Weed management in common buckwheat (Fagopyrum esculentum) David A. Wall1,2 and Majorie A. H. Smith3 1Agriculture Can. J. Plant Sci. Downloaded from cdnsciencepub.com by 181.43.242.46 on 09/10/21 For personal use only. and Agri-Food Canada, Brandon Research Centre, P.O. Box 1000A, R.R. #3, Brandon, Manitoba, Canada R7A 5Y3; 3Agriculture and Agri-Food Canada, Cereal Research Centre, 195 Dafoe Rd., Winnipeg, Manitoba, Canada R3T 2M9. Received 30 September 1998, accepted 3 February 1999. Wall, D. A. and Smith, M. A. H. 1999. Weed management in common buckwheat (Fagopyrum esculentum). Can. J. Plant Sci. 79: 455–461. There are currently no broadleaf herbicides registered in Canada for use in common buckwheat (Fagopyrum esculentum Moench.). Greenhouse and field studies were conducted from 1994 to 1997 at Morden and Brandon, Manitoba, to investigate potential herbicide treatments for buckwheat when cultural methods fail to control broadleaf weeds. Dose-response studies conducted in the greenhouse showed that buckwheat tolerance to desmedipham was three to five times greater than that of volunteer canola (Brassica napus L. and B. rapa L.), lamb’s-quarters (Chenopodium album L.), redroot pigweed (Amaranthus retroflexus L.) and wild mustard (Sinapis arvensis L.). A separate experiment showed that buckwheat cultivars did not differ in tolerance to desmedipham. Another greenhouse study demonstrated that low dosages of bentazon caused severe injury in buckwheat. In field studies, 500 to 700 g a.i. ha–1 of desmedipham effectively suppressed the annual broadleaf weed biomass in buckwheat without reducing seed yields. At a low dosage of 300 g a.i. ha–1 of desmedipham there were no yield losses even though no weed biomass reduction was evident. Imazamethabenz at 150 to 250 g a.i. ha–1 and metribuzin at 140 and 210 g a.i. ha–1 severely injured buckwheat and usually reduced yields. 2,4-D at 140 g a.i. ha–1 suppressed weed biomass, but caused severe injury and often reduced buckwheat yield. At 70 g a.i. ha–1, 2,4-D suppressed weed growth as effectively as the higher dosage with less visual injury, but sometimes reduced buckwheat yields. Thus, desmedipham is the most promising postemergence broadleaf herbicide for use in common buckwheat. Key words: Common buckwheat, crop tolerance, dose response, weed management, yield Wall, D. A. et Smith, M. A. H. 1999. Maîtrise des mauvaises herbes dans les cultures de sarrasin (Fagopyrum esculentum). Can. J. Plant Sci. 79: 455–461. Il n’existe actuellement aucun herbicide antidicotylédones homologué pour le désherbage du sarrasin (Fagopyrum esculentum Moench.). Des expériences en serre et au champ ont été réalisées à Morden (Manitoba) de 1994 à 1997 pour examiner divers traitements herbicides pour cette culture advenant l’échec des méthodes culturales contre les mauvaises herbes dicotylédones. Des études dose-réponse menées en serre ont montré que la tolérance du sarrasin au desmédiphame était 3 à 5 fois plus forte que celle des repousses de canola (Brassica napus L. et B. rapa L.), du chénopode blanc (Chenopodium album L.), de l’amarante à racine rouge (Amaranthus retroflexus L.) et de la moutarde des champs (Sinapis arvensis L.). Une expérience distincte a permis de constater que les cultivars de sarrazin ne différaient pas quant au degré de tolérance au desmédiphame. Dans une autre expérience en serre, on a constaté que même à faible dose, le bentazone endommageait gravement le sarrasin. Dans les expériences au champ, le desmédiphame aux doses de 500 à 700 g m.a. ha–1 était efficace contre les dicotylédones annuelles dans le sarrasin, sans réduire le rendement grainier de la culture. À la faible dose de 300 g m.a. ha–1, l’herbicide ne causait pas de perte de rendement, même si l’on ne constatait aucune réduction de la biomasse adventice. L’imazaméthabenz à raison de 150 à 250 g m.a. ha–1 et le métribuzin, de 140 et 210 g m.a. ha–1 provoquaient de graves endommagements au sarrasin et, habituellement, réduisaient son rendement. Le 2,4-D à la dose de 140 g m.a. ha–1 diminuait la biomasse adventice, mais au prix de graves dommages et souvent de pertes de rendement pour le sarrasin. À la dose de 70 g m.a. ha–1, le 2,4-D inhibait la croissance des mauvaises herbes aussi efficacement qu’à la dose supérieure sans laisser d’endommagement visible à la culture, bien que parfois le rendement était touché. Il semble donc que le desmédiphame soit l’herbicide antidicotylédone de post-levée le plus intéressant pour le désherbage des cultures de sarrasin. Mots clés: Sarrasin, tolérance de la culture, dose-réponse, maîtrise des mauvaises herbes, rendement In Canada, common buckwheat is grown primarily for export to Japan to be used in noodle (soba) production. In 1995, 15 500 ha of buckwheat were grown in Canada, mostly in the province of Manitoba. Weed control has always been a major problem in buckwheat production, and there are currently no herbicides registered to control broadleaf weeds in this crop. Only diclofop methyl, used to control some annual grasses, is currently registered in Canada for use in this crop (Manitoba Agriculture 1997). In Manitoba, buckwheat producers frequently report that lamb’s-quarters, redroot pigweed, wild mustard and volunteer canola are the most troublesome weeds. 2Present Abbreviations: GR50, herbicide dosage required to reduce plant fresh and dry weight by 50%; WAT, wk after treatment address: Agriculture and Agri-Food Canada, Research Centre, 107 Science Place, Saskatoon, Saskatchewan, Canada S7N 0X2. 455 Can. J. Plant Sci. Downloaded from cdnsciencepub.com by 181.43.242.46 on 09/10/21 For personal use only. 456 CANADIAN JOURNAL OF PLANT SCIENCE Buckwheat is grown in solid stands, which precludes inter-row cultivation as a cultural method of weed control. Therefore, most producers rely on cultivation prior to seeding to destroy emerged weeds. Good seedbed preparation to ensure rapid emergence and stand establishment assists in controlling weeds since good stands of buckwheat compete strongly and smother weeds (Campbell and Gubbels 1986). When cultural weed control fails, producers often resort to the use of a non-registered product to control broadleaf weeds and rescue the crop. Several herbicides registered for use in other crops that may control broadleaf weeds in buckwheat have been identified. Desmedipham, a herbicide used in sugarbeet, has shown the greatest potential to selectively control some annual broadleaf weeds in buckwheat when applied at 500 to 1000 g a.i. ha–1 (Friesen and Campbell 1986; Friesen 1988). However, at the 1998 suggested retail price of $358 10 L–1 (150 g a.i. L–1), desmedipham is a relatively expensive herbicide when applied at these rates. In sugarbeets, desmedipham is frequently applied only over the row in combination with inter-row cultivation to reduce treatment costs. When used as an overall treatment in buckwheat, desmedipham may not be cost effective at dosages currently recommended for sugarbeets. Both 2,4-D and metribuzin are broad-spectrum broadleaf herbicides registered for use in a variety of crops. There are many anecdotal reports of farmers using low dosages of 2,4-D (70 to 140 g a.i. ha–1) as a low-cost option to control or suppress annual broadleaf weeds in buckwheat. Wall (1996) showed that 2,4-D at dosages of 76 to 151 g a.i. ha–1 always resulted in severe foliar injury, but seed yields were not always reduced. Similar results were obtained by Friesen and Campbell (1986) when buckwheat was treated postemergence with 130 g a.i. ha–1 of metribuzin (approximately 62% of the recommended rate for lentil crops in Manitoba). Imazamethabenz is registered for wild mustard control in sunflowers and some cereals. In drift simulation studies, Wall (1995a) reported that buckwheat tolerated imazamethabenz dosages up to 120 g a.i. ha–1 without yields being affected. The objective of this research was to investigate the potential of low dosages of desmedipham to selectively control annual broadleaf weeds in buckwheat. A further objective was to investigate the use of low dosages of 2,4-D, bentazon, imazamethabenz, and metribuzin as potential selective herbicides in buckwheat. MATERIALS AND METHODS Greenhouse Studies DESMEDIPHAM DOSE-RESPONSE. A dose-response experiment was conducted in the greenhouse to investigate the relative tolerance of common buckwheat and four annual broadleaf weed species to postemergence applications of desmedipham. Mancan buckwheat, canola (B. napus ‘AC Excel’), wild mustard, lamb’s-quarters and redroot pigweed were seeded separately in 10-cm-diameter pots filled with a mixture of soil:peat:sand:vermiculite (3:2:1:1). One week after emergence, plants were thinned to three per pot. Desmedipham was applied at 0, 125, 250, 375, 500, 625, 750, 1000, and 1250 g a.i. ha–1 when buckwheat and weeds were in the two- to three-leaf stage. Herbicide treatments were applied in a spray chamber (Model RC-5000-500EP Agrochemical Research Sprayer, Mandel Scientific Co. Ltd., Hanlon Business Park, 2 Admiral Place, R.R. #6, Guelph, Ontario, Canada M9W 6N9) equipped with a flatfan (Teejet 8001EVS Tip, Spraying Sytems Co., Wheaton, IL 60187, USA) nozzle delivering 200 L ha–1 of total solution at 207 kPa. The experimental design was a split-plot with six replicates. Main and subplot effects were species and herbicide dosage, respectively. Plants were harvested 3 WAT and total aboveground fresh and dry weights per pot were determined. Plants were dried for 24–48 h at 60°C. Data were subjected to analysis of variance (ANOVA) and nonlinear regression. Where a significant dose response was indicated by ANOVA (at P < 0.05), nonlinear regression was used to describe the relationship between herbicide dosage and plant response. Data were fitted to the following nonlinear model: Y = a + e(b + cX) where a, b, and c are estimated nonlinear regression parameters and X is desmedipham dosage (g a.i. ha–1). Nonlinear regression parameters were fitted to the data by a modified Marquardt algorithm for least-square estimation of nonlinear parameters, using PRISMTM Scientific Graphics Package V 1.0 (GraphPad Software, San Diego, CA). The herbicide dosage required to reduce plant fresh and dry weight by 50% (GR50) was calculated for each species. The experiment was repeated once. In a separate dose-response experiment, four cultivars of common buckwheat were compared for their tolerance to desmedipham. Cultivars Mancan, Manor, Manitoba and BS87368 were seeded in 10-cm pots and treated with desmedipham using the procedures and dosages described previously. The experimental design was a split-plot with six replicates. Main and subplot effects were cultivars and herbicide dosages, respectively. Plants were harvested 3 WAT and total plant fresh and dry weight per pot were determined. Data were analyzed by ANOVA and nonlinear regression as previously described. The experiment was repeated twice. BENTAZON DOSE-RESPONSE. A dose-response experiment was conducted in the greenhouse to investigate the tolerance of common buckwheat to bentazon. Mancan buckwheat was seeded in 10-cm-diameter pots as previously described and thinned to three plants 1 wk after emergence. Bentazon was applied at 0, 60, 120, 240, 360, 480, 600, 840, 1080 and 1320 g a.i. ha–1 when buckwheat was in the two- to threeleaf stage. The recommended dosage for use against lamb’squarters and redroot pigweed is 1080 g a.i. ha–1 (Manitoba Agriculture 1997). Herbicide treatments were applied as previously described. The experimental design was a randomized complete block with six replicates. Plants were harvested 3 WAT and the total plant fresh and dry weights per pot were determined. Data were analyzed by ANOVA WALL AND SMITH — WEED MANAGEMENT IN COMMON BUCKWHEAT 457 Table 1. Significant dates relating to buckwheat crop growth and weed control Morden Event Previous crop Buckwheat seeding Crop emergence Herbicide application Crop leaf stage Weed leaf stagez Weed sampling (WAT) Buckwheat harvest Can. J. Plant Sci. Downloaded from cdnsciencepub.com by 181.43.242.46 on 09/10/21 For personal use only. zWild yCrop Brandon 1994 1995 1996 1997 1995 1996 1997 Flax 23 June 29 June 8 July 2 2–3 5 30 Sept. Wheat 31 May 7 June 26 June 4 6 6 22 Sept. Wheat 4 June 11 June 21 June 2 2–4 7 19 Sept. Fallow 30 May 4 June 19 June 2 4–6 6 17 Sept. Fallow 2 June 9 June 30 June 6 Cotyledon 7 –y Fallow 10 June 17 June 28 June 3 2–4 6 26 Sept. Wheat 24 June 2 July 14 July 3–4 2–4 4 24 Sept. mustard, redroot pigweed and lamb’s-quarters were all at similar stages at herbicide application. not harvested due to severe wind damage. and nonlinear regression as previously described. The experiment was repeated twice. Field Studies Field studies were conducted at the Agriculture and AgriFood Canada Research Centres at Morden (1994 to 1997) and Brandon (1995 to 1997) in Manitoba, Canada. The soil at the Morden site was a Neuhorst clay loam while at Brandon it was an Assinaboine complex clay loam. The crop grown in the previous year varied each year (Table 1). The experiment site was fertilized as required and cultivated using a power harrow, then a diamond harrow. Wild mustard, lamb’s-quarters and redroot pigweed were broadcast prior to cultivation to supplement the natural weed population and ensure adequate weed pressure. Mancan buckwheat was seeded 3 cm deep at 55 kg ha–1 in late May or June (Table 1) using a double disc press drill. At Morden, row spacing was 18 cm in 2 by 8 m plots, and at Brandon, row spacing was 25 cm in 2 by 5 m plots. Experiments consisted of 10 herbicide treatments plus weed-free (hand-weeded) and weedy checks, arranged in a randomized complete block design with four replicates. Herbicides were applied about 10–15 d after emergence of buckwheat and weeds (Table 1), with a plot sprayer equipped with flat-fan nozzles [Teejet 8002SS (Morden) or XR8001 (Brandon) Tip, Spraying Systems Co., Wheaton, IL 60187, USA], delivering 139 L ha–1 of total solution at 262 kPa at Morden and 100 L ha–1 at 207 kPa at Brandon. Treatments consisted of 2,4-D amine at 70 and 140 g a.i. ha–1, desmedipham at 300, 500 and 700 g a.i. ha–1, imazamethabenz at 150, 200 and 250 g a.i. ha–1 and metribuzin at 140 and 210 g a.i. ha–1. In 1994, desmedipham was applied only at 500 and 700 g a.i. ha–1. At application, buckwheat growth stage varied from two-leaf (10 cm) to six-leaf (35 cm), while wild mustard, redroot pigweed and lamb’s-quarters were usually at the two-leaf (2 cm) to four-leaf (3–5 cm) stage (Table 1). Sethoxydim, when required, was applied at 250 g a.i. ha–1 to control annual grasses. Crop injury was visually estimated at 2 and 4 WAT on a 0 to 100% scale, where 0 and 100% injury corresponded to no visible effect and complete mortality. Injury levels higher than 15% were considered to be severe, since injury above 15% almost always results in yield losses (Anonymous 1992). Crop injury ratings were not recorded at Brandon in 1995. Weed control was determined between 4 and 7 WAT (Table 1), depending on year and location, by counting the number of surviving wild mustard, lamb’squarters and redroot pigweed plants in four randomly selected 0.25 m–2 quadrats in each plot. Total broadleaf weed dry weight (g m–2) was also determined. Buckwheat was swathed when 75% of the seeds had turned brown. After drying, the buckwheat was mechanically harvested, seed was dried to a uniform moisture content, cleaned and the seed yield (g m–2) determined. Swathing and harvesting were not possible at Brandon in 1995 because the buckwheat was severely wind damaged and crop yields could not be determined. Data from each trial were analyzed using ANOVA. Weed counts and dry weights were transformed prior to analysis using {square root (X + 0.5)}, and means were back transformed for presentation. Where treatment effects were significant according to ANOVA, means were separated at P = 0.05 by Duncan’s multiple range test. RESULTS AND DISCUSSION Greenhouse Studies DESMEDIPHAM DOSE-RESPONSE. The experiments did not differ significantly according to ANOVA and the data were therefore combined. Responses were similar for fresh and dry weights, so only dry weight data are presented. Plant species differed significantly in their responses, but the interaction between species and herbicide dosage was also significant; therefore, the response of each species was analyzed separately. Buckwheat dry weights were reduced as desmedipham dosage increased, but the effect was less than that observed in the other test species, as indicated by the GR50s (Fig. 1). Wild mustard, lamb’s-quarters, and redroot pigweed dry weights decreased rapidly with dosages up to 375 g a.i. ha–1. There was little change in plant weight with further increases in desmedipham dosage. Canola response to increasing dosage was intermediate to buckwheat and the three weed species. The GR50 for buckwheat dry weight response was 1005 g a.i. ha–1, approximately 3.1 times higher than the GR50 for canola, and 6.0 times higher than the average for the three weed species (Fig. 1). The results of this experiment suggests that there are sufficient differences in tolerance to desmedipham between buckwheat and some Can. J. Plant Sci. Downloaded from cdnsciencepub.com by 181.43.242.46 on 09/10/21 For personal use only. 458 CANADIAN JOURNAL OF PLANT SCIENCE Fig. 1. Effect of desmedipham dosage on dry weight of common buckwheat, GR50 = 1005 g a.i. ha–1; canola, GR50 = 329 g a.i. ha–1; wild mustard, GR50 = 171 g a.i. ha–1; lamb’s-quarters, GR50 = 175 g a.i. ha–1; and redroot pigweed, GR50 = 155 g a.i. ha–1. Combined data from two greenhouse experiments, 1995. common weed species for this herbicide to be used for selective weed control in this crop. When tolerance to desmedipham was compared among the four buckwheat cultivars, fresh weight differed significantly among cultivars in two of the three trials (P < 0.006). Herbicide dosage was significant in all trials (P < 0.0001). However, there were no significant interactions between cultivars and herbicide dosage (P > 0.1). Similar results were obtained when dry weights were analyzed (data not presented). Therefore, there is no evidence that cultivars are responding differently to desmedipham dosage. BENTAZON DOSE-RESPONSE. Buckwheat was very sensitive to bentazon under greenhouse conditions and plant fresh weight decreased rapidly with increasing dosage (data not presented). For buckwheat fresh weight, GR50 values for the three experiments were 624, 316 and 383 g a.i. ha–1. Under field conditions, Wall (1995b) reported that 600 g a.i. ha–1 of bentazon provided poor control of weeds in dry beans. This dosage likely represents the minimum dosage that would be required to control weeds in buckwheat under field conditions. At this dosage, buckwheat fresh weight was reduced by 48, 68, and 64% in the three experiments. Based on the severe injury to buckwheat that occurred under greenhouse conditions at the minimum dosage required for weed control in the field, bentazon was not considered to be a suitable candidate herbicide for field evaluation. Field Studies In all years of the study, the average temperature for the May to September period (growing season) was within 0.7°C of the 30-yr (1961–1990) average of 16.5°C at Morden and 15.4°C at Brandon. At Morden in 1994, May and September mean temperatures were above normal, but July and August were cooler than normal. In 1995, 1996 and 1997, temperature patterns over the growing season were similar at Morden and Brandon. In all 3 yr, May was cooler and June was warmer than normal, whereas temperatures during the remainder of the growing season were near normal. Total rainfall at Morden over the growing season was within 10% of the 30-yr average of 335 mm in all years except in 1995 (39% higher). In 1995, June and July were drier than normal but August received more than three times the monthly average. At Brandon, total rainfall over the growing season was 65, 84 and 62% of the 30-yr average of 312 mm in 1995, 1996 and 1997, respectively. In 1995 and 1996, July and August were both drier than average. At both locations, the 1997 growing season was relatively dry except for July, which received about twice the average monthly rainfall. Injury levels at 2 WAT were similar to those at 4 WAT in all trials, except at Brandon in 1996, when all ratings at 2 WAT were twice those at 4 WAT. Therefore, only the ratings at 4 WAT are presented. For most treatments, injury tended to be higher in 1994 and 1995 than 1996 and 1997 at Morden (Table 2). Injury by desmedipham at 4 WAT was generally acceptable except in 1994 and 1995 (highest dosage only) at Morden, whereas injury by the other herbicides was generally severe in all trials (Table 2). Desmedipham caused some leaf burn and necrosis around the leaf margins and stunting, similar to that reported by Friesen (1988). Although there were no significant differences in injury levels among dosages, there was a trend for the severity of injury to increase with increasing dosage. Metribuzin injury was more severe than injury from any other treatment at Morden except in 1997 (Table 2). Symptoms consisted of foliar chlorosis and necrosis and severe stand reduction. Stand reductions were more severe in 1994 than in other years. Injury by 2,4-D at 140 g a.i. ha–1 was severe, and greater than at 70 g a.i. ha–1 in all trials. 2,4D produced typical phenoxy injury symptoms including severe stem epinasty, leaf cupping and adventitious root formation. Imazamethabenz injury was similar at the three dosages (Table 2), and caused slight leaf chlorosis and stunting. Injury was severe in all trials except at Morden in 1996. Weather conditions around the time of spraying may have contributed to the lower levels of injury seen in 1996. Imazamethabenz performs well under warm, dry conditions (Manitoba Agriculture 1997). At Morden in 1996 cool, wet conditions prevailed for several days after spraying. In 1997, 22 mm of rain was received several hours after spraying, and may have resulted in lower injury levels by both imazamethabenz and metribuzin that year. In most trials, populations of wild mustard and lamb’squarters were very low, with mean densities ranging from 0 to 11 and from 0 to 5 plants m–2, respectively, and control by herbicide treatments could not be effectively evaluated (data WALL AND SMITH — WEED MANAGEMENT IN COMMON BUCKWHEAT 459 Table 2. Visual injury (%) to common buckwheat by postemergence herbicides at 4 WAT Can. J. Plant Sci. Downloaded from cdnsciencepub.com by 181.43.242.46 on 09/10/21 For personal use only. Treatment Weedy Weed-free 2,4-D 2,4-D Desmedipham Desmedipham Desmedipham Imazamethabenz Imazamethabenz Imazamethabenz Metribuzin Metribuzin CV (%) Morden Brandon Dosage (g a.i. ha–1) 1994 1995 70 140 300 500 700 150 200 250 140 210 0d 0d 16c 36b – 16c 25bc 30b 26bc 29bc 74a 85a 0e 0e 28c 44b 8de 10de 19cd 25c 25c 28c 46b 69a 0d 0d 12bc 19b 5cd 6cd 7cd 10bcd 4cd 12bc 56a 56a 0f 0f 7de 15bc 4ef 6def 9cde 12bcd 18ab 23ab 15bc 23ab 0d 0d 15c 31a 1d 4d 4d 25b 35a 35a 15c 16c 0c 1c 1c 31a 0c 0c 0c 25b 26ab 31a 21b 23b 27 38 37 38 26 29 1996 1997 1996 1997 a–fMeans within columns followed by a common letter are not significantly different according to Duncan’s multiple range test at P = 0.05. not presented). Wild mustard was controlled by herbicides only at Morden in 1994, when all treatments reduced density of this weed (0–2 plants m–2 compared to 24 plants m–2 for the weedy check). Wild mustard densities were moderate to high in 1997 at Morden (26–55 plants m–2), and no treatment was effective. Densities of lamb’s-quarters were high only at Brandon in 1995, and were controlled by metribuzin and 140 g a.i. ha–1 of 2,4-D. Redroot pigweed populations were moderate to heavy in all trials (Table 3), except at Morden in 1995 (0–7 plants m–2; data not presented), but no herbicide consistently controlled its density. Redroot pigweed was controlled by desmedipham and metribuzin in 1994 at Morden and in 1997 at Brandon, and by 2,4-D in 1996 (140 g a.i. ha–1 only) and 1997 at Brandon (Table 3). In this study, the higher densities of redroot pigweed compared to wild mustard and lamb’s-quarters may have been a result of the relative ability of these three weeds to compete for light as the buckwheat stand developed. Wild mustard grows mainly in habitats with a high light intensity (Mulligan and Bailey 1975), and shading inhibits the germination of lamb’s-quarters (Basset and Crompton 1978). Redroot pigweed, on the other hand, germinates at high temperatures throughout the season regardless of light level (Weaver and McWilliams 1980) and would not have been inhibited by the buckwheat canopy to the same extent. Total broadleaf weed biomass was significantly reduced by 2,4-D at 140 g a.i. ha–1 in three of the seven trials and by desmedipham at 500 and 700 g a.i. ha–1 in four trials. Metribuzin at 210 g a.i. ha–1 reduced weed dry weight in 1994 and 1997 at Morden (Table 3), but was otherwise not effective in suppressing weed biomass. Imazamethabenz did not reduce weed dry weight except at the highest dosage in one trial. In fact, dry weights were significantly higher than in the weedy checks at Brandon in 1996 and 1997 (Table 3). Severe buckwheat injury and stand reductions likely allowed these weed populations to increase, indicating the extent to which good stands of untreated buckwheat can smother weed growth (Campbell and Gubbels 1986). This study indicates that, even though weed densities are not controlled, 140 g a.i. ha–1 of 2,4-D and 500 and 700 g a.i. ha–1 of desmedipham can effectively suppress total broadleaf weed biomass. Weed densities of the three weed species fluctuated widely among years and between locations, while total weed dry weights were more consistent. However, specific comparisons of weed densities and biomass cannot be made because dry weights included the total of all broadleaf weeds, but densities were recorded only for the three most commonly reported weed species in buckwheat. Treatment differences in buckwheat seed yields were significant in all trials. However, due to the amount of variation among plots, yields in many herbicide treatments did not differ from either the weed-free or the weedy check (Table 4). There was a general consistency among trials in the effects of desmedipham and imazamethabenz on buckwheat seed yield, while the effect of 2,4-D and metribuzin on yield varied considerably among trials. Crop yield in plots treated with desmedipham at all three dosages were equal to or better than that of any other herbicide treatment (Table 4). Desmedipham treated plots outyielded imazamethabenz treated plots in four of six trials, and plots treated with metribuzin and the higher dosage of 2,4-D in three of six trials. Buckwheat yield losses due to desmedipham were significant only in 1994 at Morden. While desmedipham provided good weed control in 1994, lower yields may reflect sustained effects of early season injury. The 1994 test was seeded late in the season and this may have reduced yield potentials due to the limited time available for crop recovery, particularly with the cool temperatures in July and August that year. Buckwheat treated with imazamethabenz generally sustained high yield losses at all dosages (Table 4). At Morden in 1996 and 1997, when imazamethabenz injury was relatively low, buckwheat yields did not differ from the weedfree check for two of the three dosages. Buckwheat yields were significantly lower than the weedy check in the four trials where injury was severe. These high yield losses are likely due to reduced competitive ability, which resulted in the higher weed biomass observed in these trials. 31 51 38 30 33 52 39 26 25 41 23 19 38 transformed prior to analysis using a square root transformation {sqrt(X + 0.5)}. a–fMeans within columns followed by a common letter are not significantly different according to Duncan’s multiple range test at P = 0.05. 70 140 300 500 700 150 200 250 140 210 CV (%) zData 16ab 7b 1c 0b 21ab 7b 9bc 2b 12abc 7b 9bc 3b 11bc 3b 33ab 68a 30ab 102a 43a 51a 22ab 8b 25ab 13b 52ab 129a 0e 0e 21bcd 50bcd 19cd 30cd 13cd 72abc 9de 50bcd 12cd 20d 37abc 73abc 35abc 84ab 37abc 54bcd 72a 70abc 65a 59bcd 6ab 0c 2abc 3abc 2abc 1bc 1bc 6ab 10a 7ab 10a 6ab 55ab 0e 3de 7de – 1de 1de 65a 52ab 32abc 28bc 11cd 91bc 0f 41d 26de 57cd 11e 11e 108ab 168a 121ab 44d 38d 17bcd 0f 19bcd 6e 22bcd 10de 11cde 24abc 30ab 40a 15bcde 21bcd 90ab 0c 91ab 56b 52b 43b 43b 115ab 173a 151ab 65ab 60ab 23a 0b 34a 31a 27a 36a 22a 27a 28a 33a 33a 18a 25bcd 0e 24bcd 24bcd 14d 18cd 26bcd 25bcd 30bc 31bc 54a 34b 19ab 0d 16bc 15bc – 2d 1d 37a 20ab 21ab 6cd 2d Weedy Weed-free 2,4-D 2,4-D Desmedipham Desmedipham Desmedipham Imazamethabenz Imazamethabenz Imazamethabenz Metribuzin Metribuzin Brandon 1996 1996 1995 1996 1996 Treatment Dosage (g a.i. ha–1) 1994 Morden 1997 1995 Brandon 1997 1994 Morden 1997 1995 Total weed dry weight (g m–2)z Redroot pigweed density (plants m–2)z Table 3. Effect of herbicide treatment on redroot pigweed density and total broadleaf weed dry weight Can. J. Plant Sci. Downloaded from cdnsciencepub.com by 181.43.242.46 on 09/10/21 For personal use only. 22cd 0f 14de 10de 12de 3ef 3ef 35bc 84a 52b 16cd 19cd CANADIAN JOURNAL OF PLANT SCIENCE 1997 460 Plots treated with 2,4-D at 70 g a.i. ha–1 yielded as well as the weed-free check in four of six trials (Table 4). In comparison, plots treated with 2,4-D at 140 g a.i. ha–1 tended not to yield as well, and yielded significantly less in two trials. In some trials, yields of buckwheat treated with 2,4-D at 70 or 140 g a.i. ha–1 were less than the weedy check. Reduced seed yields were commonly associated with severe injury from 2,4-D applications, but severe injury was not always a predictor of yield reductions. In 1994, 2,4-D caused severe injury, but the lower dosage did not reduce seed yields. Wall (1996) noted that environmental conditions at the time of and following spraying is an important factor in the ability of buckwheat to recover from severe injury. Metribuzin yielded equal to desmedipham in three trials, but severely reduced yields in 1994 and in 1996 (210 g a.i. ha–1 only) at Morden (Table 4). The yield losses can be attributed to the severe injury and stand reduction in 1994 and to both severe injury and lack of weed control in 1996. However, metribuzin-treated buckwheat sustained severe injury in 1995 without significant yield loss, while the lower dosage reduced yields in 1996 at Brandon even though injury was relatively low. Weather factors at and just after application are known to influence the effectiveness of metribuzin (Manitoba Agriculture 1997) and may help to explain some of these differences. Freeman (1982) found that metribuzin injury in potatoes was related to temperature and the amount of sunshine just before, during, and just after the time of spraying, and that weather during the rest of the season affected recovery from injury. When left untreated, moderate to heavy broadleaf weed populations can affect buckwheat yields, justifying the need for control measures. Comparison of the untreated check plots among trials showed that yield loss was associated with broadleaf weed biomass (Tables 3 and 4). Yield losses in the weedy checks, relative to the weed-free checks, ranged from 20 to 40% at Morden in 1994, 1996 and 1997, when weed dry weights were more than 50 g m–2. When weed dry weights were less than 22 g m–2, weeds had no competitive effect on crop yields. Neither metribuzin nor imazamethabenz can be recommended as candidates for selective broadleaf weed control in buckwheat because they cause severe injury to the crop and can seriously affect yields. At 140 g a.i. ha–1, 2,4-D usually provided good suppression of weed growth, but severely injured buckwheat and often reduced yields. Although 2,4-D at 70 g a.i. ha–1 suppressed weed growth almost as effectively as the higher dosage with less buckwheat injury, yield losses were still significant in two of the six trials. At 300 and 500 g a.i. ha–1, desmedipham caused an acceptable level of injury to buckwheat, and yields were similar at the two dosages, even though the lowest dosage did not suppress weed biomass as well. Friesen (1988) found that 500 g a.i. ha–1 of desmedipham, when applied at the four-leaf stage of buckwheat, selectively controlled wild mustard without crop yield losses. This study showed that desmedipham at 500 g a.i. ha–1 effectively suppressed moderate to heavy broadleaf weed infestations, and suggests that this herbicide may reduce redroot pigweed densities. Desmedipham remains WALL AND SMITH — WEED MANAGEMENT IN COMMON BUCKWHEAT 461 Table 4. Effect of herbicide treatment on buckwheat seed yield (g m–2) Can. J. Plant Sci. Downloaded from cdnsciencepub.com by 181.43.242.46 on 09/10/21 For personal use only. Treatment Weedy Weed-free 2,4-D 2,4-D Desmedipham Desmedipham Desmedipham Imazamethabenz Imazamethabenz Imazamethabenz Metribuzin Metribuzin CV (%) Brandonz Morden Dosage (g a.i. ha–1) 1994 1995 1996 1997 1996 1997 70 140 300 500 700 150 200 250 140 210 184b 230a 203ab 171b – 185b 186b 104c 118c 105c 99c 58d 164ab 209a 162ab 112c 203a 189ab 189ab 155b 111c 108c 183ab 168ab 108cde 180a 126bcd 131abcd 133abcd 136abcd 161ab 128bcd 157abc 146abcd 100de 63e 83c 141ab 143a 129abc 107abc 106abc 149a 94bc 96bc 85c 130abc 127abc 116a 116a 77cd 54e 113ab 97ab 110ab 73d 61de 53e 93bc 94abc 156bcd 162bc 152cd 135d 188a 195a 180ab 110e 104e 101e 147cd 137cd 15 8 23 25 7 12 zCrop not harvested in 1995 due to severe wind damaged. a–dMeans within columns followed by a common letter are not significantly different according to Duncan’s multiple range test at P = 0.05. the most promising postemergence treatment evaluated both in terms of effectiveness and crop safety. ACKNOWLEDGMENT D.A. Wall wishes to acknowledge the technical support of Abe Fehr and Desiree Czerkawski. Anonymous. 1992. Herbicide and plant regulator field trials. Pages 51–82 in Manual for field trials in plant protection. 3rd ed. Plant Protection Division, Ciba-Geigy Ltd., Basle, Switzerland. Bassett, I. J. and Crompton, C. W. 1978. The biology of Canadian weeds. 32. Chenopodium album. L. Can. J. Plant Sci. 58: 1061–1072. Campbell, C. G. and Gubbels, G. H. 1986. Growing buckwheat. Agric. Can. Techn. Bull. 1986-7E, Ottawa, ON. 8 pp. Freeman, J. A. 1982. The influence of weather on the response of potato cultivars to metribuzin. J. Am. Soc. Hortic. Sci. 107: 189–194. Friesen, G. H. 1988. Wild mustard, Sinapis arvensis, control in common buckwheat, Fagopyrum esculentum, with desmedipham and fluorochloridone. Weed Technol. 2: 175–178. Friesen, G. H. and Campbell, C. G. 1986. Common buckwheat (Fagopyrum esculentum) tolerance to herbicides. Weed Sci. 34: 435–439. Manitoba Agriculture. 1997. Guide to crop protection 1997. Manitoba Agriculture, Carman, MB. 254 pp. Mulligan, G. A. and Bailey, L. G. 1975. The biology of Canadian weeds. 8. Sinapis arvensis L. Can. J. Plant Sci. 55: 171–183. Wall, D. A. 1995a. Response of four annual broadleaf crops to simulated imazamethabenz spray drift. Can. J. Plant Sci. 75: 751–757. Wall, D.A. 1995b. Bentazon tank-mixtures for improved redroot pigweed (Amaranthus retroflexus) and common lambsquarters (Chenopodium album) control in navy beans (Phaseolus vulgaris). Weed Technol. 9: 610–616. Wall, D. A. 1996. Effect of sublethal dosages of 2,4-D on annual broadleaf crops. Can. J. Plant Sci. 76: 179–185. Weaver, S. E. and McWilliams, E. L. 1980. The biology of Canadian weeds. 44. Amaranthus retroflexus L., A. powellii S. Wats. and A. hybridus L. Can. J. Plant Sci. 60: 1215–1234.