Index | Search | Home | Table of Contents

McClary, D.C., A.N. Hang, G.C. Gilliland, J.M. Babcock, T.A. Lumpkin, A.G. Ogg, and L.K. Tanigoshi. 1993. Herbicides for azuki production. p. 590-594. In: J. Janick and J.E. Simon (eds.), New crops. Wiley, New York.

Herbicides for Azuki Production*

D.C. McClary, A.N. Hang, G.C. Gilliland, J.M. Babcock, T.A. Lumpkin, A.G. Ogg, and L.K. Tanigoshi

    1. Greenhouse Study
    2. Field Studies
    1. Greenhouse Study
    2. Field Studies
  4. Table 1
  5. Table 2
  6. Table 3

Research and limited production of azuki, Vigna angularis (Willd.) Ohwi & Ohashi, has been conducted in Washington State since 1987. Farmers reported that several herbicides used commonly in Phaseolus spp. bean production were phytotoxic to azuki. Inadequate weed control were also observed in commercial azuki fields and is a limiting factor to commercialization. Greenhouse and field herbicide evaluation studies were initiated to identify herbicides that provide effective control of common eastern Washington weeds and which were not phytotoxic to azuki.


Greenhouse Study

The trial was conducted from Dec. 7, 1988 to Jan. 19, 1989, and consisted of a total of 13 treatments, one pot per replication, four pots per treatment in a completely randomized design. Azuki seeds first were screened for uniformity of size and freedom from mechanical damage and then approximately 500 unblemished seeds of uniform size were imbibed with water for two days to identify those that were hard-seeded. Five imbibed seeds without radicle protrusion were sown in each 20 cm diameter pot filled with herbicide-treated air-dried Warden silt loam (coarse-silty, mixed, mesic, Xerollic, Camborthid) for the seven preplant treatments; initial planting for five preemergence treatments, the nontreated check and one post emergence treatment were sown into air-dried soil only. The preemergence treatments were incorporated by hand to label specification for depth while the post emergence treatment was applied after the first trifoliates was fully expanded with a CO2 backpack sprayer calibrated to deliver an equivalent of 252 liters/ha. Pots were maintained in a greenhouse that provided 12 hours of supplemental lighting using 400 watt high pressure sodium vapor bulbs and temperatures of 20°C (day) and 150°C (night). Subirrigation was used until seedling emergence to prevent soil crusting and then pots were irrigated overhead on demand. Fertilized water was applied once a week; pots were randomly rotated twice a week. Dates of individual hypocotyl (damaged and healthy) emergence, final healthy seedling number, plant height, leaf area index, shoot, and root total dry matter weight were recorded. Hypocotyl emergence data per pot was converted into speed of emergence rates using Maguire's (1962) formula Ni/Ti where Ni = ith number of seedlings present (i = 1) and Ti = ith days after planting (i = 1). Rates were then divided by the hypothetically perfect speed of emergence rate of five based on all five azuki seeds per pot germinating and emerging one day after sowing.

Field Studies

Two trials were conducted in 1990, using overhead or rill irrigation, at Prosser, WA (elevation: 300 m; location: 460° 15'N, 1190° 50' W) on Warden silt loam. Four preplant incorporated herbicide treatments were tested along with checks in a completely randomized design replicated three times for the overhead irrigation trial; a total of 12 treatments were evaluated under rill irrigation using a randomized complete block design with four replications. Split plots for both trials within each treatment measured 2.3 m wide (four 56 cm rows) by 4.6 m long and were seeded with either 'Erimo' and 'Hatsune', two commercial Japanese azuki cultivars. Treatments were applied to both trials on 22 May, to preirrigated soil at a spray volume of 127 liters/ha and then incorporated to a depth of 7.6 cm. Planting on May 29 was at the rate of 237,000 seeds/ha. Glyphosate was applied preemergence on 8 June, in 127 liters/ha spray volume to four overhead irrigation herbicide treatments. Bentazon was applied on 22 June, to 2 rill trial treatments when first seedling trifoliates were emerging. Approximately 46 cm of total water was available to plants in the overhead trial during the growing season (40 h x 1.42 cm/h x 80% efficiency) while rill irrigation trial azuki had about 31 cm available water (102 cm for seven 36 h set x 30% efficiency). 'Hatsune' was harvested on Sept. 20 and 'Erimo' on 4 Oct.; both cultivars were thrashed on Oct. 10-12 with a Hedge 140 combine. Weekly emergence counts, days to anthesis, qualitative assessments of phytotoxicity and efficacy, the total dry matter weight ratio of azuki-to-weeds, seed yield, and ratio of treatment-to-handweeded check seed yield were recorded.


Greenhouse Study

Azuki treated with chloramben, metolachlor, EPTC + ethalfluralin, and ethalfluralin (0.84 kg ai/ha) emerged faster compared to nontreated plants (Table 1). The checks also had significantly lower emergence percentages than the pendimethalin, ethalfluralin (1.45 kg ai/ha), and chloramben treatments. These results suggest that some herbicides might improve early azuki stand establishment by suppressing soil pathogens. Chloramben, however, caused distortions and enlargements of some seedling hypocotyls. Ethalfluralin (1.45 kg ai/ha), metolachlor + ethalfluralin, EPTC + ethalfluralin, EPTC, and EPTC + trifluralin treatments stunted azuki seedlings as compared to nontreated plants. Metolachlor + ethalfluralin treated azuki had significantly lower root weights as compared to nontreated seedlings. EPTC alone, in combination with ethalfluralin or trifluralin, and the four week post planting application of bentazon reduced azuki growth as compared to the checks. Pendimethalin (1.12 kg ai/ha), imazethepyr (0.053 kg ai/ha), ethalfluralin (0.84 kg ai/ha), and trifluralin (0.84 kg ai/ha) were not phytotoxic to azuki seedlings grown in the greenhouse.

Field Studies

Overhead irrigation herbicide trial. Average azuki seed yields ranged from 211 kg/ha in the nonweeded checks to 3,165 kg/ha for handweeded treatments. Pendimethalin+imazethepyr was significantly less effective in controlling all weeds, especially barnyard grass, Echinochloa crusgalli (L.) Beauv., than ethalfluralin+imazethepyr (Table 2). There were no other significant differences among treatments for seedling emergence, days to anthesis, number of harvest plants, phytotoxic effects, the ratio of azuki-to-weed dry matter weights, azuki seed yield, and the ratio of treatment-to-handweeded check yield. The insignificant differences in phytotoxicity and yield between treatments with or without imazethepyr demonstrate that azuki has a high tolerance for this herbicide which was mistakenly applied at a rate ten times the recommended label specifications. 'Erimo' had a significantly higher ratio of treatment-to-handweeded check seed yield than 'Hatsune'; 'Hatsune' produced significantly higher yields than 'Erimo' in the handweeded checks but yield differences between varieties under the four different herbicide treatments were insignificant.

Rill irrigation herbicide trial. Azuki seed yields ranged from 665 kg/ha for the nonweeded checks to 2,203 kg/ha for the handweeded checks. There were significantly fewer plants 5.3 m2 at harvest in ethalfluralin + imazethepyr and pendimethalin + bentazon treated plots as compared to the handweeded checks (Table 3). There were no significant differences in early season seedling counts among these three treatments suggesting that ethalfluralin + imazethepyr and pendimethalin + entazon caused some seedlings to die from either direct herbicide related phytotoxicity or diseases aided by herbicide damage/stress to seedlings. Both imazethepyr treatments, pendimethalin + imazethepyr, ethalfluralin + imazethepyr, and pendimethalin + bentazon treatments caused excessive azuki injury. Weed control in imazethepyr (0.53 kg ai/ha) and trifluralin plots were significantly less effective than that in handweeded check plots, especially for E. crusgalli.

A partitioning of yield with orthogonal contrasts demonstrated that azuki treated with either ethalfluralin or pendimethalin alone yielded significantly better than azuki in which these two herbicides were applied in combination with either imazethepyr or bentazon; azuki treated with either ethalfluralin alone or in combination with imazethepyr or bentazon yielded significantly more than azuki treated with pendimethalin in combination with both of these other herbicides as well as alone. Differences in emergence counts, days from sowing to anthesis, the ratio of azuki-to-weed total dry matter weight and yield ratio of treatments-to-handweeded checks among treatments were not significant at the 5% level. 'Hatsune' had a significantly higher ratio of treatment-to-handweeded check seed yield than 'Erimo' in this trial which is the opposite of the overhead irrigation results; significant varietal differences in water requirements during the growing season might account for this result and is being studied at this time.

Washington State azuki producers can use ethalfluralin (1.46 kg ai/ha), which is registered for use on dry edible beans such as azuki, to minimize phytotoxicity and maximize weed control in azuki production but not in combination with bentazon. The mistaken preplant incorporation of imazthapyra at rates ten times labeled rates without any phytotoxicity suggests that this herbicide has potential for azuki production, but must be field tested first at recommended rates.


Maguire, J.D. 1962. Speed of germination: Aid in selection and evaluation for seedling emergence and vigor. Crop Sci. 2:176-177.
*We acknowledge the IMPACT (International Marketing Program for Agricultural Commodities and Trade) Center, Washington State University, Pullman WA for providing major financial support to conduct this research project and ongoing herbicide evaluation studies.
Table 1. Results of greenhouse herbicide treatments on azukiz.

Treatments Speed of emergencey No. final plants Plant height (cm) Leaf area index (cm2) Root weight (g) Shoot weight (g)
Ethalfluralin (0.84) 0.40ab 4.5abc 21.7abcd 111bcd 0.055ab 0.40a
Ethalfluralin (1.45) 0.36abc 4.8ab 14.5d 98bcd 0.049abc 0.27b
Trifluralin (0.84) 0.37abc 4.0abcd 25.3abc 180a 0.051ab 0.39a
Chloramben (2.52) 0.46a 4.8ab 27.3ab 135abc 0.057a 0.38a
EPTC (3.36) 0.23d 3.0de 1.7e 3e 0.038abcd 0.03c
EPTC (3.36) + Ethalfluralin (0.84) 0.41ab 4.3abcd 2.6e 3e 0.031bcde 0.05c
EPTC (3.36) + Trifluralin (0.84) 0.28bcd 3.0de 0.9e 0e 0.020de 0.01c
Metolachlor (1.96) 0.41ab 2.5e 18.5cd 85cd 0.025cde 0.22b
Pendimethalin (1.12) 0.34abcd 5.0a 22.9abcd 159ab 0.039abcd 0.42a
Imazethepyr (0.053) 0.37abc 4.5abc 29.6a 160ab 0.060a 0.46a
Metolachlor (1.96) + Ethalfluralin (0.84) 0.22d 2.3e 4.3e 22e 0.001e 0.04c
Pendimethalin (1.12) + Imazethepyr (0.053) 0.31bcd 4.0abcd 19.9bcd 151ab 0.031bcde 0.38a
Post emergence
Bentazon (1.12) 0.30bcd 3.5bcde 18.5cd 53de 0.025cde 0.19b
Check 0.26cd 3.3cde 23.9abc 139abc 0.044abcd 0.41a
zMeans within each column following by the same letter are not significantly different by LSD at 0.05.
ySpeed of emergence per pot divided by the perfect emergence rate of five.

Table 2. Results of herbicide treatments under overhead irrigation on azukiz.

Treatments (kg ai.ha-1)y Plant number (5.3 m2) Phytotoxicity rating (%)x Efficacy rating (%)x Azuki TDM/Weed TDM Yield (kg/ha) Treatment/
handwd ck. (kg ai/ha)
Ethalfluralin (1.45) + Glyphosate (0.43) 24a 3.2b 95.3ab 1043ab 3404a 1.14a
Ethalfluralin (1.12) + Imazethepyr (0.35) + Glyphosate (0.43) 21a 38.6a 99.1a 2189a 2378b 0.77b
Pendimethalin (0.84) + Imazethepyr (0.35) + Glyphosate (0.43) 23a 21.5ab 89.9b 3b 2861ab 0.90ab
Trifluralin (0.84) + Imazethepyr (0.35) + Glyphosate (0.43) 18a 6.8ab 97.5ab 34b 3087ab 1.01ab
Handweeded check 26a --- 100a 179ab 3331a 1.00
zMeans within each column followed by the same letter are not significantly different by LSD at 0.05. Only efficacy data is significantly different using protected LSD procedures requiring an initial analysis of varian (ANOVA) type 1 error probability of less than 0.1.
yEthalfluralin, imazethepyr, pendimethalin, and trifluralin were preplant incorporated; Glyphosate (kg ai.ha-1) with a 5% nonionic surfactant (X-77) was applied preemergent to azuki.
xQualitative analysis of:
Phytotoxicity: (0% = no crop injury; 100% = complete crop injury).
Herbicide efficacy: (0% = no weed control; 100% = complete weed control).

Table 3. Results of herbicide treatments under rill irrigation on azukiz.

Treatments (kg ai.ha-1)y Plant number (5.3 m2) Phytotoxicity rating (%)x Efficacy rating (%)x AzukiTDM/Weed TDM Yield (kg/ha) Treatment/
handwd ck. (kg ai/ha)
Ethalfluralin (1.45) 14abcde 7c 98.3ab 686b 2333a 1.16a
Ethalfluralin (1.12) +Imazethepyr (0.35) 10.4e 30ab 99.8a 1004ab 1706ab 0.79abc
Ethalfluralin (0.84) +Bentazon (0.84) 17.3ab 17bc 98.7ab 567b 2044a 1.05abc
Imazethepyr (0.53) 16abcd 49a 86.4b 9b 1252b 0.72abc
Imazethepyr (0.35) 12.8cde 33ab 98.4ab 1061ab 1994a 0.98abc
Pendimethalin (1.12) 15abcd 11bc 98.2ab 1196ab 1988a 1.06ab
Pendimethalin (0.84) +Imazethepyr (0.35) 13bcde 33ab 97.4ab 912b 1130b 0.65bc
Pendimethalin (0.84) +Bentazon (0.84) 11.8de 32ab 98.3ab 10b 1230b 0.61c
Trifluralin (0.84) 17.9a 7c 92.9b 753b 2029a 1.07ab
Handweeded check 17abc --- 100a 2801a 2203a 1.00+
zMeans within each column followed by the same letter are not significantly different by LSD at 0.05.
yAll herbicides were preplant incorporated except for Bentazon which was applied post emergent.
xQualitative analysis of:
Phytotoxicity: (0% = no crop injury; 100% = complete crop injury).
Herbicide efficacy: (0% = no weed control; 100% = complete weed control).

Last update May 1, 1997 aw