Tyrannosaurus rex has a certain reputation to maintain.

It pierced and gripped the flesh of its prey with long, dagger-like teeth, could rip through 230kg (500lbs) of meat in one bite, and, despite its enormous size, is thought to have been not much slower than 100m world recorder holder Usain Bolt over the same distance. In short, it was pretty much the most effective killing machine ever to have stomped the Earth.

It’s an image that may now need a tweak, if the conclusions of a new study are to be believed.

I think this extends the spectrum of behaviours we can attribute to T. rex and possibly other dinosaurs

Bruce Rothschild, a palaeopathologist at the University of Kansas, US, who examined tooth marks made by T. rex on the bones from its prey, believes the species sometimes took time out from bloodthirsty chasing, shredding, crunching and swallowing to engage a playful side.

His analysis suggests T. rex and possibly other dinosaurs played – or engaged in behaviours for no obvious purpose other than recreation - in the same way that some modern reptiles and birds do.

Rothschild began by noting differences in previous reports of tooth-marked dinosaur bones found at feeding sites and those found on their own.

At feeding sites they were more likely to be from body parts with high meat or bone marrow content. They also exhibited similarities to the remains left by contemporary predators when eating, including skeleton dismemberment, splintered bones, spiral fractures and deep tooth marks.

In contrast tooth-marked bones found in isolation were more likely to be from body parts of little nutritional value, and the markings did not correspond to those associated with feeding.

Rothschild was especially interested in markings on ball-shaped bones called occipital condyles, which are situated at the back of the lower skull and have no marrow content.

It’s plausible but there are many other explanations for how those bite marks got there, including Tyrannosaurus ripping the heads off its prey as it was taking it apart to feed

He re-analysed a previous study of markings on eight of these from Ceratopsids, a family of dinosaurs with facial horns and neck frills including the Triceratops, and examined three others from the collections of the Royal Tyrrell Museum, in Alberta, Canada, and of the University of Kansas.

T. rex, which became extinct around 65 million years ago, is believed to be responsible for the teeth marks on these bones because there is no fossil evidence of any other predatory species that were large enough, living at the right time, and in the places the bones were found.

Rothschild said differences in the types of bones and the marks on them provided important clues to the behaviours of the dinosaurs that made them.

“If I’m eating a piece of chicken, I avoid the cartilage at the end of a bone because it’s not very tasty, and if you look at the bones found at feeding sites you see the same thing,” he said.

“But if you look at isolated bones, especially these condyles, the patterns and characteristics are completely different, and the bones are not broken up, meaning we’re looking at different behaviours.”

Rothschild then adopted the deductive reasoning approach favoured by the fictional detective Sherlock Holmes, who famously said: “When you have eliminated all which is impossible, then whatever remains, no matter how improbable, must be the truth.”

With decay and weathering already having been ruled out as the cause of the markings by previous studies, Rothschild concluded, in a paper published in the journal Ethnology, Ecology and Evolution, that: “Deductive reasoning leaves a single conclusion: the activity resulting in the tooth marks on the bones discussed herein fulfils the definition of play.”

It is believed to be the first time a researcher has proposed the idea that dinosaurs liked to play.

Such behaviour has been identified in modern reptiles including the Orinoco crocodile, several species of monitor lizards and Komodo dragons, as well as in some birds, especially birds of prey.

“I think this extends the spectrum of behaviours we can attribute to T. rex and possibly other dinosaurs,” said Rothschild. “We see playful behaviours in many animals including our pets, so it’s certainly possible.”

However Steve Brusatte, a palaeontologist at the University of Edinburgh, UK, said the study did not provide enough evidence to back up the claim.

“This is nowhere near conclusive evidence. It’s plausible but there are many other explanations for how those bite marks got there, including Tyrannosaurus ripping the heads off its prey as it was taking it apart to feed.”