A US study has bruised the fearsome reputation of a popular dinosaur.
Coelophysis, a carnivore that lived more than 200 million years ago, has often been presented in books and museum exhibits as a cannibal.
The view is based on Coelophysis fossils that have preserved stomach contents interpreted as being the chewed up remains of its own kind.
But now a re-examination has suggested those contents may be crocodile, a Royal Society journal reports.
Indeed, one purported meal could even be an accident of geology - one Coelophysis simply dying on top of another, giving the impression the underlying animal's remains were inside the top creature's gut.
"Ideas always need testing; they've been around for a while and we've got to take a close look at how we think these animals really lived," said Sterling Nesbitt from the American Museum of Natural History (AMNH), New York; and Columbia University.
Nesbitt and colleagues report their findings of a re-examination of Coelophysis fossils in Biology Letters.
The creature's cannibalistic reputation stems from the classic find in 1947 of literally hundreds of skeletons of the species Coelophysis bauri at a site near Ghost Ranch in north-central New Mexico, US.
A whole group of animals had died en masse in some catastrophe. Two of the more complete specimens when examined revealed collections of bones inside their body cavities.
The clear inference was that these bones represented last meals - and the gruesome assessment at the time was that this dino diet consisted of other Coelophysis.
But Mr Nesbitt's speciality is early crocodile-like animals, and he says one of the specimens contains tell-tale bones that puncture the cannibal legend.
"There are a few bones in there which we call diagnostic. One is a femur. It's what fits into the hip socket; and in dinosaurs and crocodiles, it does it in completely different ways.
"Then there's part of the pelvis. All dinosaurs have what we call an open-hip socket and this specimen has a closed-hip socket, just like a crocodilian-hipped animal."
The second specimen, Nesbitt's team contends, does not even show gut contents. The researchers say their new analysis demonstrates the way two animals have been fossilised, one on top of the other, gives the illusion that one ate the other.
Mr Nesbitt believes his team's findings put a big question mark against the popular image of Coelophysis - all the books, TV programmes and museum displays may have to change their content.
"It's not completely outrageous to say these guys were cannibals, it's just the evidence to say that they were, is no longer there now.
"The Coelophysis skeletons were the major piece of evidence we have that dinosaurs were cannibalistic. There are examples of animals that have bite marks on them from the same species but you can never be sure if that was cannibalism or just scavenging."
This type of reassessment is nothing new in palaeontology. As new finds are made and old specimens looked at again, ideas have to be updated.
The most famous dino of them all, T. rex, is itself the centre of a robust debate about its lifestyle. Many scientists believe its terrifying reputation has also been overstated; it may have been more of a scavenger than a predator, these researchers argue.
An animatronic Coelophysis - complete with dino dinner hanging from its mouth - can be seen at London's Natural History Museum, in its Dino Jaws exhibition.
The special presentation runs until April next year.