Weed Control With Preemergence Herbicides in Azuki Bean

Three field experiments were completed over a three-year period (2019 to 2021) in Ontario, Canada to develop weed management programs in azuki bean with herbicides (pendimethalin, S -metolachlor, halosulfuron, and imazethapyr) applied alone and in combination, and metribuzin, applied preemergence (PRE). At 2 and 4 weeks after emergence (WAE), there was ≤ 8% azuki bean injury from the herbicide treatments evaluated, with the exception of the treatments that included S -metolachlor which caused up to 19% azuki bean injury. Pendimethalin (1080 g ai ha -1 ) and S -metolachlor (1600 g ai ha -1 ) controlled green foxtail 83-94% but provided poor control of common lambsquarters, wild mustard, redroot pigweed, common ragweed, and flower-of-an-hour. Imazethapyr (75 g ai ha -1 ) controlled common lambsquarters, wild mustard, redroot pigweed, and flower-of-an-hour 90-100% but provided 76-82% control of common ragweed and green foxtail. Halosulfuron (35 g ai ha -1 ) controlled wild mustard 100%, common ragweed 81-84%, common lambsquarters 77-83%, flower-of-an-hour 72-75%, redroot pigweed 59-72%, and green foxtail 19-23%. The tankmix of pendimethalin + S -metolachlor controlled green foxtail and common lambsquarters 87-97% but the control was only 23- 83% on wild mustard, redroot pigweed, common ragweed, and flower-of-an-hour. The tankmixes of pendimethalin + imazethapyr and pendimethalin + S -metolachlor + imazethapyr provided 90-100% control of common lambsquarters, wild mustard, redroot pigweed, flower-of-an-hour, and green foxtail, and 78-87% control of common ragweed. The tankmixes of pendimethalin + halosulfuron and pendimethalin + S -metolachlor + halosulfuron controlled common lambsquarters and wild mustard 91-100%, green foxtail 76-95%, flower-of-an-hour 70-94%, redroot pigweed 68-91%, and common ragweed 78-79%. Metribuzin (280 g ai ha -1 ) controlled common lambsquarters, wild mustard, redroot pigweed, common ragweed, flower-of-an-hour, and green foxtail up to 94, 98, 81, 58, 98, and 61% respectively; control improved to 99, 100, 97, 84, 99, and 83%, respectively when the rate was increased to 560 g ai ha -1 . Generally, weed density and dry biomass reflected the level of weed control. Weed interference reduced azuki bean yield by 91% in this study. Generally, azuki bean yield reflected the level of weed control.


Introduction
Azuki bean [Vigna angularis (Willd.) Ohwi & Ohashi] is a small, red bean with a sweet flavor that is mainly used in confectionery food products that are popular with consumers in the orient (Dikshit et al., 2005;Lumpkin et al., 1994). Although azuki bean production occurs mainly in China, Japan, Korea, and India, the production area is limited to about 700,000 ha annually in those countries (Pandiyan et al., 2021). Demand for azuki beans from the Pacific Rim countries has resulted in an increase in azuki bean production in Ontario, Canada for the export market as the province has a suitable environment for azuki bean growth and it is a profitable crop that can be incorporated in current field crop production rotations in the province (Greig, 2019;Lynch, 2019). Azuki bean production increased from 3100 ha in 2017 to 8100 ha in 2019 in Ontario (Greig, 2019). Azuki bean is a short, bushy plant and therefore is vulnerable to weed interference and the associated yields losses which can be as high as 71%, similar to other dry bean market classes (McClary et al., 1993;Pandiyan et al., 2021;Soltani et al., 2018). Presently, only two soil-applied herbicides, pendimethalin and imazethapyr, are registered for weed management in azuki bean production in Ontario (OMAFRA, 2021). More research is needed to find new herbicide options that provide efficacious control of troublesome weeds in azuki bean under Ontario environmental conditions. Halosulfuron, S-metolachlor, and metribuzin are additional soil-applied herbicides that have potential for weed management in azuki bean production in Ontario (OMAFRA, 2021).
Metribuzin is a herbicide from the triazinone class that inhibits electron transport in photosystem II, resulting in an increase reactive oxygen species and subsequent lipid peroxidation in susceptible plants (Trebst, 2008). Metribuzin controls problematic weeds including common ragweed, common lambsquarters, wild mustard, redroot pigweed, velvetleaf, cocklebur (Xanthium strumarium L.), and suppression of annual grasses such as barnyardgrass and green foxtail (OMAFRA, 2021;Shaner, 2014).
Azuki bean producers need to combine grass herbicides such as pendimethalin and S-metolachlor with broadleaved herbicides such as imazethapyr and halosulfuron to obtain acceptable broad-spectrum weed control in their production. Few studies have evaluated weed management programs that include metribuzin and two-way or three-way tankmixtures of pendimethalin, S-metolachlor, imazethapyr, and halosulfuron applied preemergence in azuki bean. These tankmixtures can potentially increase weed control efficacy and provide consistent broad-spectrum control of troublesome weeds in azuki bean.
The purpose of this study was to evaluate crop injury and weed control efficacy in azuki bean with pendimethalin, S-metolachlor, imazethapyr, halosulfuron applied alone and in combination, and metribuzin applied PRE.

Materials and Methods
A field study was completed in 2019, 2020, and 2021 at the Huron Research Station near Exeter, Ontario, Canada (43°19′1.21″N, 81°30′3.87″E). The soil was a Brookston clay loam (Orthic Humic Gleysol, mixed, mesic, and poorly drained). Seedbed preparation consisted of fall moldboard plowing followed by seedbed preparation in the spring with a field cultivator with rolling basket harrows.
Azuki bean injury was assessed 2 and 4 weeks after crop emergence (WAE) and weed control for each species was assessed 4 and 8 WAE on a scale of 0-100% with 0 representing no azuki bean injury/weed control and 100 representing complete azuki bean death/weed control. At 8 WAE, weed density and dry weight were determined from two 0.5 m 2 quadrats placed randomly in each plot. Weeds were counted, cut at the soil surface, and were separated by species and dried at 60 °C to constant moisture, and then weighed. A small plot combine was used to harvest azuki beans at maturity. Azuki bean yield was adjusted to 13% seed moisture content.
Data analysis was conducted using the GLIMMIX procedure in SAS (Ver. 9.4, SAS Institute Inc., Cary, NC) and the level of significance was set at 0.05 for all comparisons. Herbicide treatment was the fixed effect and site-year (environment), environment by treatment interaction, and replicate within environment were the random effects for the generalized linear mixed model. The distribution selected for each parameter was the one that most closely met the assumptions of analysis, based on visual examination of studentized residual plots for the assumption of homogeneity of variance, and the normal probability plot and Shapiro-Wilk statistic for normality. The Gaussian distribution was used for azuki bean injury 2 WAE, redroot pigweed control, common ragweed control, and azuki bean yield. Azuki bean injury 4 WAE, lambsquarters control, wild mustard control, and flower-of-an-hour control were arcsine square-root transformed prior to analysis with the Gaussian distribution. The lognormal distribution was used for density and dry biomass for all weed species and azuki bean moisture at harvest. The least-square means were subjected to the Tukey-Kramer adjustment prior to pairwise comparisons of least-square means. Treatments with zero variance across all environments were excluded from the analysis, including the non-treated and weed-free controls for crop injury and weed control, the weed-free control for weed density and dry biomass, as well as certain herbicide treatments for density and dry biomass. The P-value that is shown in the LSMEANS table still allowed comparisons of each treatment with the value of zero. The least-square means were back-transformed, if necessary, for presentation.