Ever since Thomas Malthus issued his dire prediction in 1789 that population growth would always exceed food supply, scientists have worked to prove him wrong. So far, they’ve helped farmers keep pace by developing bigger and better varieties of crops and other agricultural innovations.

Now researchers are taking an even more audacious step: reprogramming plants to make photosynthesis more efficient. And it seems to be paying off.

Tobacco plants that were genetically engineered to optimize photosynthesis outgrew their conventional relatives by as much as 40 percent, said a study in the journal Science.

Scientists have homed in on photosynthesis because it offers one of the few remaining options for drastically boosting crop yields. Plant breeders have already selected varieties that produce more of whatever we want to eat — be it leaves, fruit, roots or seeds — when grown under ideal conditions.

Despite its ability to build towering redwoods and vast coral reefs, photosynthesis is a fairly inefficient process. Only a tiny fraction of available light gets used to produce sugars and other carbohydrates.

“The photosynthetic system has evolved to be very flexible, rather than fully optimal,” said Christine Foyer, a plant biologist at the University of Leeds in England.

Part of the problem is that plants spend a lot of energy compensating for a bug in their operating system. It involves an enzyme called RuBisCO whose job is to grab carbon dioxide molecules and send them down the assembly line.

The process worked great when photosynthesis first evolved billions of years ago, because there was no oxygen in the atmosphere. But once it built up — thanks to photosynthetic plankton — RuBisCO began latching on to the wrong gas. The resulting compound was not just useless but toxic.

So plants had to find a way to convert it into something safe and functional. But they picked a way that is “very complex and very energy-intensive,” said plant biologist Don Ort of the University of Illinois at Urbana-Champaign, the study’s senior author.

The cumbersome process, known as photorespiration, reduces crop yields by 20 percent to 50 percent. “That fixing of an oxygen is really like anti-photosynthesis,” Ort said.

So his team updated the photorespiration algorithm. They took tobacco plants and inserted new genes into their DNA that created a shortcut for processing the unwanted compound. They tried three alternatives and also silenced a gene.

In greenhouse experiments, the engineered plants put on almost 25 percent more biomass than their unaltered counterparts. Field trials — the gold standard for testing crops — had even better results, with some plants outproducing their relatives by 40 percent.

“Some of that, we think, was due to compound interest,” Ort said. Young plants grew faster and increased their leaf area, which allowed them to photosynthesize even more.

The researchers have started making the same changes in food crops such as soybeans and potatoes. However, it won’t work on crops such corn and sugarcane, which have a different way of fixing carbon.

Ort’s team is also collaborating with another group that engineered tobacco plants to utilize more light, resulting in a 15 percent increase in productivity. “We’re now in the process of what we call stacking those two traits,” Ort said. Models suggest that the benefits will add up, boosting productivity by more than 50 percent. But, Ort cautioned, “until you do the experiments, you don’t know.”

The study’s results are a great start, said Heike Sederoff, a plant biologist at North Carolina State University in Raleigh. But she said researchers will still have to assess whether the new trait persists across generations. She said the potential of these crops also will depend partly on how attitudes and regulations evolve.