If you are conducting business with the Government of Saskatchewan by mail, please be advised that delivery may be delayed due to rotating postal strikes. Various measures are in place to ensure service to Saskatchewan residents and businesses during postal strike action.

Google Translate Disclaimer

A number of pages on the Government of Saskatchewan`s web site have been professionally translated in French. These translations are identified by a yellow text box that resembles the link below and can be found in the right hand rail of the page. The home page for French-language content on this site can be found here:

Renseignements en Français

Where an official translation is not available, Google™ Translate can be used. Google™ Translate is a free online language translation service that can translate text and web pages into different languages. Translations are made available to increase access to Government of Saskatchewan content for populations whose first language is not English.

The results of software-based translation do not approach the fluency of a native speaker or possess the skill of a professional translator. The translation should not be considered exact, and may include incorrect or offensive language Government of Saskatchewan does not warrant the accuracy, reliability or timeliness of any information translated by this system. Some files or items cannot be translated, including graphs, photos, and other file formats such as portable document formats (PDFs).

Any person or entities that rely on information obtained from the system does so at his or her own risk. Government of Saskatchewan is not responsible for any damage or issues that may possibly result from using translated website content. If you have any questions about Google™ Translate, please visit: Google™ Translate FAQs.

What Is in a Number? How to Interpret Herbicide Resistance Test Results

Crop Production News - 2018

By Clark Brenzil, PAg, Provincial Specialist – Weed Control, Regina

May 2018

There are two primary labs across the prairies that provide herbicide resistance testing services; AgQuest in Minto, Man., and the Saskatchewan Agriculture Crop Protection Laboratory in Regina, Sask.. Both labs use methods that are scientifically valid and both are capable of providing a proportional number of resistant individuals found in the test, but what does that number really mean and how should it impact your weed management decisions?

The natural mutation rates for herbicide resistance traits in weeds are typically very low, on the order of 1 in 100,000 to 1 in 100,000,000 individuals (plants) within a weed species for different herbicide resistance groups, also known as modes-of-action. While these may seem like long odds at a glance, consider that a modest population of 40 plants per square meter scaled up to a quarter-section is equal to 24 million individual plants of a single weed species on that quarter section. Therefore, at an estimated mutation rate of 1 in 1,000,000 for a Group 1 herbicide, there is the potential to be 24 resistant individuals in that quarter section each season.

Once that mutant survives the application of a herbicide, it has a good chance to continue to maturity and produce at least some seed. A weed like wild oats typically produces from 100 to 150 seeds per plant, potentially as low as 30 seeds in a more competitive crop and as high as 500 without competition.

Figure 1 shows the speed of resistance progression as a percentage of the population of a single weed species at different mutation rates. For the purpose of discussion, let’s assume that each weed seed always grows into a plant that produces 50 seeds; these seeds all germinate the following year and also survive the herbicide applied.  If we assume that all the susceptible weeds are controlled, this will continue until resistant individuals make up the entire population. It’s an overly simplified model that is intended to illustrate a point: increase the number of seeds a weed produces and the pace of herbicide resistance buildup is quicker.

The model in Figure 1 assumes that there is no herbicide rotation or other weed management that reduces the survival of the seed and its progression into a mature plant, and that just the one herbicide is applied. It also assumes that all of the seeds emerge at the same time and are equally exposed to the herbicide treatment.

Figure 1:  Herbicide-resistant weed populations change slowly at first and
then very rapidly in the final two years before complete herbicide failure.
The assumption above is that each plant produces 100 seeds and 50 of those
produce plants the following year.

What Figure 1 illustrates is that weed populations grow at a faster pace than the pace we like to believe. While we perceive these populations as having a more linear growth rate, populations actually grow at a logarithmic rate. Most people would be hard-pressed to notice resistance at less than 10 per cent of the population, and the model shows that populations can transition from what seems like nothing to full blown resistance in as little as a couple of years.  

In reality, there are many things that reduce the likelihood that a weed seed will produce a mature plant. Seed predators such as insects, mice, birds and general decay cause significant loss of seeds before germination. There are also a certain number of seeds, seedlings and plants that succumb to a fatal genetic mutation and either fail to germinate altogether or are compromised to a degree that they cannot reproduce. In nature, weeds do not germinate simultaneously and may then be subject to different management impacts from tillage to treatment with a different herbicide group.

However, dormancy in weed seeds acts as a type of genetic preservation effect, where the herbicide resistance mutation remains isolated from alternative weed control measures taken in a diverse rotation. Seeds containing the original mutation are able to re-emerge two, three or four years later when the same herbicide group is being used for weed control. The number of individuals that emerge after that break will be lower than the first year after seed is shed but will be a greater number than the single original seed, allowing the increase in the proportion of resistant individuals to continue, but at a slower pace.

When testing for herbicide resistance, the reporting of the results as a percentage can give the impression to the client of a level of precision in either the test itself, or of the management options available that do not exist.

The precision of the test itself relies on the proper sampling of the weed to be tested. The sample may have been taken directly from a patch that had survived the herbicide application, or it may have been drawn from a bin sample or as the crop was cleaned. Sampling from the patch provides much greater precision than the other two sources, allowing the client to determine what measures to take in order to prevent further spread of the patch.

At the testing lab, a subsample of the original sample provided is used to assess a manageable amount of seeds, which is usually around 100 seeds. Because of the dormancy discussed above, only a portion of that seed germinates, bringing the number down 25 to 50 per cent. Typical error rates for gauging experimental significance are five per cent, meaning any differences smaller than five per cent are not considered real, but are a result of the normal variability of a biological system. If both accountable error (seed germination) and unaccountable error (randomness in the system) are combined, a scale of a number out of 10, or 10 per cent increments, is as accurate as could be hoped for in the testing process.

If we return to Figure 1 and look at the speed of progression of resistance in a weed population and relate that to our testing accuracy, we can see (if we are using a 1 to 10 scale instead of a 1 to 100 scale), that for most cases, resistance moves from barely perceptible to perceptible to fully involved in short order. This has been confirmed by researchers in Arkansas following a small patch of a glyphosate-resistant weed found in a cotton field; the whole field was involved in three years. By the time we find a solid patch, resistant individuals that are less obvious have been spread across the entire field with harvest equipment.

Because of this rapid progression, the Crop Protection Lab in Regina has recently moved away from reporting specific percentages of the sample that presented as positive for herbicide resistance and moved toward reporting in the general levels high, medium and low.  Low is less than 20 per cent, high is greater than 75 per cent and medium is between those values. The reason for this is to provide a better characterization of the management implications for the level of resistance in the sample rather than focusing on the specifics of the result, which present false utility to the client in their decision making. A test with high resistance is an obvious indication that resistance evolution is well underway. A result of low resistance could be an indicator that resistance is just evolving and the producer should take immediate measures to keep it at a low level, or it could be the result of error in the test. The medium result is much more contentious and suggests that there are a limited number of applications of the tested herbicide group left before it is no longer effective on that weed. The modelling suggests as little as one use remaining for suppression of that weed.

A further argument could be made that even four levels (none, low, medium, high) is too much detail and a test result should only indicate whether there are definitively resistant individuals in the sample, none at all, or if the presence is marginal or within the range of error, say less than 10 per cent. This would give a more direct message to the producer that the useful life of that herbicide group on the tested weed species is coming to an end and herbicides within that group should either be avoided or mixed with a herbicide from another herbicide group to extend its life or when the group’s utility on other labelled weeds is needed.

The Bottom Line: If a herbicide resistance test is indicating that there is resistance evolving in your field, don’t try to split hairs too finely. A difference of a few percentage points (and some would argue tens of percentage points) is not a real difference. If resistance is indicated, immediate steps are necessary to change weed management practices to address that weed using an approach other than the herbicide group the weed is exhibiting resistance to. The writing is on the wall – will you read it?


We need your feedback to improve saskatchewan.ca. Help us improve