Published online 3 May 2006
Published in Agron J 98:815-822 (2006)
DOI: 10.2134/agronj2005.0234
© 2006 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
Production Papers
Wheat Seeding Rate Influences Herbicide Performance in Wild Oat (Avena fatua L.)
John T. O'Donovana,*,
Robert E. Blackshawb,
K. Neil Harkerc and
George W. Claytonc
a Agric. and Agri-Food Canada, Beaverlodge, AB T0H 0C0 Canada
b Agric. and Agri-Food Canada, Lethbridge, AB T1J 4B1 Canada
c Agric. and Agri-Food Canada, Lacombe, AB T4L 1W1 Canada
* Corresponding author (O'DonovanJ{at}agr.gc.ca)
Received for publication August 15, 2005.
 |
ABSTRACT
|
|---|
Field experiments were conducted at three locations in Alberta for 3 yr to determine if spring wheat (Triticum aestivum L.) seeding rate (75 and 150 kg ha–1) influenced the effects of recommended and reduced herbicide rates on wild oat (Avena fatua L.) shoot biomass, wild oat seed in the soil seed bank, and wheat yield and net economic return. Wild oat biomass and seed in the soil seed bank decreased nonlinearly at both seeding rates as herbicide rates increased. The herbicides were more effective in reducing wild oat shoot biomass and seed in the soil seed bank when wheat was seeded at the higher rate. The lowest wheat yields and net economic returns occurred when no herbicides were applied and both variables increased nonlinearly with increasing herbicide rate. In most cases, wheat yield and net economic return were greater at the higher seeding rate. On average, wheat yield improved by 19% and net economic return by 16% when wheat was seeded at the higher rate. The results indicate that seeding wheat at relatively high rates can contribute positively to herbicide performance and result in better wild oat management and higher wheat yields and economic returns. In some cases, there was little difference between applying the herbicides at 75 or 100% of the recommended rate but reducing rates below 75% almost always resulted in higher wild oat shoot biomass and seed, and reduced yields and net economic returns, even at the higher wheat seeding rate.
|
INTRODUCTION
|
|---|
WILD OAT is the most serious annual weed of field crops in western Canada and can result in extensive crop yield and revenue losses. Wild oat can also be expensive to control with herbicides, and thus farmers sometimes reduce herbicide rates below those recommended on the label to reduce costs. There is evidence that herbicides for wild oat control can perform better in the presence of crop competition (Sharma and Vanden Born, 1983) suggesting that agronomic practices that enhance crop competitiveness may improve herbicide performance and result in better wild oat control. In studies with continuous barley (Hordeum vulgare L.; O'Donovan et al., 2001; Wille et al., 1998), a canola (Brassica napus L.)–barley rotation (O'Donovan et al., 2004), and a barley–field pea (Pisum sativum L.) rotation (Blackshaw et al., 2005), seeding crops at a relatively high rate enhanced the efficacy of herbicides applied at reduced rates. Similarly, in a diverse crop rotation, seeding the crops at a relatively high rate resulted in a 33% reduction in graminicide rate by limiting seed production of herbicide-resistant wild oat (Beckie and Kirkland, 2003). Similar research on wild oat management in wheat using a combination of seeding and herbicide rates is very limited.
Previous studies with wheat indicated that increased wild oat seed production and reduced crop yields and revenues can occur when herbicide rates were reduced below those recommended on the label (Holm et al., 2000; O'Donovan et al., 2003a, 2003b). In these studies, the wheat seeding rate was relatively low (67–80 kg ha–1). Wheat plant density as a function of seeding rate will vary among varieties due to different seed size. Alberta provincial guidelines recommend that spring wheat plant densities should be in the range of 100 to 200 plants m–2, and indicate that seeding rates should generally be lower when soil moisture is limited than when moisture is plentiful (Alberta Agriculture, Food and Rural Development, 2005a). In wheat fields surveyed in central Alberta, wheat plant densities were sometimes lower than those recommended, suggesting that growers were seeding at suboptimal rates (O'Donovan, 2000).
The objective of this study was to determine the effects of various rates of registered graminicides on wild oat shoot biomass, wild oat seed in the soil seed bank, and wheat yield and net economic return as influenced by low (75 kg ha–1) and high (150 kg ha–1) wheat seeding rates.
|
MATERIALS AND METHODS
|
|---|
Field Operations
Field experiments were conducted under zero tillage at Beaverlodge (119°26' W, 55°13' N), Lacombe (113°44' W, 52°28' N), and Lethbridge (112°47' W, 49°38' N), Alberta, Canada, in 2000, 2001, and 2002. Soil types were a clay loam (29% sand, 44% silt, and 27% clay) Dark Gray Solod (Molic Cryoboralf) with pH 6.1 and 6% organic matter at Beaverlodge, a sandy loam (67% sand, 19% silt, and 15% clay) Black Chernozem (Typic Haplustoll) with pH 7.1 and 5.6% organic matter at Lacombe, and a loam (36% sand, 30% silt and 34% clay) Dark Brown Chernozem (Typic Haploboroll) with pH 7.8 and 3.6% organic matter at Lethbridge. Wheat was grown on the same plots at each location each year to determine the cumulative effects of the management practices on wild oat seed in the soil seed bank after 3 yr. All experiments were conducted on unirrigated land.
The experiment was a randomized complete block with four replicates. Seed drills with knife openers were used at Beaverlodge and Lacombe, while a double-disk press drill was used at Lethbridge. Plot size was 3.7 by 20 m at Beaverlodge, 3.7 by 15.2 m at Lacombe, and 2.1 by 6.0 m at Lethbridge.
Wheat (cv. AC Crystal) was seeded at 75 and 150 kg ha–1 in 20-cm rows (2.5-cm depth) to obtain target plant densities of approximately 100 and 200 plants m–2, which represent low and high ends of the recommended wheat seeding rate range (Alberta Agriculture, Food and Rural Development, 2005a). To facilitate presentation and discussion of results, crop seeding rates will be referred to as low and high. Seeding dates in 2000, 2001, and 2002, respectively, were May 5, 8, and 13 at Beaverlodge, May 1, 2, and 10 at Lacombe, and May 13, 10, and 19 at Lethbridge. These are typical seeding dates for these locations. Fertilizer was applied each year at seeding time according to the soil test recommendations for wheat (Alberta Agriculture, Food and Rural Development, 2005b). Fertilizers were applied either as a side band (Beaverlodge and Lacombe) or a midrow band (Lethbridge).
In 2000, shortly after seeding the crop, wild oat was seeded to a depth of 2.5 cm at right angles to the wheat rows to obtain a target plant density of 
100 plants m–2. Plants resulting from these seeds and their offspring were monitored during the course of the experiment. The wild oat seed used at Beaverlodge and Lacombe originated from seed collected from a seed cleaning plant near Lacombe, while seed used at Lethbridge originated from a seed cleaning plant in that region. The populations used in the study were determined to be susceptible to clodinafop, imazamethabenz, and tralkoxydim in greenhouse studies.
Herbicides for wild oat control were applied at the three- to four-leaf stage of wheat and at the two- to four-leaf stage of wild oat. A different graminicide was used each year to conform to provincial herbicide-resistance management guidelines (Ali, 2005). Clodinafop–propargyl [2-propynyl(R)-2-(4-(5-chloro-3-fluoro-2-pyridinyloxy)-phenoxy)propionate/5-chloro-8-quinolinoxyacetic acid-1-methylhexylester] in 2000, imazamethabenz [(±)-2-[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2-yl]-4(and 5)-methylbenzoic acid] in 2001, and tralkoxydim (2-[1-(ethoxyimino)propyl]-3-hydroxy-5-(2,4,6-trimethylphenyl)-cyclohex-2-enone) in 2002 were applied at rates corresponding to 0, 0.25, 0.50, 0.75, and 1.00 times the recommended label rates, which were 55, 400, and 200 g a.i. ha–1, respectively. In 2000, herbicides for dicot weed control were applied from 4 to 5 d after clodinafop–propargyl application. These were commercial mixtures of MCPA [(4-chloro-2-methylphenoxy)acetic acid], mecoprop [(±)-2-(4-chloro-2-methylphenoxy)propanoic acid], and dicamba (3,6-dichloro-2-methoxybenzoic acid) applied at 350, 77, and 77 g a.i., respectively, at Beaverlodge; a commercial mixture of thifensulfuron-methyl (3-[[[[(4-methoxy-6-methyl-1,3,5-triazin-2-yl)amino]carbonyl]amino]sulfonyl]-2-thiophenecarboxylic acid) and tribenuron (2-[[[[(4-methoxy-6-methyl-1,3,5-triazin-2-yl)methylamino]carbonyl]amino]sulfonyl]benzoic acid) applied at 10 and 5 g a.i. ha–1, respectively, tank mixed with MCPA amine at 350 g a.i. ha–1 at Lacombe; and a commercial mixture of bromoxynil (3,5-dibromo-4-hydroxybenzonitrile) and MCPA ester, each applied at 280 g a.i. ha–1 at Lethbridge. In 2001, imazamethabenz treatments were tank mixed with thifensulfuron and tribenuron at 10 and 5 g a.i. ha–1, respectively. In 2002, tralkoxydim treatments were tank mixed with bromoxynil and MCPA ester (each at 280 g a.i. ha–1). When applicable, herbicides were applied with recommended and registered adjuvants (Ali, 2005). All graminicide treatments received the full recommended adjuvant rates. All herbicides were applied in a water volume of 110 L ha–1 with flat fan nozzles at a pressure of 275 kPa at Beaverlodge and Lacombe, and 207 kPa at Lethbridge.
In 2000, two 1-m2 areas were established in each plot. These areas were permanently marked, and all data were collected from them each year. Wild oat density in 2000 and wheat density each year were determined before herbicide application. Mature wild oat plants in each quadrat were cut at soil level before seed shatter. Wheat plants (four rows) were cut at soil level at maturity, and threshed in a stationary thresher. All samples were dried to a constant weight. Each fall the wild oat seed was returned to the soil surface in the marked quadrats.
Soil Seed Bank Assessment
The effect of the treatments on the amount of wild oat seed present in the soil seed bank was determined after crop harvest at the conclusion of the experiment in the fall of 2002. Soil was randomly sampled to a depth of 5 cm at eight locations in each plot using a 9-cm-diam. soil coring device. Samples from each plot were bulked and dried for 5 to 7 d at 30°C to prevent seed germination, and relatively large extraneous material (e.g., straw and rocks) removed with an 8-mm sieve. Fine soil was removed using a 1-mm sieve. The rest of the sample was gently washed with water until only wild oat seed remained, which was then dried at 30°C, counted, and expressed on a square-meter basis.
Data Analysis
Data were analyzed using regression analysis. The appropriate model was selected based on the nature of the response, and models that provided the best descriptions of the data are presented.
Relationships between herbicide rate and either wild oat shoot dry weight (biomass) or wild oat seed in the soil seed bank were described using a nonlinear dose response model:
 | [1] |
where y is estimated wild oat biomass (g m–2) or seed m–2 as a function of herbicide rate (x), k is wild oat shoot biomass or wild oat seed in the soil seed bank in the absence of herbicide application, and b and g are scale parameters describing the shape of the response. Data were fitted to Eq. [1] and [2] using the derivative-free PROC NONLIN (SAS Institute, 1999).
The relationship beween wheat yield and herbicide rate at each seeding rate was described using the nonlinear model
 | [2] |
where y is the estimated crop yield (g m–2) as a function of herbicide rate (x), a is the wheat yield at zero herbicide rate, and b and c describe the slope of the regression line.
Cumulative net economic return (across the 3 yr) was calculated from the model
 | [3] |
where E is the net economic return, Y is the cumulative wheat seed yield (kg ha–1), P is the market price of wheat, S is the cost of seed, and H and A are the cost of the graminicide at the rate applied and its application cost, respectively. Assumptions were P = $0.14 kg–1, S = $0.22 kg–1, H = $38.69, $32.18, and $40.22 ha–1 for the full recommended rates of clodinafop, imazamethabenz and tralkoxydim, respectively, and A = $12.38 ha–1. All monetary values are in Canadian dollars. At Beaverlodge and Lacombe, Eq. [2] best described the relationship between net economic return (y) and herbicide rate (x) while at Lethbridge the the relationship was best described by the quadratic model
 | [4] |
where a is the y intercept (net return at zero herbicide rate) and b and c are regression coefficients that describe the slope of the line.
Herbicide rates expressed as a proportion (0, 0.25, 0.50, 0.75, and 1.00) of the full recommended rate were used as the independent variable in the regression analysis. Entire regression curves for each variable were compared using an extra sum-of-squares F test:
 | [5] |
where SScom and DFcom are the sum-of-squares and degrees of freedom for the combined regression anlysis, and SSsep and DFsep are the sum-of-squares and degrees of freedom for the separate regression analyses. Probability (P) values were computed from the F statistic and differences between curves were deemed significant at 
< 0.05.
|
RESULTS AND DISCUSSION
|
|---|
Wheat plant density as a function of seeding rate varied considerably among years and locations and in most cases was below that targeted (Table 1). In most cases, however, doubling the seeding rate resulted in approximately twice the number of wheat plants. Densities were highest overall at Lethbridge and lowest at Beaverlodge. The reason for the differences among locations and years is not completely clear but may have been due to variable soil conditions including soil moisture at seeding time. In a previous study with wheat, plant density as a function of seeding rate (averaged across several years) was 22 and 31% lower at Beaverlodge than at Lacombe and Lethbridge, respectively (O'Donovan et al., 2005). It is possible that the relatively low pH gray luvisol soils at Beaverlodge were not conducive to optimum wheat seedling establishment. Wild oat plant densities in 2000 also varied among locations, averaging 50 ± 1.7, 33 ± 1.0, and 310 ± 12.2 plants m–2 at Beaverlodge, Lacombe, and Lethbridge, respectively. The relatively high wild oat population at Lethbridge may have been due to residual wild oat seed in the soil seed bank.
View this table:
[in this window]
[in a new window]
|
Table 1. Wheat plant density as a function of seeding rate (75 and 150 kg ha–1) at three locations in 2000, 2001, and 2002. Data are means ± standard error. Plants were counted at the two- to four-leaf stage of wheat.
|
|
Wild Oat Shoot Biomass and Seed in the Soil Seed Bank
Where no herbicides were applied, wild oat shoot biomass and the amount of wild oat seed in the soil seed bank were often several orders of magnitude higher than when herbicides were applied. For this reason, graphs describing these variables (Fig. 1
–4
) are presented for data collected only from plots where herbicides were applied to facilitate easier interpretation of differences among seeding rates at the different herbicide rates. The k values in the equations (Fig. 1–4) represent the amounts of wild oat biomass and seed that occurred in the unsprayed controls. The k values indicate that, in most cases, wild oat biomass and wild oat seed in the soil seed bank were less at the high than the low wheat seeding rates when herbicides were not applied. Averaged across locations and years, increasing the wheat seeding rate alone reduced wild oat biomass by 18% and wild oat seed in the soil seed bank by 46%. Reduced weed biomass and seed production at high compared with low crop seeding rates has been documented in numerous other studies with wheat and barley (e.g., Blackshaw et al., 2000; O'Donovan et al., 1999).
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 1. Relationship between wild oat shoot dry weight and herbicide rate at low and high wheat seeding rates (SR) at Beaverlodge in 2001 and 2002. Symbols (± standard errors) represent actual wild oat shoot dry weight. Lines represent dry-weight estimates from the equation y = k/[1 + exp(bg)xb], where y is estimated wild oat shoot dry weight as a function of herbicide rate (x), k is wild oat shoot dry weight in the absence of herbicide application, and b and g are scale parameters describing the shape of the response. Numbers above symbols are actual shoot dry-weight data at low (top) and high (bottom) wheat seeding rates.
|
|
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 2. Relationship between wild oat shoot dry weight and herbicide rate at low and high wheat seeding rates (SR) at Lacombe in 2000, 2001, and 2002. Symbols (± standard errors) represent actual shoot dry weight. Lines represent shoot dry weight estimates from the equation y = k/[1 + exp(bg)x b], where y is estimated wild oat shoot dry weight as a function of herbicide rate (x), k is wild oat shoot dry weight in the absence of herbicide application, and b and g are scale parameters describing the shape of the response. Numbers above symbols are actual shoot dry weight data at low (top) and high (bottom) wheat seeding rates.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 3. Relationship between wild oat shoot dry weight and herbicide rate at low and high wheat seeding rates (SR) at Lethbridge in 2000, 2001, and 2002. Symbols (± standard errors) represent actual shoot dry weight. Lines represent crop shoot dry weight estimates from the equation y = k/[1 + exp(bg)x b], where y is estimated wild oat shoot dry weight as a function of herbicide rate (x), k is wild oat shoot dry weight in the absence of herbicide application, and b and g are scale parameters describing the shape of the response. Numbers above symbols are actual shoot dry weight data at low (top) and high (bottom) wheat seeding rates.
|
|
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 4. Relationship between wild oat seed number in the soil seed bank and herbicide rate at low, and high wheat seeding rates (SR) in 2002. Symbols (± standard errors) represent actual wild oat seed number. Lines represent estimated wild oat seed number from the model y = k/[1 + exp(bg)x b], where y is estimated wild oat seed number as a function of herbicide rate (x), k is wild oat seed number in the absence of herbicide application, and b and g are scale parameters describing the shape of the response. Numbers above symbols are actual seed numbers at low (top) and high (bottom) wheat seeding rates.
|
|
The regression models describing the effects of herbicide rate on wild oat biomass and wild oat seed in the soil seed bank were significant (P < 0.001), and the F test indicated that the curves differed significantly (P < 0.01) among years and crop seeding rates at each location. At Beaverlodge in 2000, all rates of clodinafop resulted in virtually complete wild oat control and little or no wild oat biomass was collected from the herbicide-treated plots, thus precluding fitting regression equations to the data.
During the other years at Beaverlodge and in all years at the other locations, wild oat biomass decreased nonlinearly at both seeding rates as herbicide rates increased (Fig. 1
–3). Reductions in wild oat biomass were steepest between 25 and 50% of the recommended rates. In most cases, at each herbicide rate, there was less wild oat biomass at the high than the low wheat seeding rate (Fig. 1–3). In almost all instances, applying the herbicides at 25 or 50% of the recommended rates resulted in greater wild oat biomass compared with 75 or 100% of the recommended rates. There was little difference in wild oat biomass between the latter rates, especially at the high crop seeding rate.
Regression models describing the cumulative effects of herbicide rate on the amount of wild oat seed in the soil seed bank were highly significant (P < 0.001), and the responses (Fig. 4) were generally similar to those described for wild oat biomass each year. The F test indicated that responses to herbicide rate differed significantly (P < 0.001) between crop seeding rates. In the absence of herbicide application, wild oat seed in the soil seed bank decreased, on average, by 46% when the seeding rate was increased; but this still resulted in high numbers of wild oat seed in the soil seed bank (see k values, Fig. 4) and would thus be unacceptable as a wild oat management strategy. At both seeding rates, the amount of wild oat seed decreased with increasing herbicide rate and at each herbicide rate there was considerably less wild oat seed in the soil seed bank at the high compared with the low wheat seeding rate (Fig. 4). For example, at the full herbicide rate, wild oat seed decreased, on average, from 253 to 106 seeds m–2 when the seeding rate was increased. Thus seeding wheat at relatively high rates can contribute positively to herbicide performance and result in better wild oat management; however, application of the full herbicide rate resulted in the lowest numbers of wild oat seed in the soil seed bank at all locations, and would probably provide the most sustainable weed management in a continuous wheat system for the long term.
Wheat Seed Yield and Net Economic Return
All regression equations were significant (P < 0.05). In all cases, wheat yields were lowest when no herbicides were applied (Fig. 5
–7
) and in some cases yields were zero or close to zero by 2002. Wheat yields increased nonlinearly as herbicide rates increased. In most cases, there was little difference in yield between applying herbicides at 75 or 100% of the recommended rates. With the exception of Lethbridge in 2002, the F test indicated that the curves differed significantly among the crop seeding rates at each location each year. In most cases, wheat yields were superior at the high compared with the low seeding rate across the range of herbicide rates (Fig. 5–7). On average, wheat yield increased by 19% when the seeding rate was increased. Exceptions occurred with clodinafop at Beaverlodge in 2000, where the positive effects of the high seeding rate on wheat yield was not evident at all herbicide rates (Fig. 5), while at Lethbridge in 2001, imazamethabenz applied at 75 and 100% of the recommended rate resulted in higher yields at the low than the high seeding rates (Fig. 7). The reason for these anomalies is unclear.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 5. Relationship between wheat yield and herbicide rate at low and high wheat seeding rates (SR) at Beaverlodge in 2000, 2001, and 2002. Symbols (± standard errors) represent actual wheat yields. Lines represent wheat yield estimates from the equation y = a + b[1 – exp(–cx)], where y is the estimated wheat yield as a function of herbicide rate (x), a is the wheat yield at zero herbicide rate, and b and c describe the slope of the regression line.
|
|
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 6. Relationship between wheat yield and herbicide rate at low and high wheat seeding rates (SR) at Lacombe in 2000, 2001, and 2002. Symbols (± standard errors) represent actual wheat yields. Lines represent wheat yield estimates from the equation y = a + b[1 – exp(–cx)], where y is the estimated wheat yield as a function of herbicide rate (x), a is the wheat yield at zero herbicide rate, and b and c describe the slope of the regression line.
|
|
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 7. Relationship between wheat yield and herbicide rate at low and high wheat seeding rates (SR) at Lethbridge in 2000, 2001, and 2002. Symbols (± standard errors) represent actual wheat yields. Lines represent wheat yield estimates from the equation y = a + b[1 – exp(–cx)], where y is the estimated wheat yield as a function of herbicide rate (x), a is the wheat yield at zero herbicide rate, and b and c describe the slope of the regression line. Data for 2002 were pooled since curves did not differ significantly between seeding rates.
|
|
Net economic returns based on fixed prices and costs were variable among years and influenced by crop yield and both wheat seeding rate and herbicide rate (data not shown). Regression equations describing cumulative net economic return (across the 3 yr) as a function of herbicide rate were significant (P < 0.01). In most cases, net economic return increased with increasing herbicide rate. The F test indicated that the curves differed significantly (P < 0.05) among the crop seeding rates at Lacombe and Lethbridge and, at most of the herbicide rates, net return was greater at the high than the low wheat seeding rate. At Lethbridge, there was little or no difference in net return between the seeding rates at the two highest herbicide rates, and there was an economic advantage to applying the herbicides at 75 rather than 100% of the recommended rate (Fig. 8
). At Beaverlodge, the F test indicated that the P value (0.052) was marginally larger than the arbitrarily adopted significance level ( < 0.05); however, the trend, in terms of higher net returns at the higher wheat seeding rate, was similar to Lacombe and Lethbridge; thus, it is very likely that the differences at Beaverlodge were real. On average, net return increased by 16% when wheat was seeded at the higher rate. Overall, there was little or no economic advantage to reducing the herbicide rates below those recommended at either wheat seeding rate, and reducing rates below 75% almost always resulted in reduced revenue.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 8. Relationship between cumulative net economic return (across 3 yr) and herbicide rate. Symbols (± standard errors) represent actual net returns. At Beaverlodge and Lacombe, lines represent net return estimates from the equation y = a + b[1 – exp(–cx)], where y is the estimated net return as a function of herbicide rate (x), a is the net return at zero herbicide rate, and b and c describe the slope of the regression lines. At Lethbridge, lines represent net return estimates from the equation y = a + bx + cx2, where a is the y intercept (net return at zero herbicide rate) and b and c are regression coefficients that describe the slope of the line. Monetary amounts are in Canadian dollars.
|
|
Under the conditions of this study, the results clearly indicate that wild oat control with herbicides was necessary to optimize wheat yield and economic return regardless of wheat seeding rate. At both seeding rates, the lowest wheat yields and net economic returns occurred when no herbicides were applied and both variables increased nonlinearly with increasing herbicide rate. The results confirm those of previous studies where reducing graminicide rates below those recommended sometimes resulted in reduced wheat yields or economic returns (O'Donovan et al., 2003a, 2003b). Other studies in wheat have reported variable results with reduced herbicide rates in terms of maintaining yields and improving economic returns (Spandl et al., 1997; Stevenson et al., 2000; Stougaard et al., 1997). In general, reduced herbicide rates performed better at relatively low wild oat densities in both barley (Belles et al., 2000; Wille et al., 1998) and wheat (Holm et al., 2000). In another study with continuous barley, applying tralkoxydim for wild oat control at rates below those recommended often resulted in similar yields and higher economic returns compared with applying the full recommended rate, especially at relatively high barley seeding rates (O'Donovan et al., 2001). Barley has been shown to be generally more competitive with weeds than wheat (Dew, 1972). Thus reducing herbicide rates below those recommended may be more feasible in barley, especially when crop competitiveness is enhanced by increasing the seeding rate or growing competitive varieties. Studies in Europe comparing varying herbicide rates in spring (Salonen, 1992) and winter (Christensen, 1994) wheat and barley support this conclusion.
Previous studies have suggested that enhancing crop competitiveness either through growing competitive varieties or increasing crop seeding rates could improve the feasibility of applying herbicides at lower than recommended rates in spring barley (O'Donovan et al., 2001), spring wheat (Hucl, 1998; Lemerle et al., 1996), winter wheat (Christensen, 1994), canola (Brassica napus L.; Harker et al., 2003; O'Donovan et al., 2004), and in several crops grown in rotation (Beckie and Kirkland, 2003; Blackshaw et al., 2005). The results of this study also indicate that seeding wheat at a relatively high rate did indeed improve the performance of herbicides applied at variable rates for wild oat control; however, there was little or no overall economic advantage to reducing herbicide rates below those recommended. While there was sometimes little difference between applying the herbicides at 75 or 100% of recommended rates, reducing rates below 75% almost always resulted in higher wild oat biomass, wild oat seed in the soil seed bank, and reduced yields and net economic returns. Thus there may be significant risk associated with reducing herbicide rates for wild oat control in continuous wheat, even at relatively high wheat seeding rates.
|
ACKNOWLEDGMENTS
|
|---|
This research was supported in part by grants from the Western Grains Research Foundation and the Matching Investment Initiative of Agriculture and Agri-Food Canada. We are grateful to Greg Semach, Jim Drabble, Randall Brandt, Bob Pocock, and Larry Michielsen for excellent technical assistance.
|
REFERENCES
|
|---|
- Alberta Agriculture, Food and Rural Development. 2005a. Using 1000 kernel weight for calculating seeding rates and harvest losses. Available at http://www1.agric.gov.ab.ca/$department/deptdocs.nsf/all/agdex81?opendocument (accessed 12 Oct. 2005; verified 15 Feb. 2006). AAFRD, Edmonton, AB, Canada.
- Alberta Agriculture, Food and Rural Development. 2005b. Fertilizer requirements of cereals. Available at http://www1.agric.gov.ab.ca/$department/deptdocs.nsf/all/crop4689 (accessed 12 Oct. 2005; verified 15 Feb. 2006). AAFRD, Edmonton, AB, Canada.
- Ali, S. 2005. Crop protection 2005. Agdex 606-1. Alberta Agriculture, Food and Rural Development, Edmonton, AB, Canada.
- Beckie, H.J., and K.J. Kirkland. 2003. Implications of reduced herbicide rates on resistance enrichment in wild oat (Avena fatua). Weed Technol. 17:138–148.
- Belles, D.S., D.C. Thill, and B. Shafi. 2000. PP-604 rate and Avena fatua density effects on seed production and viability in Hordeum vulgare. Weed Sci. 48:378–384.
- Blackshaw, R.E., J.R. Moyer, K.N. Harker, and G.W. Clayton. 2005. Integration of agronomic practices and herbicides for sustainable weed management in a zero-till barley (Hordeum vulgare)–field pea (Pisum sativum) rotation. Weed Technol. 19:190–196.
- Blackshaw, R.E., G.P. Semach, and J.T. O'Donovan. 2000. Wheat seed rate affects redstem filaree (Erodium cicutarium) interference and seed production in a zero tillage cropping system. Weed Technol. 14:389–396.
- Christensen, S. 1994. Crop weed competition and herbicide performance in cereal species and varieties. Weed Res. 34:29–36.[CrossRef]
- Dew, D.A. 1972. An index of competition for estimating crop loss due to weeds. Can. J. Plant Sci. 52:921–927.
- Harker, K.N., G.W. Clayton, R.E. Blackshaw, J.T. O'Donovan, and F.C. Stevenson. 2003. Seeding rate, herbicide timing and competitive hybrids contribute to integrated weed management in canola (Brassica napus). Can. J. Plant Sci. 83:433–440.
- Holm, F.A., K.J. Kirkland, and F.C. Stevenson. 2000. Defining optimum rates and timing for wild oat control in spring wheat (Triticum aestivum). Weed Technol. 14:167–175.
- Hucl, P. 1998. Response to weed control by four spring genotypes differing in competitive ability. Can. J. Plant Sci. 78:171–173.
- Lemerle, D., B. Verbeek, and N.E. Coombes. 1996. Interaction between wheat (Triticum aestivum) and diclofop to reduce the cost of annual ryegrass (Lolium rigidum) control. Weed Sci. 44:634–639.
- O'Donovan, J.T. 2000. Weed scouting protocols for reducing herbicide costs and managing resistance. Final Report. Alberta Agric. Res. Inst., Edmonton, AB, Canada.
- O'Donovan, J.T., R.E. Blackshaw, K.N. Harker, G.W. Clayton, and R. McKenzie. 2005. Variable plant establishment contributes to differences in competitiveness with wild oat among wheat and barley varieties. Can. J. Plant Sci. 85:771–776.
- O'Donovan, J.T., K.N. Harker, R.E. Blackshaw, and R.N. Stougaard. 2003a. Effect of variable tralkoxydim rates on wild oat (Avena fatua) seed production, wheat (Triticum aestivum) yield, and economic return. Weed Technol. 17:149–156.
- O'Donovan, J.T., K.N. Harker, R.E. Blackshaw, and R.N. Stougaard. 2003b. Influence of variable rates of imazamethabenz and difenzoquat on wild oat (Avena fatua) seed production, and wheat (Triticum aestivum) yield and profitability. Can. J. Plant Sci. 83:977–985.
- O'Donovan, J.T., K.N. Harker, G.W. Clayton, D. Robinson, J.C. Newman, and L.M. Hall. 2001. Barley seeding rate influences the effects of variable herbicide rates on wild oat (Avena fatua). Weed Sci. 49:746–754.
- O'Donovan, J.T., J.C. Newman, K.N. Harker, R.E. Blackshaw, and D.W. McAndrew. 1999. Effect of barley plant density on wild oat interference, shoot biomass and seed yield under zero tillage. Can. J. Plant Sci. 79:655–662.
- O'Donovan, J.T., J.C. Newman, K.N. Harker, and G.W. Clayton. 2004. Crop seeding rate influences the performance of variable herbicide rates in a canola–barley–canola rotation. Weed Technol. 18:733–741.
- Salonen, J. 1992. Efficacy of reduced herbicide doses in spring cereals of different competitive ability. Weed Res. 32:483–491.
- SAS Institute. 1999. SAS/STAT user's guide. Version 8. SAS Inst., Cary, NC.
- Sharma, M.P., and W.H. Vanden Born. 1983. Crop competition aids efficacy of wild oat herbicides. Can. J. Plant Sci. 63:503–507.
- Spandl, E., B.R. Durgan, and D.W. Miller. 1997. Wild oat (Avena fatua) control in spring wheat (Triticum aestivum) and barley (Hordeum vulgare) with reduced rates of postemergence herbicides. Weed Technol. 11:591–597.
- Stevenson, F.C., F.A. Holm, and K.J. Kirkland. 2000. Optimizing wild oat (Avena fatua) control with ICIA 0604. Weed Technol. 14:608–616.
- Stougaard, R.N., B.D. Maxwell, and J. Harris. 1997. Influence of application timing on the efficacy of reduced rate postemergence herbicides for wild oat (Avena fatua) control in spring barley (Hordeum vulgare). Weed Technol. 11:283–289.
- Wille, M.J., D.C. Thill, and W.J. Price. 1998. Wild oat (Avena fatua) seed production in spring barley (Hordeum vulgare) is affected by the interaction of wild oat density and herbicide rate. Weed Sci. 46:336–343.
TOP
ABSTRACT
INTRODUCTION
