Some scientists say using only herbicides to kill weeds is not a long-term solution. They think using evolution to stop weeds would be more effective. Christopher Berkey for The New York Times

Depending on your point of view, barnyardgrass is a nightmare or a marvel.

That's because it's a supremely triumphant weed. Barnyardgrass is particularly devastating on rice farms, where losses sometimes reach 100 percent. It has evolved resistance to many herbicides that farmers rely on to control weeds, and each plant can produce up to a million seeds, which nestle into the soil, waiting for a chance to regrow.

All told, barnyardgrass and the world's many other weeds result in a 10 percent reduction in the productivity of crops. In the United States alone, they cause an estimated $33 billion in losses each year. Herbicides can reduce the toll, but within a few years of the introduction of a new chemical, resistances evolve.

Now some scientists argue that we can find more effective ways to fight weeds by studying their evolution.

"They're amazingly successful plants. They've evolved to take advantage of us," said Ana L. Caicedo of the University of Massachusetts.

Barnyardgrass has changed dramatically from its nonweed ancestors, evolving a tolerance to the waterlogged soils in rice fields. It has also evolved to look just like rice.

"The biotech folks would have no clue about how to make one plant look like another plant," said R. Ford Denison, an evolutionary biologist at the University of Minnesota. "And yet a thousand years of selection on a small patch of the earth was enough to give it crop mimicry - and flood tolerance."

Certain traits help wild species become weeds - they already grow fast, for example, and make lots of seeds.

Other weeds evolved from the union of wild plants and crops. In the 1970s, wild beets in Europe released pollen that fertilized sugar beets growing on farms.

Even crops can turn into weeds. "We domesticated a plant from the wild, and somehow it de-domesticated itself - which I think is pretty exciting," Dr. Caicedo said.

Among those crops gone wild is a weed known as red rice. Domesticated rice was bred to hold onto its seeds when farmers harvested it. Red rice evolved fragile seeds that fell to the ground, sometimes going dormant. Those dormant seeds can sprout later. "It's a fantastic trait for a weed to have. You're hedging your bets," Dr. Caicedo said.

The DNA of de-domesticated weeds acquires new mutations to different genes. "You've got a new bag of genetic tricks," said Norman Ellstrand of the University of California, Riverside.

The past century brought a slew of weed-killing chemicals that soon became ineffective. Today 217 species of weeds are resistant to at least one herbicide, according to the International Survey of Herbicide Resistant Weeds.

In the 1970s, there was great hope for a new herbicide called glyphosate, sold by Monsanto as Roundup. Early studies revealed no resistance evolving in weeds, raising hopes that, at last, farmers had escaped from evolution. In the 1980s, Monsanto increased glyphosate's popularity by introducing genetically modified crops carrying a gene that gave them resistance to the herbicide. Instead of using several different herbicides, many farmers could now use just one.

And yet through evolution, the weeds have become resistant to glyphosate anyway.

Earlier this year, the agricultural consulting firm Stratus reported that half of American farms had glyphosate-resistant weeds in 2012, up from 34 percent the year before.

Some researchers have argued that weeds could be foiled by combining two resistance genes in one plant, so that farmers could apply two herbicides at once. The odds of a weed having resistance to both chemicals would be tiny.

In the journal Trends in Genetics, a team of French and American scientists present a counterargument: spraying one chemical can drive the evolution of an all-purpose stress response system that can defend against more than one herbicide.

David Mortensen, a weed biologist at Pennsylvania State University, predicted that such plants would create a new generation of resistant weeds.

He and his colleagues are investigating controlling weeds by planting crops like winter rye that can kill weeds by blocking sunlight and releasing toxins. "You want to spread the selection pressure across a number of things that you're doing, so that the selection pressure is not riding on one tactic," he said.

Scientists have already explored evolution-based strategies for battling other enemies, like bacteria that evolve resistance to antibiotics.

"We should be looking at this more carefully," Dr. Ellstrand said. "And we're just getting to it now."

The New York Times

(China Daily 07/28/2013 page10)