 soybean resulted in 0.5% yield
reduction (Auch and Arnold 1978), whereas another study
reported that the same rate caused an 83% yield reduction
when applied to V3 soybean (Andersen et al. 2004). Dicamba
applied at 11 g ha
21
increased yield 21% when applied at late
pod, but reduced yield 42% at early bloom (Auch and Arnold
1978). In a drought-stressed year, 0.4 g ae ha
21
caused 10%
yield loss, whereas in a year with adequate rainfall 15 g ha
21
was required for a 10% yield reduction (Weidenhamer et al.
DOI: 10.1614/WS-D-12-00203.1
* First and third authors: Graduate Research Assistant and Professor,
Department of Botany and Plant Pathology, 915 W. State Street, Purdue
University, West Lafayette, IN 47907; second author: Field Scientist, Dow
AgroSciences, 9330 Zionsville Road, Indianapolis, IN 46268. Corresponding
author’s E-mail: wgj@purdue.edu
Weed Science 2013 61:526–536
526 NWeed Science 61, October–December 2013
1989). To better understand changes in soybean yield caused
by dicamba, an analysis of yield components can elucidate
alterations in yield production.
Yield components can be distinguished by primary traits
affecting yield (seed mass and seed number m
22
), secondary
traits (seeds per pod and pod number m
22
), tertiary traits
(pods per reproductive node and reproductive node m
22
) and
quaternary traits (node number m
22
and percent reproductive
nodes) (Board and Modali 2005). Researchers have analyzed
yield components to measure the effect of planting date
(Robinson et al. 2009), to compare old vs. new soybean
cultivars (Kahlon et al. 2011), and to measure the effect of
2,4-D on soybean (Robinson et al. 2013). In these studies it
was determined that reproductive node number, pod number,
and seed number caused higher yield. Evaluating yield
components may enhance the understanding of how dicamba
exposure affects yield when applied at different growth stages
or in varying environments. The objectives were to charac-
terize the response of glyphosate-tolerant soybean growth
stages and dicamba dose on soybean injury, yield components,
and total seed yield, and to determine if early-season soybean
injury ratings can be used to determine seed yield loss.
Materials and Methods
Field experiments were planted at the Throckmorton
Purdue Agriculture Center (TPAC) located near Lafayette,
IN (40.3001, 286.9056), on May 18, 2009, and on June 8,
2010 and at the Dow AgroSciences Midwest Research Center
(MRC) located near Fowler, IN (40.6336, 286.1101), on
June 6, 2009. The 2010 field experiment at MRC was flooded
and data were not collected. Soil type at TPAC in 2009
was an Octagon silt loam (fine-loamy, mixed, active, mesic
Mollic Oxyaquic Hapludalfs); at TPAC in 2010 soil was a
Throckmorton silt loam (fine-silty, mixed, superactive, mesic
Mollic Oxyaquic Hapludalfs), and soil type at MRC in 2009
was a Darroch silt loam (fine-loamy, mixed, superactive,
mesic Aquic Argiudolls). Beck’s brand ‘342NRR’ soybean was
seeded in 38-cm rows at 430,000 seeds ha
21
. All fertility
practices were conducted according to Purdue University
recommendations (Gerber et al. 2012). Conventional tillage
was utilized at all field sites, except at TPAC in 2009 where no
tillage was used.
Plots were kept weed-free by applying PRE and POST
herbicides and by subsequent hand-weeding as required.
Weed control at TPAC during both years consisted of
metribuzin (158 g ai ha
21
) plus chlorimuron (26 g ai ha
21
)
applied just prior to soybean emergence. POST weed control
at TPAC in 2009 consisted of glyphosate (1,060 g ae ha
21
)
plus ammonium sulfate (2,037 g 100 L
21
), applied on June
19; at TPAC in 2010, glyphosate (1,120 g ha
21
) plus
ammonium sulfate (2,037 g 100 L
21
) plus clethodim
(102 g ae ha
21
) were applied on July 11. Clethodim was
used to control glyphosate-tolerant volunteer maize. At the
MRC location in 2009, trifluralin (1,400 g ae ha
21
) plus
imazethapyr (70 g ae ha
21
) were incorporated twice prior to
planting. Detailed information of herbicides utilized can be
found in Table 1.
The experimental design was a randomized complete block
with a factorial arrangement of treatments. Treatments were
application timing at V2, V5, or R2 stages of soybean and
dicamba (diglycolamine salt) rates of 0, 0.06, 0.2, 0.6, 1.1,
2.3, 4.5, 9.1, and 22.7 g ha
21
. Soybean plants within a plot
were considered to have reached a certain stage of
development when at least half of the plants reached that
stage (Fehr and Caviness 1977). Plot size was 3.1 m wide and
9.1 m long and consisted of a 3.1-m-long and 1.5-m-wide
buffer to reduce the possibility of off-target movement into
adjacent plots.
All dicamba treatments were applied in 140 L ha
21
carrier
volume using a CO
2
-pressurized backpack sprayer with a 3.1-
m-wide boom and XR11002 flat fan nozzles (TeeJet Spraying
Systems Company, Wheaton, IL 60189) at 138 kPa. Wind
speeds were less than 5 km h
21
when treatments were applied.
Visual estimates of percentage of soybean injury were
recorded 14 and 28 DAT using a scale of 0 to 100%, where 0
5no crop injury and 100 5complete plant death (Table 2;
Figure 1). Plant height was recorded from three arbitrarily
selected plants at the R8 growth stage. At maturity, 10 plants
from the middle two rows of each treatment were arbitrarily
selected to determine the following yield components as
outlined by Board and Modali (2005): seed mass (g 100
seeds
21
), seeds m
22
, seeds pod
21
, pods m
22
, main-stem
reproductive nodes m
22
, pods reproductive node
21
, main-
stem nodes m
22
, and percentage of reproductive nodes. Plots
were harvested with a plot combine and seed yield was
adjusted to 130 g kg
21
moisture. Oil and protein
concentrations were determined from machine-harvested seed
using near-infrared reflectance spectroscopy at the Purdue
University Grain Quality Laboratory.
Statistical Analysis. All data were normalized to a percentage
of the untreated check, except soybean injury. Thereafter, data
were subjected to ANOVA using the SAS PROC MIXED
procedure (SAS software for Windows, version 9.2. SAS
Institute Inc., Cary, NC 27513) to test for significant effects
of experiment, dicamba rate, and treatment timing, and the
interaction of dicamba rate by treatment timing. When the
ANOVA did not reveal any differences between experiment,
treatment timing, or dicamba rate by treatment timing, the
data were pooled for further analysis. However, if a difference
was observed, experiment or treatment timings were separated
Table 1. Sources of material for herbicides.
Common name Trade name Rate in active ingredient Manufacturer
Metribuzin plus chlorimuron Canopy DF 0.643 g g
21
DuPont Crop Protection, Wilmington, DE; www.dupont.com
0.107 g g
21
Glyphosate Roundup PowerMax 660 g L
21
Monsanto Company, St. Louis, MO; www.monsanto.com
Clethodim Tapout N116 g L
21
Helena Chemical Company, Collierville, TN;, www.helenachemical.com
Trifluralin Treflan HFP N480 g L
21
Dow AgroSciences, Indianapolis, IN; www.dowagro.com
Imazethapyr Pursuit N240 g L
21
BASF Corporation Agricultural Products, Research Triangle Park, NC; www.agro.
basf.com
Dicamba Clarity N480 g L
21
BASF Corporation Agricultural Products, Research Triangle Park, NC; www.agro.
basf.com
Robinson et al.: Soybean response to dicamba N527
according to the analysis. Further analysis was performed using
nonlinear regression (dose-response model) to predict seed
yield, plant height, and yield component loss due to dicamba
injury using the drc package in R (Knezevic et al. 2007) (R,
Version 2.10.1, www.r-project.org). A lack-of-fit test (nonsig-
nificant P value when P $0.05) was used to determine if the
data were well described by the model, because R
2
values do not
adequately measure the fit of nonlinear models (Spiess and
Neumeyer 2010). The three-parameter Weibull model was
used in two different parameterizations (Equations 1 and 2) to
describe the response of dicamba (Ritz and Strebig 2011; Seber
and Wild 1989).
Y~d1{exp {exp blog x{log eðÞ
fg
½ðÞ½1
Y~dexp {exp blog x{eðÞ
fg
½ðÞ½2
Yrepresents the response variable, bthe relative slope of the
curve, dthe upper limit, ethe inflection point, and xthe rate of
dicamba. Effective dose (ED) causing 10, 20, and 50% loss
(ED
10
,ED
20
, and ED
50
) was calculated from the nonlinear
regression for visual soybean injury values associated with
soybean injury from herbicides (Kniss and Lyon 2011). The
ED
5
,ED
10
,andED
20
were calculated for seed yield and yield
components. These ED values were chosen because they have
previously been used to estimate herbicide doses causing
unacceptable seed yield loss (Knezevic et al. 1998; Kniss and
Lyon 2011).
The SAS GLM procedure was used to correlate soybean
injury and seed yield. Quadratic relationships (Equation 3)
were developed for visual rating data collected 14 and 28
DAT and were subject to ttests to determine if similarities
existed between DAT and treatment timings. Yrepresents
seed yield, b
0
the intercept, b
1
the linear coefficient, b
2
the
quadratic coefficient, and xthe value of soybean injury. The
correlation coefficient (R
2
) values were used to describe the fit
of the model because this model was derived as a generalized
linear model where each bis linear.
^
YY ~^
bb0z^
bb1xz^
bb2x2½3
Path analysis and correlation of yield components were
applied to soybean yield as described by Kahlon et al. (2011).
Path analysis is commonly used by plant breeders to
distinguish relative importance of yield components and has
been used to determine the effect of weeds on yield
components of crops (Donald and Khan 1996). Analyses
were performed with the primary (seeds m
22
and seed mass),
secondary (pods m
22
and seeds pod
21
), tertiary (reproductive
nodes m
22
and pods reproductive node
21
), and quaternary
(nodes m
22
and percentage of reproductive nodes) traits.
Experiment and treatment timing were pooled. The SAS IML
procedure was utilized to determine the direct path
coefficients (Kang 1994). Indirect path coefficients were a
result of multiplying the path coefficient with the correlation
coefficient. The correlation coefficient and the residual
error were also calculated (Kang 1994). A bidirectional path
analysis was used because of the simultaneous development of
yield components and the ability they have to interact and
compensate with one another. The criteria used for
identifying an important trait affecting its response variable
were when that trait had a large positive correlation with the
response variable, a large positive direct effect, and a small
indirect effect (Kahlon et al. 2011).
Results and Discussion
Environment. Environmental conditions varied by location
and year (Table 3). A higher average temperature occurred in
2010 than in 2009. Precipitation was lower at TPAC in 2009
during reproductive stages (July through September) when
compared to TPAC in 2010 and MRC in 2009. Because of
this, seed yield at TPAC in 2009 was lower than in the other
experiments. Average soybean yield in the untreated checks at
TPAC in 2009 was 2.2 Mg ha
21
as compared to 4.4 Mg ha
21
at TPAC in 2010 and MRC in 2009.
Visual Symptoms. Nonlinear regressions were fit to describe
soybean injury according to the ANOVA (Table 4, Figure 2).
At 14 DAT the ED
20
values ranged from 0.676 to
0.937 g ha
21
dicamba (Table 5). At 28 DAT the ED
20
values ranged from 0.359 to 1.37 g ha
21
dicamba. It was
observed that rates greater than or equal to 2.3 g ha
21
caused
apical meristem death. Regrowth typically occurred at axillary
buds either at the cotyledonary node or the unifoliate node,
causing the growth of two main branches, with one branch
generally becoming dominant. This slow regrowth could help
explain why soybean plants treated with dicamba can have a
delayed progression through growth stages compared to
untreated plants.
Table 2. Rating scale for visual estimate of soybean injury affected by synthetic auxin herbicides.
Rating Description
0 No injury, plant growth is normal.
10 Slight reduction in height or canopy volume, cupped or bubbled leaves on less than or equal to the upper 10% of the plant, bent petioles, and chlorosis or necrosis.
20 Moderately crinkled leaflets (extended across less than or equal to the upper 20% of the plant), curled petioles, reduced height and canopy volume, cupped
terminal leaflets.
30 Moderate to high reduction of height and canopy; compacted internodes and plants begin to have an abnormal appearance; malformation with drawstring,
fiddleneck, or cupped effects on less than or equal to the upper 30% of the plant; many petioles curled and main stems may be bent.
40 Highly stunted plants (less than or equal to 40% of the plant), petioles curled and main stems bent or starting to curl, upper leaves exhibit severe malformation
and expansion of new leaves suppressed, plant may have patches of necrotic tissue.
50 Very high reduction of plant height (less than or equal to 50% of the plant) with little likelihood of recovery from the apical meristem, new growth suppressed,
formation of pods reduced or malformed, some leaf and stem tissue becomes necrotic, petioles and stems show severe twisting.
60 Severe height and canopy reduction, including any new growth from axillary buds; leaves severely cupped or fiddlenecked on less than or equal to 60% of the
plant; petioles and stems twisted, swollen, and splitting; more extensive die-back of tissue.
70 Severe to very severe reduction of plants, new growth callused and inhibited, most leaves severely deformed and mostly necrotic, extensive petiole bending.
80 Very severe soybean injury, less than or equal to 80% of the plants mainly prostrate, petioles twisted with leaves drooping, leaves chlorotic or necrotic, stems
severely twisted, swollen, and split.
90 Plant dying, less than or equal to 90% of the plants mainly prostrate, leaves and stems mostly chlorotic or necrotic, all petioles severely twisted, swollen, or split.
100 All plants dead.
528 NWeed Science 61, October–December 2013
These models describing visual soybean injury from
dicamba exposure predicted similar injury ratings to that
reported by other researchers (Al-Khatib and Peterson 1999;
Andersen et al. 2004; Kelley et al. 2005). However, the model
was different from the report of Sciumbato et al. (2004), who
used a more conservative scale for visual soybean injury rating.
Seed Yield and Plant Height Loss. Nonlinear regressions
were fit to describe soybean seed yield and plant height loss
according to the ANOVA (Table 4). The ED
10
for seed yield
at TPAC in 2009 was 0.169 g ha
21
dicamba (Table 6). The
ED
10
for seed yield at MRC in 2009 and TPAC in 2010
ranged from 0.529 to 1.1 g ha
21
dicamba. The ED
10
for
plant height ranged from 0.577 to 2.21g ha
21
dicamba. The
model prediction of plant height reduction was similar to the
amount of reduction reported by Kelley et al. (2005) at
5.6 g ha
21
and Wax et al. (1969) at 17.5 g ha
21
dicamba.
Similar to the results at MRC in 2009 and TPAC in 2010,
other researchers reported that dicamba treatments of
5.6 g ha
21
at V1 to V3 soybean stages caused less seed yield
loss when compared to dicamba applications at V7 to early
bloom (Auch and Arnold 1978; Kelley et al. 2005). Retention
of yield when dicamba was applied to V2 soybean at MRC
in 2009 and TPAC in 2010 could be attributed to soybean
having more time to overcome dicamba injuries before
reproduction began. In a study where main-stem nodes were
removed at V2, V6, and R3 soybean stages, the V2 timing had
the least yield loss (Conley et al. 2009), which demonstrated
Figure 1. Visual injury of soybean at 28 d after 0.23, 0.56, and 1.1 g ha
21
dicamba treatments at Lafayette, IN, in 2009 (V2 and R2 growth stages pictured) and
Fowler, IN, in 2009 (V5 growth stage). At 0.23 g ha
21
dicamba, soybean injury in each experiment averaged (a) 20% at the V2 growth stage, (b) 15% at the V5 growth
stage, and (c) 17% at the R2 growth stage; at 0.56 g ha
21
dicamba, soybean injury averaged (d) 25% at the V2 growth stage, (e) 35% at the V5 growth stage, and (f) 28%
at the R2 growth stage; whereas 1.1 g ha
21
dicamba, caused an average of injury of (g) 30% at the V2 growth stage, (h) 53% at the V5 growth stage, and (i) 43% at the
R2 growth stage).
Robinson et al.: Soybean response to dicamba N529
the ability of soybean to compensate from injury occurring
during early vegetative growth.
MRC in 2009 and TPAC in 2010 had similar yield loss to
previous research; however, at most rates there was much
higher yield loss at TPAC in 2009 than previous research has
reported from dicamba exposure (Auch and Arnold 1978;
Kelley et al. 2005; Wax et al. 1969; Weidenhamer et al.
1989). The greater yield reduction at TPAC in 2009 is likely
the result of less precipitation, which minimized the ability of
the soybean plant to compensate for the early season injury.
Similarly, Weidenhamer et al. (1989) reported yield loss was
11.5 times greater for equivalent rates in a drought-stressed
year compared to a year with typical rainfall. Drought stress in
soybean has been reported to decrease photosynthesis (Liu et
al. 2004), total dry matter (Board and Modali 2005), flower
production, pod number (Kokubun and Takahashi 2001),
and seed number (Sionit and Paul 1977). Drought stress to
soybean plants could have inhibited the ability of soybean to
detoxify or sequester dicamba, allowing it to stay active in the
plant for a longer period of time, while the concomitant stress
of drought could have increased abortion of flowers and pods.
Many plants exude auxin herbicides through their roots
(Dexter et al. 1971; Lingle and Suttle 1985; Robocker 1976);
however, water stress can lead to reduced translocation, which
may allow auxin herbicides to remain active longer in the
plant. The severity of drought stress varies by the time and
duration during soybean reproductive development (Kokubun
and Takahashi 2001; Sionit and Paul 1977). The interaction
of drought stress by dicamba rates on soybean seed yield and
yield components is not known and warrants further study to
improve understanding and more accurately predict seed yield
loss in such situations.
Characterization of Yield Components. Nonlinear regres-
sions were fit to describe each yield component according to
the analysis of variance (Table 4). Treatment timing and rate
affected the response of most yield components. The number
of seeds m
22
, pods m
22
, seeds pod
21
, reproductive nodes
m
22
, and nodes m
22
were reduced as dicamba rates increased
(Figure 3). The response of seed mass, pods reproductive
node
21
, and percentage of reproductive nodes increased or
decreased with increasing dicamba rates and treatment
timings, but generally had less change than the previously
mentioned yield components. The ED
10
of seeds m
22
ranged
from 0.061 to 11.9 g ha
21
dicamba (Table 6). Pods m
22
were reduced 10% with 2.44 to 3.65 g ha
21
dicamba when
applied at the V5 or R2 stage; however, there was less than
10% reduction when dicamba was applied at the V2 stage
with the dicamba rates evaluated in this experiment.
Reproductive nodes m
22
had an ED
10
range of 0.247 to
1.44 g ha
21
and the ED
10
of nodes m
22
was between 0.162
to 2.04 g ha
21
dicamba. Seed mass changed little, except a 5%
reduction was seen when dicamba treatments occurred at the
V5 soybean stage. Seeds pod
21
were reduced 10% with 7.35
to 14.2 g ha
21
dicamba when applied at the V5 or R2 stage;
however, there was less than 10% reduction in seeds pod
21
Table 3. Total monthly precipitation and mean monthly air temperature (May through September) at the Midwest Research Center (MRC in 2009) located in Fowler,
IN, during 2009 and at the Throckmorton Purdue Agriculture Center (TPAC) located in Lafayette, IN, during 2009 and 2010.
Precipitation Temperature
MRC in 2009 TPAC in 2009 TPAC in 2010 MRC in 2009 TPAC in 2009 TPAC in 2010
---------------------------------------------------------------- m m --------------------------------------------------------------- ------------------------------------------------------------------- C -----------------------------------------------------------------
May 123 61 62 17.0 17.3 18.1
June 69 93 106 22.1 22.4 23.5
July 67 37 65 20.3 20.7 24.4
August 57 30 44 20.5 21.4 24.1
September 11 3 24 18.3 18.6 19.7
Table 4. Analysis of variance of soybean injury, seed yield and plant height, yield components, and seed composition treated with different rates of dicamba ranging
from 0 to 22.7 g ha
21
at the V2, V5, and R2 soybean growth stages at the Throckmorton Purdue Agricultural Center, Lafayette, IN in 2009 and 2010 and at the
Midwest Research Center, near Fowler, IN in 2009.
Source Experiment Dicamba rate Treatment timing Dicamba rate 3treatment timing
-------------------------------------------------------------------------------------------------------------------- P value -------------------------------------------------------------------------------------------------------------------
Soybean injury
14 DAT
a
0.5780 ,0.0001 ,0.0001 0.2316
28 DAT 0.0017 ,0.0001 ,0.0001 ,0.0001
Seed yield and plant height
Seed yield 0.0219 ,0.0001 0.0389 0.0009
Plant height 0.2429 ,0.0001 ,0.0001 0.0008
Yield components
Seeds m
22
0.0069 ,0.0001 0.0393 0.0002
Seed mass 0.6660 0.0160 ,0.0001 ,0.0001
Pods m
22
0.0117 ,0.0001 0.3264 0.0024
Seeds pod
21
0.6760 ,0.0001 0.0004 0.0001
Main stem reproductive nodes m
22
0.5177 ,0.0001 0.0140 ,0.0001
Pods reproductive node
21
0.0517 ,0.0001 ,0.0001 ,0.0001
% Reproductive nodes 0.0860 0.0006 0.0066 0.0041
Main stem nodes m
22
0.7036 ,0.0001 0.0040 ,0.0001
Seed composition
Oil concentration 0.1274 ,0.0001 ,0.0001 ,0.0001
Protein concentration 0.1738 ,0.0001 ,0.0001 ,0.0001
a
Abbreviation: DAT, days after treatment.
530 NWeed Science 61, October–December 2013
when dicamba was applied at the V2 stage with the dicamba
rates evaluated in this experiment. The number of pods
reproductive node
21
did not have greater than 5% change at
any treatment timing. Percentage of reproductive nodes was
only reduced at the R2 treatment timing, with an ED
10
of
16.1 g ha
21
dicamba.
The results from this study found less change in seed mass
due to dicamba injury at both the vegetative and reproductive
stages than was reported by Kelley et al. (2005) and Wax et al.
(1969). Pod loss was greater at TPAC in 2009 than was
reported by Kelley et al. (2005), which might be accredited to
drought stress at this experimental site. In contrast to these
findings, Wax et al. (1969) found no reduction in the number
of pods, but this was likely a result of a bushy soybean cultivar
planted at 76- or 102-cm row spacing, which enabled the
plant to branch out and produce more pods. Seeds pod
21
responded to dicamba in a similar manner as reported by
Kelley et al. (2005), but differently from Wax et al. (1969)
who found no reduction in seeds pod
21
from dicamba
applications. The cultural practices or genetics of the soybean
may be the reason why no reduction in seeds pod
21
was
observed.
The reduction of nodes and, subsequently, reproductive
nodes can limit the production of other yield components and
yield. The rapid loss of reproductive nodes was likely the
result of apical meristem death and the delay of axillary bud
growth. A reduction in nodes and reproductive nodes have
been found to be yield limiting factors in comparing old with
new cultivars (Kahlon et al. 2011).
Seed Composition. Dicamba rate, treatment timing, and
dicamba rate by treatment timing had significant effects on
harvested seed oil and protein concentration (Table 4).
Nonlinear regression was fit to the V2, V5, and R2 treatment
timings for both oil and protein concentration. At 22.7 g ha
21
dicamba, oil concentration was reduced by 2.1 and 2.6% at
the V2 and V5 treatment timings, respectively (data not
shown). The R2 treatment timing had a maximum reduction
in oil concentration of 15% at 22.7 g ha
21
dicamba (data not
shown). Seed protein concentration at the V2 and V5
treatment timings was reduced by 0.4% and increased by
0.2%, respectively, at 22.7 g ha
21
(data not shown). Dicamba
treatments at the R2 timing increased protein concentration
by up to 4.7% at 22.7 g ha
21
dicamba (data not shown).
These results are consistent with those of Wax et al. (1969)
who reported that dicamba injury during the vegetative
growth had little effect on soybean seed oil and protein
concentration whereas dicamba injury during soybean
reproduction resulted in changes in seed oil and protein
concentration.
Effect of Yield Components on Total Seed Yield. The direct
path effect of seeds m
22
on seed yield was 5.9 times greater
than seed mass on seed yield (Figure 4). The indirect path
effect of seed mass on seed yield through seeds m
22
was
strongly negative (20.18) and eliminated any benefit of the
direct path effect of seed mass on seed yield, whereas the
indirect path effect of seeds m
22
was 20.03, indicating that
little yield compensation occurred. As a result the direct path
effect of seeds m
22
on seed yield was similar to the correlation
of seeds m
22
and seed yield. The direct effect and the
correlation of seeds m
22
on seed yield had large positive
Figure 2. Nonlinear regressions of soybean injury (A) 14 and 28 d after
treatment at the Midwest Research Center (MRC) near Fowler, IN; (B) 14 and
28 d after treatment at the Throckmorton Purdue Agriculture Center (TPAC)
near Lafayette, IN in 2009; and (C) 28 d after treatment the TPAC in 2010 as
affected by dicamba treatments applied to soybean at the V2, V5, or R2 growth
stages. Regression parameters and effective dose values can be found in Table 6.
Robinson et al.: Soybean response to dicamba N531
values, whereas the indirect effect was small. This indicated
that seeds m
22
had the greatest effect on seed yield among the
primary traits.
The direct path effects of the secondary traits (seeds pod
21
and pods m
22
) on seeds m
22
was 3.7 times greater for pods
m
22
compared to seeds pod
21
(Figure 5). The indirect effect
of seeds pod
21
on seeds m
22
through pods m
22
was 0.30. The
indirect effect increased the compensation and benefit of the
direct path effect of seeds pod
21
on seeds m
22
. The indirect
effect of pods m
22
was small (0.08), and as a result the
correlation of pods m
22
on seeds m
22
was similar to the direct
path effect. Because the path coefficient and correlations had
high values and the indirect effect was low, pods m
22
was the
most influential secondary trait on seeds m
22
.
The tertiary traits, pods reproductive node
21
and repro-
ductive nodes m
22
on pods m
22
, had a direct effect that was
2.0 times greater for the reproductive nodes m
22
compared to
the pods reproductive node
21
when dicamba was applied
(Figure 5). The indirect effect of pods reproductive node
21
on
pods m
22
through reproductive nodes m
22
was strongly
negative (20.88). The compensation of yield components
eliminated the beneficial direct effect of pods reproductive
node
21
on pods m
22
as was seen in the correlation. In contrast,
the indirect effect of reproductive nodes m
22
on pods m
22
through pods reproductive nodes
21
was 20.43. The small
indirect effect and the large positive direct effect and correlation
values indicated that the influence of reproductive nodes m
22
on pods m
22
was the strongest of the tertiary traits.
The direct effect of nodes m
22
compared to the percentage
of reproductive nodes (quaternary traits) on reproductive
nodes m
22
was 4.9 times greater (Figure 5). There was no
indirect path effect of nodes m
22
on reproductive nodes m
22
through percentage of reproductive nodes, thus no seed yield
compensation occurred and the direct path effects and
correlations of nodes m
22
to reproductive nodes m
22
were
the same. The indirect path effect of percentage of
reproductive nodes on reproductive nodes m
22
was 0.01.
Little compensation of quaternary traits occurred from
dicamba treatments; however, because nodes m
22
had a large
positive direct effect and correlation on reproductive nodes
m
22
, nodes m
22
were the most important quaternary trait.
Changes in seed yield from dicamba treatments were caused
by the number of main-stem nodes m
22
, main-stem
reproductive nodes m
22
, pods m
22
, and seeds m
22
.In
contrast to the finding of Kahlon et al. (2011) there was
strong yield compensation at the tertiary traits, pods
reproductive node
21
and reproductive nodes m
22
. The
difference might be because of the effects of dicamba on
soybean reproductive structures. As plants became stressed
from exposure to dicamba, the main contributors of seed yield
were seeds m
22
, pods m
22
, reproductive nodes m
22
, and
nodes m
22
. This was likely the result of the other yield
components compensating to allow the main contributing
yield components to maximize their growth to increase seed
yield.
Implications of Visual Injury Estimates on Seed Yield Loss.
Regression analyses of estimated visual soybean injury were
correlated to seed yield loss. Visual estimates of soybean injury
at 14 and 28 DAT were pooled because they were not
different from each other (P 50.1739) (data not shown). The
V2 treatment timing was different from the V5 treatment
timing (P 50.0015) and the R2 treatment timing (P 5
0.0014), whereas the V5 treatment timing and the R2
treatment timing were not different (P 50.5527) (data not
shown). Predicting yield loss from visual soybean injury
ratings was significant at the V2 treatment timing
(P ,0.0001) and the V5 and R2 treatment timings
(P ,0.0001) (Figure 5). The V2 treatment timings had an
R
2
50.92 and the V5 and R2 treatment timings had an R
2
5
0.91. Soybean injury at 14 and 28 d after dicamba exposure
may be used to estimate the amount of seed yield loss. Visual
soybean injury ratings are easy, quick, and a cost-effective
means to measure potential seed yield loss from dicamba
exposure. Nevertheless, human variability in rating and the
response of different soybean cultivars to dicamba under
various environments will reduce the consistency of estimating
seed yield loss.
Visual injury to soybean plants from dicamba can be easy to
detect when dicamba has drifted or when a tank contamina-
tion occurs on a dicamba-sensitive soybean, because of the
sensitivity of soybean to dicamba. When dicamba drifts or is
Table 5. Nonlinear curve regression parameters of visual estimates of soybean injury at 14 and 28 d after treatment used in dicamba treatments applied to soybean at
the V2, V5, or R2 growth stages at the Throckmorton Purdue Agricultural Center, Lafayette, IN, in 2009 and 2010 and at the Midwest Research Center, near Fowler,
IN, in 2009.
DAT
a
Growth stage
b
Regression parameters
c
ED
10
6SE
d
ED
20
6SE ED
50
6SEbde
------------------------------------------------------ g a e ha
21
----------------------------------------------------
14 V2, V5 20.465 85.2 1.05 0.203 60.02 0.676 60.09 4.05 61.2
14 R2 20.520 72.9 1.07 0.285 60.03 0.937 60.2 6.97 63.6
MRC in 2009 and TPAC 2009
28 V2 20.344 125 3.50 0.238 60.04 0.884 60.3 4.55 63.0
28 V5 20.654 84.7 0.540 0.169 60.02 0.399 60.04 1.44 60.3
28 R2 20.609 75.7 0.424 0.134 60.02 0.359 60.04 1.80 60.5
TPAC in 2010
28 V2 20.822 55.8 1.05 0.544 60.05 1.37 60.2 15.6 67.8
28 V5 20.472 80.2 0.904 0.192 60.02 0.653 60.1 4.41 62.0
28 R2 20.584 76.9 1.25 0.368 60.03 1.02 60.2 5.30 61.8
a
Abbreviations: DAT, days after treatment; ED, effective dose, MRC, Midwest Research Center; TPAC, Throckmorton Purdue Agricultural Center.
b
Treatments of dicamba were applied at the MRC in 2009 and at the TPAC in 2009 and 2010.
c
Dose-response parameters where brepresents the relative slope of the curve, dthe upper limit, and ethe inflection point ofthe equation Y5d(exp[2exp{b(log x2e)}]).
d
ED
5
,ED
10
ED
20
, effective dose causing 10, 20, and 50% visual soybean injury.
532 NWeed Science 61, October–December 2013
volatilized across a field the plants with the greatest amount of
injury cause excessive concern, but symptoms will gradually
lessen the farther away from the source of dicamba. The
difference in rate needed to achieve 20% soybean injury was
two to three times greater than what was needed for 10%
soybean injury (Table 5). Although foliar symptoms may
appear to be severe, the yield loss may be minimal when only
leaves and petioles are malformed. Greater yield loss would
likely be associated with a reduction in the number of nodes,
which limits other reproductive structures from forming. This
would occur when dicamba causes an inhibition of apical
meristem growth or death of the apical meristem.
Plant height may be a good way to quickly estimate the
potential seed yield loss that is caused by dicamba exposure
(Weidenhamer et al. 1989). However, the challenge to this
approach is determining when the appropriate time is to
measure height and associate that with yield loss. Visual
soybean injury was greatest at 28 DAT, but Weidenhamer
et al. (1989) found a strong relationship of seed yield loss by
height loss at 60 DAT. A reduction in plant height would
likely be associated with reduced leaf area, fewer main-stem
nodes, and fewer pods or less seed mass because of reduced
photosynthesis. Seed number is often highly correlated with
soybean yield (De Bruin and Pedersen 2008). Thus, a
reduction in plant height can reduce many yield components
causing seed yield loss.
Soybean yield cannot be described by a single yield
component, rather all yield components influence yield to
varying degrees when affected by nontarget herbicides. The
amount of herbicide and the development stage of the plant at
Table 6. Nonlinear regression parameters of seed yield, plant height, and yield components as affected by dicamba treatments applied to soybean at the V2, V5, or R2
growth stages at the Throckmorton Purdue Agricultural Center (TPAC), Lafayette, IN, in 2009 and 2010 and at the Midwest Research Center (MRC), near Fowler, IN,
in 2009.
Component Growth stage
a
Regression parameters
b
ED
5
6SE
c
ED
10
6SE ED
20
6SEbde
------------------------------------------- g a e ha
21
------------------------------------------
MRC in 2009, TPAC in 2010
Seed yield V2 0.549 14.5 1.22 0.142 60.2 0.73 60.8 —
V5 1.00 45.5 7.19 0.528 60.3 1.10 60.4 2.38 60.7
R2 0.901 60.0 15.6 0.242 60.2 0.529 60.3 1.17 60.7
TPAC 2009
V2, V5, R2 0.520 111 15.9 0.042 60.03 0.169 60.1 0.706 60.7
Plant height V2 1.43 25.5 3.61 1.24 60.6 2.21 60.5 4.87 61.9
V5 0.747 50.8 4.39 0.211 60.09 0.577 60.2 1.74 60.5
R2 0.911 44.0 5.37 0.526 60.2 1.21 60.3 3.10 60.7
MRC in 2009, TPAC in 2010
Seeds m
22
V2 0.175 79.4 1,166,500 0.186 63 11.9 6166 —
V5 1.30 41.0 7.77 1.62 60.04 2.92 60.05 5.71 60.1
R2 1.06 58.6 10.9 1.10 60.4 2.22 60.7 4.74 61.9
TPAC 2009
V2,V5,R2 0.369 318 690 0.009 60.06 0.061 60.4 0.416 62.9
Seed mass V2 28.61 1.38 5.63 — — —
V5 5.12 5.19 0.289 0.365 60.6 — —
R2 2.32 28.88 4.05 — — —
MRC in 2009, TPAC in 2010
Pods m
22
V2 0.110 16.3 212 0.025 60.6 — —
V5 1.45 31.1 7.04 2.11 61.0 3.65 61.2 7.17 62.7
R2 1.01 46.5 9.98 1.15 60.5 2.44 61.1 5.64 63.2
TPAC 2009
V2,V5,R2 0.377 417 1,718 0.014 60.04 0.089 60.3 0.577 61.8
Seeds pod
21
V2 21.36 4.25 1.64 — — —
V5 20.570 15.6 3.41 2.72 612 14.2 6117 —
R2 20.460 43.3 16.8 3.16 63.1 7.35 611 —
Reproductive nodes m
22
V2 20.213 104 13.6 0.073 60.09 0.247 60.3 1.28 62.3
V5 21.69 73.1 2.16 1.21 60.4 1.44 60.3 1.86 60.3
R2 20.384 173 20.1 0.748 60.4 1.32 61.0 2.72 63.2
Pods reproductive node
21
V2 2.55 233.0 6.11 — — —
V5 4.26 264.0 3.68 — — —
R2 2.10 236.8 8.20 — — —
Nodes m
22
V2 0.340 106 146 0.020 60.07 0.162 60.6 1.45 65
V5 1.83 69.1 3.49 0.847 60.3 1.27 60.4 1.94 60.4
R2 1.27 62.8 8.15 1.14 60.6 2.04 60.8 3.82 61
Percent reproductive nodes V2 14.3 22.45 0.995 — — —
V5 214.2 3.77 1.32 — — —
R2 20.454 223 195 10.3 616 16.1 629 —
a
Treatments of dicamba were applied at the Midwest Research Center (MRC) in 2009 and at the Throckmorton Purdue Agriculture Center (TPAC) in 2009 and 2010.
b
Dose response parameters where brepresents the relative slope of the curve, dthe upper limit, and ethe inflection point of the equation. The equation Y5d
(exp[2exp{b(log x2e) }]) described seeds pod
21
, main-stem reproductive nodes m
22
, and percentage of reproductive nodes; the equation Y5d(1 2exp[2exp{b(log x
2log e)}]) was fit to soybean seed yield, plant height, seeds m
22
, seed mass, pods m
22
, seeds pod
21
, pods reproductive node
21
, and main-stem nodes m
22
.
c
Abbreviation: ED
5
,ED
10
ED
20
, effective dose. causing 5, 10, or 20% loss.
Robinson et al.: Soybean response to dicamba N533
Figure 3. Nonlinear regressions of (A) seed yield loss, (B) plant height loss, (C) seeds m
22
loss, (D) seed mass loss, (E) pod m
22
loss, (F) seed pod
21
loss, (G)
reproductive nodes m
22
loss, (H) pod reproductive node
21
loss, (I) nodes m
22
loss, and (J) percentage of reproductive nodes loss as affected by dicamba treatments
applied to soybean at the V2, V5, or R2 growth stages. Studies were conducted at the Throckmorton Purdue Research Center (TPAC) near Lafayette, IN, in 2009 and
2010 and at the Midwest Research Center (MRC) near Fowler, IN, in 2009. Regression parameters and effective dose values can be found in Table 6.
534 NWeed Science 61, October–December 2013
the time of the event will affect the degree of damage. The
yield components in this study that were most affected were
the number of main-stem nodes m
22
, main-stem reproduc-
tive nodes m
22
, pods m
22
, and seeds m
22
. A similar response
of the number of nodes, reproductive nodes, pods, and seed
on seed yield formation has been observed in other studies
(Kahlon et al. 2011; Robinson et al. 2013), which indicates
that these are the principal yield components that can be
influenced during plant growth. Studies of soybean seed yield
reduction need to take an approach to quantify the number of
nodes, reproductive nodes, pods, and seed lost to understand
the whole effect of unintended herbicide exposure on soybean.
One of the challenges in estimating soybean seed yield and
yield loss is that soybean plants may compensate for a decrease
in one yield component by increasing another, making the
analysis of all the important yield components necessary. The
other yield components, seed mass and percentage of
reproductive nodes, are more stable because they are likely
genetically controlled.
Soybean oil and protein concentration have been shown to
be correlated with the maximum air temperature during the
R6 growth stage (Robinson et al. 2009). A reduction in leaf
area from herbicide drift can increase the canopy temperature
and can result in an increase in soybean oil and a decrease in
protein content.
In summary, soybean plants were damaged by dicamba and
as a result the number of nodes, reproductive nodes, pods, and
seeds were reduced, causing a loss in seed yield. The seed yield
loss could be estimated by observing the visual symptomology,
but variability in human ratings, knowledge of when the drift
incident occurred, and variable responses of soybean cultivars
to dicamba can make this method less reliable. Because
soybean plants are extremely sensitive to dicamba, the effects
of tank contamination, residues remaining in the spraying
equipment (Boerboom 2004), and volatility of dicamba will
need to be taken into consideration. Additionally, tank
mixtures of glyphosate and dicamba have caused greater injury
and yield loss than dicamba alone in some cases when
treatments occurred on glyphosate-resistant soybean (Kelley et
al. 2005). Further research examining this relationship may
help explain if greater injury and yield loss will occur when
soybean plants inadvertently come into contact with glypho-
sate and dicamba tank mixtures. Furthermore, mitigating off-
site movement of dicamba by using recommended spraying
techniques and being aware of when sensitive vegetation is
nearby will be of utmost importance to avoid unintended
exposure to dicamba.
Figure 4. Path analysis of soybean yield components affected by dicamba near Lafayette, IN, in 2009 and 2010 and near Fowler, IN, in 2009. Path coefficient (P),
correlation coefficient (r), and the indirect effect of the path coefficient by correlation coefficient (r 3P) values are shown. The bold lines indicate an important trait
affecting its response variable while the green and blue lines are the indirect effects. The important trait affecting its response variable was determined when there was a
large positive correlation, large positive direct effect, and a small negative indirect effect.
Figure 5. Relationship of visual estimates of soybean injury (0 to 100%) at 14
and 28 d after treatment and seed yield loss (%) of soybean plants treated with
dicamba (0 to 22.7 g ha
21
) at the V2, V5, or R2 soybean growth stages. Visual
estimates of injury were taken at 14 and 28 d after treatment. Studies were
conducted near Lafayette, IN, in 2009 and 2010 and near Fowler, IN, in 2009.
Both regressions were significant (P ,0.0001).
Robinson et al.: Soybean response to dicamba N535
Literature Cited
Al-Khatib, K. and D. Peterson. 1999. Soybean (Glycine max) response to
simulated drift from selected sulfonylurea herbicides, dicamba, glyphosate, and
glufosinate. Weed Technol. 13:264–270.
Andersen, S. M., S. A. Clay, L. J. Wrage, and D. Matthees. 2004. Soybean foliage
residues of dicamba and 2,4-D and correlation to application rates and yield.
Agron. J. 96:750–760.
Auch, D. E. and W. E. Arnold. 1978. Dicamba use and injury on soybeans
(Glycine max) in South Dakota. Weed Sci. 26:471–475.
Behrens, M. R., N. Mutlu, S. Chakraborty, R. Dumitru, W. Z. Jiang, B. J.
LaVallee, P. L. Herman, T. E. Clemente, and D. P. Weeks. 2007. Dicamba
resistance: enlarging and preserving biotechnology-based weed management
strategies. Science 316:1185–1188.
Behrens, R. and W. E. Lueschen. 1979. Dicamba volatility. Weed Sci.
27:486–493.
Board, J. E. and H. Modali. 2005. Dry matter accumulation predictors for
optimal yield in soybean. Crop Sci. 45:1790–1799.
Boerboom, C. 2004. Field case studies of dicamba movement to soybeans:
University of Wisconsin. http://www.soils.wisc.edu/extension/FAPM/2004
proceedings/Boerboom3.pdf. Accessed September 3, 2009.
Conley, S. P., P. Pedersen, and E. P. Christmas. 2009. Main-stem node removal
effect on soybean seed yield and composition. Agron. J. 101:120–123.
De Bruin, J. L. and P. Pedersen. 2008. Soybean seed yield response to planting
date and seeding rate in the Upper Midwest. Agron. J. 100:696–703.
Dexter, A., F. Slife, and H. Butler. 1971. Detoxification of 2,4-D by several plant
species. Weed Sci. 19:721–726.
Fehr, W. R. and C. E. Caviness. 1977. Stages of Soybean Development. Special
Report 80 Iowa Agriculture and Home Economics Experiment Station, Iowa
State University. 11 p.
Gerber, C. K., S. M. Brouder, J. J. Camberato, S. N. Casteel, T. Creswell, J. J.
Faghihi, B. C. Joern, W. G. Johnson, C. H. Krupke, J. S. Loven, G.R.W.
Nice, R. L. Nielsen, J. L. Obermeyer, G. E. Ruhl, G. C. Steinhardt, and K. A.
Wise. 2012. Corn and Soybean Field Guide. West Lafayette. IN: Agriculture
Communication Media Distribution Center. 324 p.
Kahlon, C. S., J. E. Board, and M. S. Kang. 2011. An analysis of yield
component changes for new vs. old soybean cultivars. Agron. J. 103:13–22.
Kang, M. S. 1994. Applied Quantitative Genetics. Baton Rouge, LA: M. S. Kang.
157 p.
Kelley, K. B., L. M. Wax, A. G. Hager, and D. E. Riechers. 2005. Soybean
response to plant growth regulator herbicides is affected by other
postemergence herbicides. Weed Sci. 53:101–112.
Knezevic, S. Z., P. H. Sikkema, F. Tardif, A. S. Hamill, K. Chandler, and C. J.
Swanton. 1998. Biologically effective dose and selectivity of RPA 201772 for
preemergence weed control in corn (Zea mays). Weed Technol. 12:670–676.
Knezevic, S. Z., J. C. Streibig, and C. Ritz. 2007. Utilizing R software package for
dose-response studies: the concept and data analysis. Weed Technol.
21:840–848.
Kniss, A. R. and D. J. Lyon. 2011. Winter wheat response to preplant
applications of aminocyclopyrachlor. Weed Technol. 25:51–57.
Kokubun, M. S. and S. Takahashi. 2001. Flower abortion caused by preanthesis
water deficit is not attributed to impairment of pollen in soybean. Crop Sci.
41:1517–1521.
Lingle, S. E. and J. C. Suttle. 1985. A model system for the study of 2,4-D
translocation in leafy spurge. Can. J. Plant Sci. 65:367–377.
Liu, F., C. R. Jensen, and M. N. Andersen. 2004. Drought stress effect on
carbohydrate concentration in soybean leaves and pods during early
reproductive development: its implication in altering pod set. Field Crops
Res. 86:1–13.
Maybank, J., K. Yoshida, and R. Grover. 1978. Spray drift from agricultural
pesticide applications. J. Air Pollut. Control Assoc. 28:1009–1014.
Ritz, C. and J. Strebig. 2011. Package ‘drc’. Available at http://cran.r-project.org/
web/packages/drc/drc.pdf. Accessed June 9, 2011.
Robinson, A. P., S. P. Conley, J. J. Volenec, and J. B. Santini. 2009. Analysis of
high yielding, early-planted soybean in Indiana. Agron. J. 101:131–139.
Robinson, A. P., V. M. Davis, D. M. Simpson, and W. G. Johnson. 2013.
Response of soybean yield components to 2,4-D. Weed Sci. 61:68–76.
Robocker, W. C. 1976. Translocation and metabolism of dicamba in western
bracken. Weed Sci. 24:435–438.
Sciumbato, A. S., J. M. Chandler, S. A. Senseman, R. W. Bovey, and K. L.
Smith. 2004. Determining exposure to auxin-like herbicides. I. Quantifying
injury to cotton and soybean. Weed Technol. 18:1125–1134.
Seber, G.A.F. and C. J. Wild. 1989. Nonlinear Regression. New York: J. Wiley.
768 p.
Sionit, N. K. and J. Paul. 1977. Effect of water stress during different stages of
growth of soybean. Agron. J. 69:274.
Spiess, A. and N. Neumeyer. 2010. An evaluation of R
2
as an inadequate measure
for nonlinear models in pharmacological and biochemical research: a Monte
Carlo approach. BMC Pharmacol. 10:6–17.
Wax, L. M., L. A. Knuth, and F. W. Slife. 1969. Response of soybeans to 2,4-D,
dicamba, and picloram. Weed Sci. 17:388–393.
Weidenhamer, J. D., G. B. Triplett, Jr., and F. E. Sobotka. 1989. Dicamba injury
to soybean. Agron. J. 81:637–643.
Wolf, T. M., R. Grover, K. Wallace, S. R. Shewchuk, and J. Maybank. 1993.
Effect of protective shields on drift and deposition characteristics of field
sprayers. Can. J. Plant Sci. 73:1261–1273.
Received December 20, 2012, and approved May 10, 2013.
536 NWeed Science 61, October–December 2013
