Reptilian venoms are among nature's deadliest innovations. Every year, these chemical weapons, delivered through the fangs of vipers, cobras, kraits, and more, kill up to 200,000 people, not to mention millions of rodents, birds, and other animals. These toxins are also causing painful convulsions in the halls of academia, where two rival camps of scientists are engaged in a bitter (toxic?) debate about their evolutionary origins.

Ostensibly, this is a spat over how many times venom evolved among snakes and lizards—just the once, or on several repeated occasions. It's also about whether supposedly non-venomous species like bearded dragons and Komodo dragons actually pack a toxic bite. But really, it's about the see-sawing nature of scientific truth, and the ways in which hypotheses become enshrined as facts.

Until recently, textbooks stated that only a quarter of snakes are venomous, and that they evolved their toxins separately to the only two venomous lizards: the Gila monster and beaded lizard. But in 2005, Bryan Fry from the University of Queensland proposed a different origin story for reptile venom.

For years, Fry had found toxins in the mouths of supposedly non-venomous species, like the lace monitor, bearded dragon, and Asian ratsnake. He’d also showed that the mouth glands of these and other reptiles all activate the same basic set of nine toxin-making genes.

Based on these discoveries, Fry proposed that reptile venom evolved just once, in an ancestral lizard that lived 170 million years ago. Its descendants include all snakes and some lizards, including iguanas, chameleons, and monitor lizards—a group that Fry christened the Toxicofera. Some of its members, like pythons, abandoned their venomous inheritance. Others, like cobras and Gila monsters, went to town with it and developed truly potent weapons. Many others, like the Komodo dragon and bearded dragons, retained milder cocktails but were still, contrary to popular belief, venomous.

Fry published the Toxicofera idea in Nature in 2005. In the intervening decade, it has become enshrined as established fact. Scientists have referenced it in their papers. Journalists, including myself, have written about it uncritically. Stephen Fry referred to it in an episode of the British quiz show QI. In the landmark BBC series, Life, David Attenborough told the world about the Komodo's venomous bite.

Certainly, when John Mulley, an evolutionary biologist from Bangor University, first started studying venom genes, he took Toxicofera at face value. “It meant that there were potentially lots of species to look at, since most reptiles were venomous,” he says. “It was all very exciting, but the more I looked at it, the shakier it got.”

Mulley, the graduate student Adam Hargreaves, and several others started finding that the so-called venom genes were activated throughout the bodies of Toxicoferan reptiles, rather than just in the mouth and salivary glands. The genes showed up in the genitals, brain, heart, intestine, stomach, liver, and muscles. And some of these same “toxin” genes are activated in the leopard gecko—a lizard that falls outside the Toxicofera clique. Some are even activated in human saliva.

Mulley and Hargreaves concluded that these genes aren't venom genes at all. Instead, they're probably “housekeeping” genes that are involved in general processes like making proteins or sending signals between cells. They're part of a reptile's normal physiology. Some snakes and lizards have tweaked them to disrupt other animals' physiology, making venom. “This means that the genetic material for making truly toxic compounds is present in many reptiles,” says Stephen Mackessy from the University of North Colorado. “But just because we find a toxin-like gene in gland tissue, we can't necessarily assume that it's a toxin or that the animal that produces it is venomous.”

As an example, Fry's team showed that the bearded dragon makes crotamine in its mouth glands—a substance that was previously known only from rattlesnake venom. “Why does a largely herbivorous lizard have this in its oral products? It means that its presence has nothing to do with venom,” says Scott Weinstein from the Women's and Children's Hospital in Adelaide. (He's skeptical about the Komodo dragon too. “[The Komodos] have powerful serrated teeth. I think they make their prey bleed to death through massive physical trauma.”)

Fry dismisses these arguments. Of course, toxins are recruited from ordinary physiological genes, he says. “How could it be any other way? They aren't magically created by the toxin fairy. They evolve.” So, it's obvious that toxin genes should be active in other tissues, and that doesn't refute their role in venom. Consider hyaluronidase—a  protein that attacks sugars. Sperm cells use it to break into eggs and fertilize them. But snakes and scorpions use it to destroy flesh; if you inject it into another animal at very large concentrations, it starts breaking down the local tissues. “Such proteins could perform multiple functions in the same animal,” says Fry.

Okay, says Mulley, but you'd still expect these genes to be more strongly active in venom glands than other body parts, since their owners need to manufacture toxins at large concentrations. “And we were finding them at the same level all over the body,” he says.

That doesn't matter, says Kartik Sunagar at the Hebrew University of Jerusalem, who sides with Fry. “Venom is extremely expensive to produce so animals that don't use it lose it,” he says. For example, the marbled sea snake, which eats fish eggs rather than fish, has no teeth, shriveled venom glands, and venom that's about 100 times less toxic than those of its close relatives. So the fact that Toxicoferans are still activating these supposed toxin genes at all in their mouth glands speaks volumes.

To break this stalemate, Weinstein argues, the Toxicoferan supporters need to provide evidence that the things they call venom are actually being used to subdue and kill prey. Sure, says Fry, “just give us a couple more centuries, an army of research students, unlimited funding, and we will be with you on that.”

The problem is that venoms are highly targeted weapons, adapted to take out specific prey. Tree snakes, for example, wield neurotoxins that are up to 300 times more potent against birds (which they hunt) than mammals (which they don't). So, to work out if a potential toxin is actually toxic, you often need to test it against its owner’s natural prey, rather than just lab rodents.

Besides, it's not necessary, says Fry. The Toxicofera idea wasn't just based on a few genes, but also on studies of gland anatomy, feeding behavior, and evolutionary relationships. If that’s not enough, he says, you’d also have to dispute the status of most canonically venomous animals, including spiders, scorpions, cone snails, stonefish, and stingrays. “Anyone who contends that all these lines of corroborating evidence are still not enough seems to lack an understanding of toxinology specifically and science in general,” he adds.

That's a lot of people, it seems. This September, both sides presented their evidence in a public debate, at the International Society of Toxinology World Congress. To Mulley's surprise, he won by a landslide. “We expected to be defeated, or maybe scrape a draw,” he says. “Maybe the tide is turning.”

Juan Calvete, another Toxicofera supporter from the University of Valencia, says this rivalry will only end once more reptile genomes come in. Currently, we only have information from two Toxicoferans—the venomous king cobra and the non-venomous Burmese python. Without more data, it's hard to say which genes truly represent venoms, or how such chemical weapons evolve. “We only have fragmentary information,” he says. “We simply don't know the exact mechanisms that create a toxin gene.”

This might seem like an entirely academic spat, but it's not. For starters, it has implications for scientists who are making anti-venoms against snakebites. “If Toxicofera is true, it means that venom is very complex,” says Mulley. “So an anti-venom would have to counter a huge number of these proteins. But if we're right, for most snakes, you only need to deal with 30 or so. And if we did that, we could develop better anti-venoms with fewer side effects.”

There are clinical implications, too. Weinstein recently read a report suggesting that a patient had been envenomated by a monitor lizard, when all the symptoms clearly pointed to a much deadlier Russell's viper. “Suggesting that this group of lizards is venomous and responsible for bites—and, of course, there's no anti-venom—may mean that patients are denied the right diagnosis and treatment.”

There's the matter of public confusion. “In New York, monitors and iguanas are now impossible to keep privately, partly as a result of all of this,” says Weinstein. “For me and many of my colleagues, our development as herpetologists and clinicians all stemmed from having private collections as kids. So I hate to see responsible private collectors affected by premature and unsupported conclusions.”

No matter who is ultimately right, it's clear that the Toxicofera concept is still up for debate—an uncertainty that's missing from both the mainstream media and academic papers. Partly that's because scientific publishers are notoriously unwilling to publish contradictory papers. “A lot of these Toxicofera papers are in Nature and other high-impact journals, but when we shipped our work to them, they said it was only interesting to reptile venom people,” says Mulley. “But the originals were clearly interesting to everyone!”

The media and popular audiences love scientific plot twists. We’re drawn to underdog ideas that overturn established dogma, but our love for them risks turning them into dogma themselves. “Once this stuff enters the public consciousness, it’s hard to get it out again,” says Mulley.