On a languid, damp July morning, I meet weed scientist Aaron Hager outside the old Agronomy Seed House at the University of Illinois’ South Farm. In the distance are round barns built in the early 1900s, designed to withstand Midwestern windstorms. The sky is a formless white. It’s the day after a storm system hundreds of miles wide rolled through, churning out 80-mile-per-hour gusts and prompting dozens of tornado watches and sirens reminiscent of a Cold War bomb drill. On about 23 million acres, or roughly two-thirds of the state, farmers grow corn and soybeans, with a smattering of wheat. They generally spray virtually every acre with herbicides, says Hager, who was raised on a farm in Illinois. But these chemicals, which allow one plant species to live unbothered across inconceivably vast spaces, are no longer stopping all the weeds from growing. Since the 1980s, more and more plants have evolved to become immune to the biochemical mechanisms that herbicides leverage to kill them. This herbicidal resistance threatens to decrease yields—out-of-control weeds can reduce them by 50% or more, and extreme cases can wipe out whole fields.  At worst, it can even drive farmers out of business. It’s the agricultural equivalent of antibiotic resistance, and it keeps getting worse. As we drive east from the campus in Champaign-Urbana, the twin cities where I grew up, we spot a soybean field overgrown with dark-green, spiky plants that rise to chest height.  “So here’s the problem,” Hager says. “That’s all water hemp right there. My guess is it’s been sprayed at least once, if not more than once.” “With these herbicide-resistant weeds, it’s only going to get worse. It’s going to blow up.” Water hemp (Amaranthus tuberculatus), which can infest just about any kind of crop field, grows an inch or more a day, and females of the species can easily produce hundreds of thousands of seeds. Native to the Midwest, it has burst forth in much greater abundance over the last few years, because it has become resistant to seven different classes of herbicides. Season-long competition from water hemp can reduce soybean yields by 44% and corn yields by 15%, according to Purdue University Extension. Most farmers are still making do. Two different groups of herbicides still usually work against water hemp. But cases of resistance to both are cropping up more and more. “We’re starting to see failures,” says Kevin Bradley, a plant scientist at the University of Missouri who studies weed management. “We could be in a dangerous situation, for sure.” Elsewhere, the situation is even more grim. “We really need a fundamental change in weed control, and we need it quick, ’cause the weeds have caught up to us,” says Larry Steckel, a professor of plant sciences at the University of Tennessee. “It’s come to a pretty critical point.”  On the rise According to Ian Heap, a weed scientist who runs the International Herbicide-Resistant Weed Database, there have been well over 500 unique cases of the phenomenon in 273 weed species and counting. Weeds have evolved resistance to 168 different herbicides and 21 of the 31 known “modes of action,” which means the specific biochemical target or pathway a chemical is designed to disrupt. Some modes of action are shared by many herbicides. One of the most wicked weeds in the South, one that plagues Steckel and his colleagues, is a rhubarb-red-stemmed cousin to water hemp known as Palmer amaranth (Amaranthus palmeri). Populations of the weeds have been found that are impervious to nine different classes of herbicides. The plant can grow more than two inches a day to reach eight feet in height and dominate entire fields. Originally from the desert Southwest, it boasts a sturdy root system and can withstand droughts. If rainy weather or your daughter’s wedding prevents you from spraying it for a couple of days, you’ve probably missed your chance to control it chemically.   Palmer amaranth “will zero your yield out,” Hager says. Several other weeds, including Italian ryegrass and a tumbleweed called kochia, are inflicting real pain on the farmers in the South and the West, particularly in wheat and sugar beet fields.    Chemical birth  Before World War II, farmers generally used cultivators such as plows and harrows to remove weeds and break up the ground. Or they did it by hand—like my mother, who remembers hoeing weeds in cornfields as a kid growing up on an Indiana farm. That changed with the advent of synthetic pesticides and herbicides, which farmers started using in the 1950s. By the 1970s, some of the first examples of resistance appeared. By the early 1980s, Heap and his colleague Stephen Powles had discovered populations of ryegrass (Lolium rigidum) that were resistant to the most commonly used herbicides, known as ACCase inhibitors, spreading throughout southern Australia. Within a few years, this species had become resistant to yet another class, called ALS-inhibiting herbicides.   The problem had just begun. It was about to get much worse. In the mid to late 1990s, the agricultural giant Monsanto—now a part of Bayer Crop Science—began marketing genetically engineered crops including corn and soybeans that were resistant to the commercial weed killer Roundup, the active ingredient of which is called glyphosate. Monsanto portrayed these “Roundup-ready” crops, and the ability to spray whole fields with glyphosate, as a virtual silver bullet for weed control. Glyphosate quickly became one of the most widely used agricultural chemicals, and it remains so today. It was so successful, in fact, that research and development on other new herbicides withered: No major commercial herbicide appears likely to hit the market anytime soon that could help address herbicide resistance on a grand scale.  Monsanto claimed it was “highly unlikely” that glyphosate-resistant weeds would become a problem. There were, of course, those who correctly predicted that such a thing was inevitable—among them Jonathan Gressel, a professor emeritus at the Weizmann Institute of Science in Rehovot, Israel, who has been studying herbicides since the 1960s. Stanley Culpepper, a weed scientist at the University of Georgia, confirmed the first case of Roundup resistance in Palmer amaranth in 2004. Resistance rapidly spread. Both Palmer amaranth and water hemp produce male and female plants, the former of which produce pollen that can blow long distances on the wind to pollinate the latter. This also gives the plant a lot of genetic diversity, which allows it to evolve faster—all the better for herbicide resistance to develop and spread. These super-weeds sowed chaos throughout the state. “It devastated us,” Culpepper says, recalling the period from 2008 to 2012 as particularly difficult. “We were mowing fields down.”   Staying alive Herbicide resistance is a predictable ­outcome of evolution, explains Patrick Tranel, a leader in the field of molecular weed science at the University of Illinois, whose lab is a few miles from the South Farm.  “When you try to kill something, what does it do? It tries to not be killed,” Tranel says.  Weeds have developed surprising ways to get around chemical control. One 2009 study published in the Proceedings of the National Academy of Sciences showed that a mutation in the Palmer amaranth genome allowed the plant to make more than 150 copies of the gene that glyphosate targets. That kind of gene amplification had never been reported in plants before, says Franck Dayan, a weed scientist at Colorado State University. Another bizarre way resistance can arise in that species is via structures called extrachromosomal circular DNA, strands of genetic material including the gene target for glyphosate that exist outside of nuclear chromosomes. This gene can be transferred via wind-blown pollen from plants with this adaptation.  But scientists are increasingly finding metabolic resistance in weeds, where plants have evolved mechanisms to break down just about any foreign substance—including a range of herbicides.  Let’s say a given herbicide worked on a population of water hemp one year. If any plants “escape,” or survive, and make seeds, their offspring could possess metabolic resistance to the herbicides used.  “When you try to kill something, what does it do? It tries to not be killed.” Patrick Tranel, University of Illinois There’s evidence of resistance developing to both of the chemical groups that have replaced or been mixed with Roundup to kill this weed: an herbicide called glufosinate and a pair of substances known as 2,4-D and dicamba. These two would normally kill many crops, too, but there are now millions of acres of corn and soy genetically modified to be impervious. So essentially the response has been to throw more chemicals at the problem. “If it worked last year, if you have metabolic resistance there’s no guarantee it’s going to work this year,” Hager says.  Many of these herbicides can harm the environment and have the potential to harm human health, says Nathan Donley, the environmental health science director at the Center for Biological Diversity, which is based in Tucson, Arizona. Paraquat, for example, is a neurotoxic chemical banned in more than 60 countries (it’s been linked to conditions like Parkinson’s), Donley says, but it’s being used more and more in the United States. 2,4-D, one of the active ingredients in Agent Orange, is a potential endocrine disruptor, and exposure to it is correlated with increased risk of various cancers. Glyphosate is listed as a probable human carcinogen by an agency within the World Health Organization and has been the subject of tens of thousands of lawsuits worth tens of billions. Atrazine can stick around in groundwater for years and can shrink testicles and reduce sperm count in certain fish, amphibians, reptiles, and mammals. Replacing glyphosate with herbicides like 2,4-D and dicamba, which are generally more toxic, “is definitely a step in the wrong direction,” Donley says.  Looking for solutions It’s not just chemicals. Weeds can become resistant to any type of control method. In a classic example from China, a weed called barnyard grass evolved over centuries to resemble rice and thus evade hand weeding. Because weeds can evolve relatively quickly, researchers recommend a wide diversity of control tactics. Mixing two herbicides with different modes of action can sometimes work, though that’s not the best for the environment or the farmer’s wallet, Tranel says. Rotating the plants that are grown helps, as does installing winter cover crops and, above all, not using the same herbicide in the same way every year.  Fundamentally, the solution is to “not focus solely on herbicides for weed management,” says Micheal Owen, a weed scientist and emeritus professor at Iowa State University. And that presents a “major, major issue for the farmer” and the current state of American farms, he adds.  BELL HUTLEY Farms have ballooned in size over the last couple of decades, as a result of rural flight, labor costs, and the advent of chemicals and genetically modified crops that allowed farmers to quickly apply herbicides over massive areas to control weeds. This has led to a kind of sinister simplification in terms of crop diversity, weed control practices, and the like. And the weeds have adjusted.  On the one hand, it’s understandable that farmers often do the cheapest thing they can to control weeds, to get them through the year. But resistance is a medium- to long-term problem running up against a system of short-term thinking and incentives, says Katie Dentzman, a rural sociologist also at Iowa State University. Her studies have shown that farmers are generally informed and worried about herbicide resistance but are constrained by a variety of factors that prevent them from really heading it off. The farm is too big to economically control weeds without spraying in a single shot, some farmers say, while others lack the labor, financing, or time.  Agriculture needs to embrace a diversity of weed control practices, Owen says. But that’s much easier said than done.  “We’re too narrow-visioned, focusing on herbicides as the solution,” says Steven Fennimore, a weed scientist with the University of California, Davis, based in Salinas, California. Fennimore specializes in vegetables, for which there are few herbicide options, and there are fewer still for organic growers. So innovation is necessary. He developed a prototype that injects steam into the ground, killing weeds within several inches of the entry point. This has proved around 90% effective, and he’s used it in fields growing lettuce, carrots, and onions. But it is not exactly quick: It takes two or three days to treat a 10-acre block. Many other nonchemical means of control are gaining traction in vegetables and other high-value crops. Eventually, if the economics and logistics work out, these could catch on in row crops, those planted in rows that can be tilled by machinery.  A company called Carbon Robotics, for example, produces an AI-driven system called the LaserWeeder that, as the name implies, uses lasers to kill weeds. It is designed to pilot itself up and down crop rows, recognizing unwanted plants and vaporizing them with one of its 30 lasers. LaserWeeders are now active in at least 17 states, according to the company.   You can also shock weeds by using electricity, and several apparatuses designed to do so are commercially available in the United States and Europe. A typical design involves the use of a height-adjustable copper boom that zaps weeds it touches. The most obvious downside with this method is that the weeds usually have to be taller than the crop. By the time the weeds have grown that high, they’ve probably already caused a decline in yield.  Weed seed destructors are another promising option. These devices, commonly used in Australia and catching on a bit in places like the Pacific Northwest, grind up and kill the seeds of weeds as wheat is harvested. An Israeli company called WeedOut hatched a system to irradiate and sterilize the pollen of Palmer amaranth plants and then release it into fields. This way, female plants receive the sterile pollen and fail to produce viable seeds.  “I’m very excited about this [as] a long-term way to reduce the seed bank and to manage these weeds without having to spray an herbicide,” Owen says.  WeedOut is currently testing its approach in corn, soybean, and sugar beet fields in the US and working to get EPA approval. It recently secured $8 million in funding to scale up.  In general, AI-driven rigs and precision spraying are very likely to eventually reduce herbicide use, says Stephen Duke, who studies herbicides at the University of Mississippi: “Eventually I expect we’ll see robotic weeding and AI-driven spray rigs taking over.” But he expects that to take a while on crops like soybeans and corn, since it is economically difficult to invest a lot of money in tending such “low-value” agronomic crops planted across such vast areas. A handful of startups are pursuing new types of herbicides, based on natural products found in fungi or used by plants to compete with one another. But none of these promise to be ready for market anytime soon. Field day  Some of the most successful tools for preventing resistance are not exactly high-tech. That much is clear from the presentations at the Aurora Farm Field Day, organized by Cornell University just north of its campus in Ithaca, New York.  For example, one of the most important things farmers can do to prevent the spread of weed seeds is to clean out their combines after harvest, especially if they’re buying or using equipment from another state, says Lynn Sosnoskie, an assistant professor and weed scientist at Cornell.  Combines are believed to have already introduced Palmer amaranth into the state, she says—there are now at least five populations in New York.  Another classic approach is crop rotation—switching between crops with different life cycles, management practices, and growth patterns is a mainstay of agriculture, and it helps prevent weeds from becoming accustomed to one cropping system. Yet another option is to put in a winter cover crop that helps prevent weeds from getting established.  “We’re not going to solve weed problems with chemicals alone,” Sosnoskie says. That means we have to start pursuing these kinds of straightforward practices. It’s an especially important point to hammer home in places like New York state, where the problem isn’t yet top of mind. That’s in part because the state isn’t dominated by monocultures the way the Midwest is, and it has a more diverse patchwork of land use.  But it’s not immune to the issue. Resistance has arrived and threatens to “blow up,” says Vipan Kumar, also a weed expert at Cornell. “We have to do everything we can to prevent this,” Kumar says. “My role is to educate people that this is coming, and we have to be ready.” Douglas Main is a journalist and former senior editor and writer at National Geographic.