Previous PageTable Of ContentsNext Page

Time of weed removal for canola

K Neil Harker1, George W. Clayton1 and Adrian M Johnston2

1AAFC, Lacombe Research Centre, 6000 C&E Trail, Lacombe, Alberta, Canada, T4L 1W1, E-mails: and
AAFC, Melfort Research Centre, P.O. Box 1240, Melfort, Saskatchewan, Canada, S0E 1A0, E-mail:


Roundup Ready canola (‘Quest’) was planted in Alberta, Canada in 1997 and 1998. Weeds were removed with glyphosate at three weekly intervals after crop emergence. In 1997, yield reductions were dramatic after late applications at both sites. Delaying glyphosate application by 2 weeks caused yield losses of 19% (Lacombe) and 57% (Legal). Results in 1998 were more mixed, but usually indicated a definite advantage for early glyphosate application. At Lacombe in 1998, there was no time of weed removal effect on yield due to very light weed emergence before the early application intervals. At the other sites, in plots where glyphosate was applied 3 weeks after crop emergence, yield was reduced by 27% (Beaverlodge) and 25% (Ellerslie), when compared to plots with early glyphosate application.

KEYWORDS: Competition, Duration of Interference, Yield Loss


Canola yield loss due to weed interference has been documented in several western Canadian studies (Dew 1976, O’Donovan et al. 1988 and 1989, O’Donovan 1991, and O’Sullivan et al. 1985). Improved weed management and the continual development of new and more effective canola herbicides has increased canola quantity and quality. Initially, western Canada canola growers depended rather heavily on pre-plant incorporated dinitroaniline herbicides. In recent years there has been much greater reliance on post-emergence herbicides for weed control; this trend has only strengthened with the widespread adoption of herbicide-tolerant canola. Post-emergence herbicides are sometimes relied upon for the control of all weeds in a given field. However, there has been a tendency to delay herbicide application to accommodate a single herbicide treatment for early- and late-emerging weeds. The result can be that weed interference leads to irrecoverable yield loss before the first herbicide is applied.

For post-emergence herbicides, wide “application windows” have gained broad market acceptance over the last decade. Growers appreciate the flexibility of a wide “window” assuming that they have more time before they need to spray. Indeed, many late applications are still “safe” on the crop as far as phytotoxicity is concerned, but are not safe in terms of sacrificed yields. There is a tendency to delay herbicide application until the last weed has emerged or until one more “flush” germinates. In addition, most producers agree that spraying herbicides is one of their least desirable tasks, and therefore, it is easy to procrastinate.

There are several consequences that develop when weeds are sprayed at relatively late crop stages. On the positive side, late spraying will help the field look “clean” at harvest. On the negative side, large weeds are usually more tolerant to herbicides and advanced crops are often less tolerant to herbicides. The major negative consequence that arises from late spraying is lower yields. Crop/weed interference studies have indicated that early emerging weeds are most detrimental to yield (O’Donovan et al. 1985, Peters and Wilson 1983). Therefore, late spraying may preserve some crop yield, but a significant portion may have already been irretrievably lost.

Our objective was to determine how weed removal with a non-residual herbicide at weekly intervals after crop emergence would influence canola seed yield.


Field experiments were conducted in central and northern Alberta at two locations in 1997 (Lacombe and Legal) and three locations in 1998 (Lacombe, Beaverlodge, and Ellerslie). Prior to seeding, cultivated oat was cross-seeded into the plots to supplement the natural weed infestation. ‘Quest’ canola (Roundup Ready) was planted on barley stubble with a Conserva-Pak double-shoot seeder equipped with knives on 30 cm row spacing. Fertilizer (N, P205, and K2O) was banded beside and below the seed during the seeding operation at soil test recommended rates. Treatments included three weed removal dates [1, 2, and 3 weeks after crop emergence (WACE)] and an untreated check. Weeds were removed with 450 g ai ha-1 of glyphosate applied at a water volume of 100 L ha-1. Individual plot size was 3 X 15 m. The experiment was designed as a randomized complete block with four replications. Weed and crop growth stages at the time of glyphosate application were recorded and yield was determined at harvest.


For almost all sites and years the highest canola yields were in plots where weeds were removed early. In 1997 at Lacombe, there was an 8 bu/ac yield loss that accompanied weed removal at 3

WACE (Figure 1). There was no significant yield loss when glyphosate treatment was delayed from 1 to 2 WACE. At Legal (Figure 1), yield losses were severe when glyphosate was applied at any interval beyond TWR 1. It should be noted that the TWR 2 interval at Legal was actually three days later than shown due to rainfall on the projected application date. Nevertheless, the combination of a heavy natural weed infestation with weed removal approximately 2.5 weeks after canola emergence led to yields that were not significantly higher than the untreated check. Legal was the only site in this study where later applications at TWR 2 or 3 did not protect some of the crop yield.

Figure 1. Time of weed removal effects on canola seed yield at two Alberta locations in 1997. TWR = Time of Weed Removal at 1, 2, or 3 weeks after canola emergence (WACE). Weedy = untreated check. LSD (0.05) for Lacombe and Legal = 4 and 3 bu/ac, respectively.

In 1998, time of weed removal varied substantially among sites (Figure 2). The Beaverlodge and Ellerslie sites led to very similar results. At both sites there was no significant yield loss until glyphosate application was delayed until 3 WACE. The 2 week delay until TWR 3 resulted in 27 and 25% yield losses at Beaverlodge and Ellerslie, respectively. At Lacombe, there was no time of weed removal effect even though untreated plots yielded 32% less than those treated with glyphosate. Due to unusually dry conditions at Lacombe in 1998, there were few early emerging weeds for the early glyphosate applications. One might expect that yields would actually be reduced when herbicide applications are made before the majority of the weeds emerge; such was not the case at Lacombe. It appeared that the removal of a few early emerging weeds was as important to canola yield as the later removal of many more weeds. These results support data from other studies indicating that early emerging weeds are the greatest threat to crop yield (O’Donovan et al. 1985, Peters and Wilson 1983).


Canola yield losses due to weed competition occurs at early crop stages. In situations where pre-plant incorporated and pre-emergence herbicides will control a significant portion of the weed spectrum, these herbicide could be employed to reduce the risk of weed-related yield loss. Growers should not wait for late weed flushes before they apply post-emergence herbicides. Late emerging weeds may require a second application, or the initial use of a residual herbicide. But, in order to protect high yields, the first application should be made early.

Figure 2. Time of weed removal effects on canola seed yield at three Alberta locations in 1998. TWR = Time of Weed Removal at 1, 2, or 3 weeks after canola emergence (WACE). Weedy = untreated check. LSD (0.05) for Beaverlodge, Lacombe, and Ellerslie = 5, 4, and 5 bu/ac, respectively.

Further research is needed to determine if these effects are consistent under other growing conditions, and to determine the influence late emerging weeds have on ease of harvest and weed seed banks.


The authors express their appreciation to B. Pocock, L. Michielsen, J. Drabble, M. Baig, and M. Kidnie for plot maintenance and technical assistance.


1. Dew, D. A. and Keys, C. H. 1976. An index of competition for estimating loss of rape due to wild oats. Canadian Journal of Plant Science 56:1005-1006.

2. O’Donovan, J. T., de St. Remy, E. A., O’Sullivan, P. A., Dew, D. A., and Sharma, A. K. 1985. Influence of the relative time of weed emergence of wild oat (Avena fatua) on yield loss of barley (Hordeum vulgare) and wheat (Triticum aestivum). Weed Science 33:498-503.

3. O’Donovan, J. T., Sharma, A. K., Kirkland, K. J. and de St. Remy, E. A., 1988. Volunteer barley (Hordeum vulgare) interference in canola (Brassica campestris and B. napus). Weed Science 36:734-739.

4. O’Donovan, J. T., Kirkland, K. J., and Sharma, A. K. 1989. Canola yield and profitability as influenced by volunteer wheat infestations. Canadian Journal of Plant Science 69:1235-1244.

5. O’Donovan, J. T. 1991. Quackgrass (Elytrigia repens) interference in canola (Brassica campestris). Weed Science 39:397-401.

6. O’Sullivan, P. A., Weiss, G. M., and Kossatz, V. C. 1985. Indices of competition for estimating rapeseed yield loss due to Canada thistle. Canadian Journal of Plant Science 65:145-149.

7. Peters, N. C. B. and Wilson, B. J. 1983. Some studies on the competition between Avena fatua L. and spring barley. II. Variation of A. fatua emergence and development and its influence on crop yield. Weed Research 23:305-311.

Previous PageTop Of PageNext Page