Tom Wolf
I remember the moment well. It was 1986, and I was a summer student for DuPont. One of my tasks was to help demonstrate the performance of a new group of herbicides, the sulfonylureas (SU), on broadleaf weed control in cereals. As remains commonplace, we compared these new products to existing standards, in this case Bucrtril-M (MCPA/bromoxynil); and to make it more challenging, we sprayed the plots a bit late to give the likes of redroot pigweed, lamb’s-quarters and wild buckwheat an edge.
Of course, the new SUs did very well. As we showed the crispy weed skeletons to our tour attendees, we looked for signs of excitement in the farmers’ stoic eyes. Why wouldn’t they be impressed? These new herbicides showed not just weed-killing prowess, but also crop safety, low use rates and low mammalian and environmental toxicity. At that moment, I became convinced that we were finally poised to win the age-old battle against not just broadleaf weeds in cereals, but with the recent introduction of the remarkably effective Group 1 graminicides, that battle as well! We would control weeds. The future looked bright. Better get some shades.
It is with a bit of embarrassment that I look back on those naive views now. Lamb’s-quarters, wild buckwheat, wild oats and green foxtail are alive and well in 2025. In fact, these four weeds have not been displaced from their positions as the top five most abundant weeds in Western Canada since surveys began in the 1970s, according to Julia Leeson of Agriculture and Agri-Food Canada. Not only that, Leeson’s surveys show there are a host of weeds rising in abundance on our farms (kochia, volunteer canola, annual sow-thistle), despite intensive herbicide use.
If my enthusiastic young self transported through time from 1986 to the present and asked me where it all fell apart, how would I answer?
I’d probably start with genetics. Weed genomes are remarkably diverse. They contain genes that allow them to adapt to a large variety of conditions. For one, they have dormancy traits in their seeds that permit them to schedule their emergence to conditions that increase their likelihood of success and procreation. Most can wait many years to germinate. When they do, they can tailor their growth habits to both competitive and open environments, long and short growing seasons and dry and wet conditions with vast ranges of physical size and seed production.
Weeds are also well adapted to move their seed long distances by wind, water or other vectors. In other words, if weed control measures slip up, weeds will quickly take advantage.
A second aspect of the diverse genome is the presence of different versions of the same enzymes in the cells of some small percentage of plants. Enzymes are proteins and are comprised of long chains of amino acids. The sequence of these amino acids determines how the proteins fold into clusters that ultimately function to catalyze chemical reactions in the cell. Many herbicides, including the Group 1, 2 and 9 modes of action, inhibit single enzymes that are critical to the life of the plant by binding with the given enzyme and limiting its ability to catalyze the designated chemical reaction.
A single change in the amino acid sequence can change the fold pattern and render the enzyme ineffective, causing plant death. Some sequence changes do not affect enzyme function but instead prevent the herbicide from binding on it. This sequence change confers what is known as binding site resistance. It turns out that weeds contain genes with just these types of mutations, albeit in very low numbers.
Unbeknown to my 1986 self, by repeatedly using these herbicides, we quickly selected for the resistant genotypes. Between 1988 and 2021, scientists have documented 68 cases of Group 2 resistance in Canada, according to Dr. Ian Heap’s website weedscience.org, spanning 25 weeds including redroot pigweed, lamb’s-quarters, wild buckwheat, cleavers and kochia. By comparison, the old industry standard of MCPA/bromoxynil counts five cases of resistance on five weeds between 1990 and 2019.
Binding site resistance is the most obvious reason for problems with sole reliance on herbicides for weed control. However, more subtle factors can also play a role in reducing herbicide effectiveness, such as reduced absorption and translocation, both in response to genes and environmental conditions.
We select for resistance by the act of using a herbicide. Yes, we can delay resistance by using better application practices that increase efficacy and prevent tolerance to herbicides from polygenic resistance, but the use of herbicides commits us to a one-way street of inevitable resistance in at least one weed on all farms.
Weed escapes do not have to be caused by resistance. One of the most frequent causes is poor control in sprayer wheel tracks. There are a number of reasons for this, including the physical damage of wheels on weeds, reducing plant vigour and product uptake. The displacement of sprays due to wheel aerodynamics (increasing with speed and lug depth) and the presence of dust may also play a role, particularly with dust-sensitive products such as glyphosate and diquat. All these contributing factors have become more prominent since 1986.
Secondary emergence of weeds due to increased soil-seed contact after compaction can also contribute to escapes from post-emergence herbicide application. Larger, heavier, faster-moving equipment is once again a large part of the cause.
A recent publication by the Grains Research and Development Corporation (GRDC) in Australia suggested that the use of coarser sprays and higher water volumes, combined with slower travel speeds, was most likely to decrease the loss of control in wheel tracks. They suggested that mudguards and flaps may also be beneficial. Spraying when the soil is moist may also minimize dust.
A residual herbicide provides some level of insurance that a certain level of weed control persists even when weeds survive a post-emergent application. In cereal and oilseed production, we are only now rediscovering the utility and power of these products.
Harvest seed destructors, though still not commonplace, are available and these have minimized the spread of weed seeds that plants have not shed by harvest time. Although seeds still attached to weeds may be a small portion of total seeds produced, the act of destroying them has become very relevant when one considers the ability of a modern combine straw and chaff spreader to distribute residue 40 to 50 feet. A small patch can grow very quickly when spread that far.
So, where did we go wrong in these past 40 years? The first mistake is in our words. We believed that weed control was possible and desirable. To a large degree, we still believe that. So, the first change is to be humble and admit that we are really talking about weed management. We have not been eliminating weeds. We may have been managing them, when they’re not managing us.
I would tell my younger self not to underestimate the diversity of weeds and their ability to adapt to niche conditions. Despite the tremendous efficacy of herbicides, I would warn against the mistake of relying too much on any one method of weed control. I would stress the need to develop integrated systems in which weed population densities could be reduced without the use of herbicides. We know what these are – diverse crop rotations, including the integration of either short- or long-term forage stands. Higher seeding rates and competitive crops and cultivars have shown to be very powerful ways of reducing weed density.
I recognize that in light of tremendous results achieved with herbicides, and only sporadic accounts of resistance in the early years, these practices would have been a hard sell in the 1980s, but even back then, herbicides failed and the setback in terms of yield loss and soil weed seedbank fortification was felt for years.
I would also stress the importance of early detection and intervention. A weed patch that survived a spray pass could certainly be resistant, but not necessarily. Perhaps there is an underlying soil condition that merits attention. There is no substitute for eyes on the ground, diligence and attention to detail, even when managing large acreages.
A good system is one in which some level of consistency and reliability can be achieved. In farming, as in any business, a predictable business environment is essential to long-term planning. A sudden change in the rules of the game is never welcome. Yes, adaptation is possible, but being on the back foot and playing catch-up is not a winning situation. Agronomy is no different.
On my trip to Australia this past winter, I learned that some farmers for whom herbicide resistance had become unmanageable had reverted to inversion tilling. They basically plowed their land to put the troublesome weed-seed-containing layer of soil 30 centimetres under. That allowed them to start fresh. If and when another inversion was required in the future, the weed seeds would have hopefully lost their viability.
Even if this were the answer, building up residue and organic matter after such a dramatic intervention and surviving erosion susceptibility would put the land in a vulnerable position in the meantime. A better way may indeed be to integrate all of our available tools and manage weeds in the long term.
