Response of four market classes of dry bean to acifluorfen, bentazon, and bentazon/acifluorfen applied postemergence

Three distinct studies consisting of a total of 11 field trials were carried out from 2016 to 2019 in Ontario to assess the tolerance of four market classes of dry bean to acifluorfen, bentazon, and bentazon/acifluorfen applied postemergence (POST) at various rates. In Study 1, acifluorfen (265 and 530 g ai ha−1) and bentazon (577 and 1144 g ai ha−1) caused up to 13% and 8% visible injury, respectively, but did not affect the relative stand, height, seed moisture, or yield of dry bean. Bentazon/acifluorfen (842 and 1684 g ai ha−1) caused up to 32% visible injury in dry bean but did not affect the relative stand, height, moisture, or seed yield except at the 2× rate which reduced dry bean biomass, height, and yield 43%, 11%, and 11%, respectively. The preformulated mixture of bentazon/acifluorfen caused a synergistic increase in dry bean injury. In Study 2, acifluorfen (600 and 1200 g ai ha−1) and bentazon (1080 and 2160 g ai ha−1) injured dry bean up to 12% and to 9%, respectively, but did not affect the relative stand, height, moisture, or yield. Bentazon/acifluorfen (840 and 1680 g ai ha−1) injured dry bean up to 16% but did not affect the relative stand, height, moisture, or yield of dry bean except at the high rate which reduced biomass, height, and yield 17%, 9%, and 5%, respectively. In Study 3, acifluorfen (420 and 480 g ai ha−1) and bentazon (1080 and 2160 g ai ha−1) injured dry bean up to 24 and 16%, respectively, but did not affect the relative stand, height, seed moisture content, or yield except for the relative biomass which was reduced up to 17%. Bentazon/acifluorfen (840 and 1680 g ai ha−1) injured dry bean up to 45% but did not affect the relative stand, height, seed moisture, or yield except for biomass, height, and seed yield which were decreased 44%, 11%, and 12% at the high rate. This study concludes that acifluorfen alone or preformulated with bentazon applied POST at rates assessed cannot be safely used dry bean.


| INTRODUCTION
Dry bean (Phaseolus vulgaris L.) is a major legume crop produced in Canada, mostly for exportation to the United States, Europe, and Asia (Hensall District Co-operative, 2020). Growers in Canada produced approximately 341,000 metric tons of dry bean in 2018/2019 crop years (Statista, 2020). Approximately one third of the dry bean grown in Canada is produced in Ontario. In 2017, growers in Ontario produced approximately 121,000 tons of dry bean on 55,000 ha with a value of nearly Can$93,000,000 (Ontario Ministry of Agriculture and  (Soltani et al., 2018). Despite such great losses, there are currently few herbicides registered for weed control in dry bean. Dry bean is not a large market in terms of hectarage in North America compared with other field crops, and agro-chemical companies have not invested a lot of resources to develop new herbicides for this niche market crop. New research is needed to determine the tolerance of dry bean to herbicides registered for the control of weeds in other crops such as soybean.
Acifluorfen is a Group 14 diphenyl ether contact herbicide that is currently registered for use in soybean production in Ontario (OMAFRA, 2020). It is taken up through foliage and inhibits the enzyme protoporphyrinogen oxidase (PPO or Protox) that is needed for the synthesis of chlorophyll and heme (involved in the maintenance of redox balance and the energy state in cells). Acifluorfen controls key annual broadleaved weeds in Ontario (OMAFRA, 2020).
Currently, only three postemergence herbicides are available for broadleaved weed control in dry bean, bentazon, fomesafen, and halosulfuron. Bentazon does not provide adequate control of redroot pigweed, waterhemp, common ragweed, giant ragweed, and Canada fleabane (OMAFRA, 2020). Additionally, fomesafen provides weak/inadequate control of Canada fleabane, cocklebur, common lambsquarters, and velvetleaf. Halosulfuron, applied POST, is weak on annual nightshades and common lambsquarters in dry bean production.
Acifluorfen and bentazon/acifluorfen premixed are not registered to be used in dry edible beans in Ontario. Bentazon/acifluorfen premixed can control common annual broadleaved weeds in dry bean.
Tolerance of dry bean to POST application of acifluorfen and acifluorfen/bentazon premixed needs to be determined under Ontario environmental conditions. This manuscript summarizes the results of three studies on the tolerance of dry edible beans to acifluorfen, bentazon and bentazon/acifluorfen applied POST. The purpose of this research was to ascertain the tolerance of azuki, kidney, small red, and white beans to acifluorfen, bentazon, and bentazon/acifluorfen applied POST at various rates.

| MATERIALS AND METHODS
Three distinct studies were established from 2016 to 2019 in southwestern Ontario. Study 1 included a total of four field trials carried out carried out at Exeter, ON (2016ON ( , 2017ON ( , and 2018  Experiments in each study were designed as a split-block, with treatment (TRT) assigned as the whole plot factor and dry bean type (DB) as subplots. Treatments for Studies 1-3 are shown in Tables 1-4,   5 and 6, and 7-9, respectively. The ratio of acifluorfen to bentazon in the formulation mixture evaluated was 1:2.18 which included 265 g ai ha À1 of acifluorfen + 577 g ai ha À1 of bentazon = 842 g ai ha À1 of acifluorfen/bentazon (representing the 1Â rate) and 530 g ai ha À1 of acifluorfen + 1154 g ai ha À1 of bentazon = 1684 g ai ha À1 of acifluorfen/bentazon (representing the 2Â rate). Therefore, the rates of bentazon and acifluorfen in the premix are the same rate used in the stand-alone products. The current commercial formulations of acifluorfen and bentazon (as used in this study) do not require the addition of an adjuvant (the adjuvant is already included in the commercial formulation). In contrast, an adjuvant must be added to the commercial formulation of acifluorfen/bentazon as recommended on the product label. The adjuvant used and rate applied was recommended by the manufacturer at the time the study was initiated.
The experimental plots were 6.0 m wide (eight rows spaced 0.75 m apart) and 10.0 m long at Exeter and 8.0 m long at Ridgetown.
Herbicides were sprayed POST at the 1-2 trifoliate stage (3-4 weeks after seeding) with a CO 2 -pressurized backpack sprayer (calibrated to deliver 200 L ha À1 at 240 kPa). The spray boom was 2.5 m long equipped with six ultralow drift nozzles spaced 0.5 m apart producing a spray width of 3.0 m. All plots were maintained free of weeds during the entire season.
Dry bean injury was determined at 1, 2, 4, and 8 weeks after herbicide treatment (WAT) on a scale of 0% (no visible injury) to 100% (plant death). The relative plant stand (percent of the control) and dry weight (biomass) were assessed 2 WAT by determining the number of plants in 1 m row within each subplots, harvesting the aboveground portion and drying at 60 C (48 h). Average plant height was determined at 5 WAT (10 randomly selected plants per subplot). A smallplot combine was used to harvest dry bean at maturity.
For all studies, the GLIMMIX procedure in Statistical Analysis Systems (SAS, 2013) was used to carry out a mixed model analysis of variance for each response variable. Fixed effects consisted of TRT, DB, and their interaction. Random effects consisted of environment (yearlocation combinations), TRT by DB by environment interaction, replicate nested within environment, and its interaction with TRT. The significance of fixed and random effects was ascertained with an F test and likelihood ratio test, respectively. A significance level of alpha = 0.05 was chosen for all statistical analyses. Different distributions for each response variable were assessed using studentized residual plots and the Shapiro-Wilk statistic; in each case, the distribution that best met the variance analysis assumptions (normally distributed, homogeneous residuals with a mean of zero) was selected. All response variables except dry bean visible injury were converted to a percent of the untreated control within each replicate in order to minimize year-to-year variation in absolute measurements.
For analysis of percent injury at 1, 2, 4, and 8 WAT, untreated control was excluded due to zero variance.
For Study 1, the relative dry bean stand, biomass, height, and yield followed a Gaussian distribution. An arcsine square root transformation and Gaussian distribution was used to analyze estimates of percent visible dry bean injury 1, 2, 4, and 8 WAT. For each variable, dry bean responses to the two combined bentazon/acifluorfen treatments were tested for additivity, synergism, or antagonism. Expected values for percent dry bean visible injury were calculated for each rate of the herbicide combination by replicate using Colby's equation (Colby, 1967): In this equation, E is assumed to be the expected dry bean visible injury for the herbicide combination and G and H are the dry bean Included assist (1.5 L ha À1 ). c Included assist (3.0 L ha À1 ). *Significant difference of P < 0.05, between observed and expected values based on a two-sided t test. **Significant difference of P < 0.01, respectively, between observed and expected values based on a twosided t test.
visible injury with bentazon and acifluorfen, respectively, when applied individually. For the other parameters, expected values were determined with a modified form of Colby's equation, expressing data as a percentage of the control: where E 1 is the expected the relative dry bean stand, biomass, height, moisture, and yield with the herbicide combination and G 1 and H 1 are the measured parameter values with bentazon and acifluorfen, respectively, when applied individually.
All expected values were compared with the corresponding observed value using a two-sided t test at P < 0.05. The effect of the herbicide combination was considered to be additive if the t test was nonsignificant. If the observed percent visible injury or the relative moisture was higher than expected, or the relative stand, biomass, height, or yield was lower than expected, the effect of the herbicide combination was considered to be synergistic. If the converse was true, the effect of the herbicide combination was considered antagonistic.
For Study 2, the relative crop stand, biomass, height, and yield followed a Gaussian distribution. An arcsine square root transformation and Gaussian distribution was used to analyze estimates of percent visible dry bean visible injury 1, 4, and 8 WAT. The Poisson distribution was used to analyze estimates of percent visible dry bean injury 2 WAT. Recognizing that overdispersion is a concern with the Poisson distribution, in particular, the Pearson chi-square/df was confirmed to be less than 1 for all dry bean injury. The relative crop biomass and seed moisture content were analyzed using a lognormal distribution.
For Study 3, the relative dry bean stand, height, and followed a Gaussian distribution. The Poisson distribution was used to analyze estimates of percent visible dry bean visible injury 1, 2, 4, and 8 WAT.
Recognizing that overdispersion is a concern with the Poisson T A B L E 2 Least square means and significance of main effects and interaction for stand count, above ground biomass per m of row and per plant, height, moisture, and yield relative to the untreated control of four dry bean market classes treated with bentazon, acifluorfen, and bentazon/acifluorfen in trials conducted near Exeter, ON (2016ON ( -2018  Included assist (3.0 L ha À1 ). *Significant difference of P < 0.05, between observed and expected values based on a two sided t test. **Significant difference of P < 0.01, between observed and expected values based on a two-sided t test.
distribution in particular, the Pearson chi-square/df was confirmed to be less than 1 for all dry bean injury. A lognormal distribution was used for the analysis of relative biomass and seed moisture.
For Studies 1-3, pairwise comparisons of least square means were adjusted using the Tukey-Kramer method. If the TRT by DB interaction was nonnegligible, comparisons among simple effects were T A B L E 3 Lease square means for percent injury 4 WAT of four dry bean market classes treated with bentazon, acifluorfen and bentazon/ acifluorfen near Exeter, ON (2016ON ( -2018

| Study 1
This study was carried out mainly to determine if there is a synergistic increase in injury of dry bean when bentazon is combined with acifluorfen.

| Main effects of herbicides
There was a significant impact of the herbicides on the visible injury, biomass, height, moisture, and yield (Tables 1 and 2). A synergistic increase in injury with the preformulated mixture for almost all the parameters in dry bean evaluated was observed (Tables 1-3).
Bentazon caused 5%, 3%, 2%, and 0% injury at 577 g ai ha À1 and 8%, 5%, 3%, and 1% injury at 1154 g ai ha À1 assessed 1, 2, 4, and 8 WAT, respectively. Injury decreased with time. Bentazon did not affect the relative stand, height, moisture, or seed yield (Table 2). Note: Means followed by the same letter within a column (a-e) or row (X-Z) are not significantly different according to a Tukey-Kramer multiple range test at P < 0.05. Rows without an uppercase letter have no differences among market classes. Abbreviations: SR, small red; WAT, weeks after herbicide application. a Included assist (1.5 L ha À1 ). b Included assist (3.0 L ha À1 ).
Bentazon/acifluorfen applied at 842 g ai ha À1 did not affect the relative stand, height, moisture, or yield. Bentazon/acifluorfen applied at 1684 g ai ha À1 did not decrease plant stand but decreased dry bean height and yield 40% and 11%, respectively ( Table 2). The seed moisture content was also increased by 12% indicating delayed maturity (Table 2). There was a synergistic decrease in relative height (1Â and 2Â rate) and yield (2Â rate) and a synergistic increase in seed moisture content (2Â rate) with the preformulated mixture of bentazon/ acifluorfen (Table 2).

| Main effects of herbicides
Acifluorfen applied at 600 and 1200 g ai ha À1 caused 7% and 12% injury at 1 WAT, visible injury decreased with time, and there was 0.2% and 2% visible injury at 8 WAT, respectively (Table 5).
Acifluorfen did not affect the relative dry bean stand, biomass, height, seed moisture content, or yield of dry bean evaluated (Table 5).
Bentazon applied at 1080 and 2160 g ai ha À1 injured dry bean 6% and 9% at 1 WAT, but visible injury decreased with time, and there was only 1% and 2% visible injury at 8 WAT, respectively (Table 5). Bentazon did not affect the relative dry bean stand, dry weight plant À1 , height, moisture, or yield of dry bean evaluated (Table 5). Bentazon reduced dry bean dry weight m row À1 (2Â rate).
Bentazon/acifluorfen at 840 and 1680 g ai ha À1 caused 10% and 16% injury at 1 WAT, but the visible injury decreased with time, and there was only 1% and 3% visible injury at 8 WAT, respectively (Table 1). Bentazon/acifluorfen at the 840 g ai ha À1 did not affect the relative dry bean stand, biomass, plant height, moisture, or seed yield of dry bean evaluated (Table 5). Bentazon/acifluorfen at 1680 g ai ha À1 did not decrease plant stand and biomass (g m row À1 ), but there was a reduction in dry weight plant À1 , height, and seed yield of 17%, 9%, and 5%, respectively (Table 5). The seed moisture content was also increased by 9% indicating delayed maturity (Table 5).

Visible injury
Generally, higher injury was observed in azuki bean in comparison with the other dry bean market classes with bentazon and bentazon/ acifluorfen at rates evaluated ( Table 6). Responses of other dry bean types were generally similar with all herbicide treatments evaluated at 2 and 4 WAT (Table 6). At 2 WAT, bentazon caused greater injury in kidney bean in comparison with small red and white beans (Table 6).

| Study 3
The rate of acifluorfen was reduced in Study 3 due to unacceptably high injury in Study 2.

Visible injury
Dry bean injury was variable based on the herbicide treatment and dry bean market class (Table 8). Injury was similar among bean types at 1, 2, and 8 WAT (Table 8). At 4 WAT, acifluorfen caused higher injury in kidney bean (6%) than azuki bean (3%). Generally, bentazon produced higher injury in azuki bean in comparison with the other bean types. The injury with bentazon/acifluorfen was comparable among dry bean types at 1, 2, 4, and 8 WAT (Table 8).
Other research has reported 7% visible injury with no decrease in azuki bean height or yield with acifluorfen applied POST at 600 or 1200 g ai ha À1 (Stewart et al. 2010). In the same study, POST application of bentazon (1080 or 2160 g ai ha À1 ) injured azuki bean 66% and decreased azuki bean biomass, height, and seed yield 90%, 27%, and 57%, respectively (Stewart et al., 2010). Variable responses have been reported with bentazon on P. vulgaris dry bean species. VanGessel et al. (2000) found up to 20% injury in dry bean with bentazon, but Soltani et al. (2005) reported only 3% injury with bentazon in dry bean. Additionally, Wall (1995) reported 21% decrease in dry bean yield with the POST application of bentazon. However, Blackshaw et al. (2000) and Burnside et al. (1994) showed no decrease in yield with the POST application of bentazon in dry bean.

| CONCLUSION
Acifluorfen at rates assessed caused significant initial injury but generally had no effect on the relative dry bean stand, biomass, plant height, moisture, or seed yield. Bentazon at rates assessed caused significant initial injury in azuki bean but generally did not affect the final yield of other dry bean types evaluated. However, bentazon/ acifluorfen preformulated applied POST caused significant initial injury and a reduction on of relative dry bean biomass, height, and seed yield of the dry bean studied. There was a synergistic increase in injury with the preformulated mixture of bentazon/acifluorfen for most parameters evaluated. Injury responses to acifluorfen were generally similar among dry bean studied. However, there was generally higher injury responses in azuki bean in comparison with other dry bean types evaluated with bentazon and bentazon/acifluorfen premixed at rates evaluated. The visible injury was generally comparable between kidney, small red, and white beans with all herbicide treatments studied. This study concludes that acifluorfen or bentazon/acifluorfen at rates assessed cannot be safely used in dry bean production.

ACKNOWLEDGMENT
Financial support for this research was provided by the Ontario Bean Growers.