The first dinosaurs may have originated in the Northern Hemisphere, possibly in an area that is now Britain.
This is one of the conclusions of the first detailed re-evaluation of the relationships between dinosaurs for 130 years.
It shows that the current theory of how dinosaurs evolved and where they came from may well be wrong.
This major shake-up of dinosaur theory is published in this weeks's edition of the journal Nature.
The reassessment shows that the meat eating beasts, such as Tyrannosaurus rex and Velociraptor, have been wrongly classified in the dinosaur family tree.
One of the implications is that dinosaurs first emerged 15 million years earlier than previously believed.
And the fossil evidence suggests that this origin may have occurred further north than current thinking suggests - possibly in an area that is now the UK, according to the new study's lead author, Matthew Baron of Cambridge University.
"The northern continents certainly played a much bigger role in dinosaur evolution than we previously thought and dinosaurs may have originated in the UK," he told BBC News.
The previous version of the dinosaur family tree was developed 130 years ago by Harry Govier Seeley, a palaeontologist also working at Kings College, London.
By comparing the size, shapes and arrangements of fossilised bones of different species of dinosaurs and how they changed over time, he devised a theory of how they were related and how they evolved.
He concluded that there were two main groups of dinosaurs: those whose hip bones were like those of modern-day birds, which Seeley called Ornithischia, and those whose hip bones were more reptile-like, which he named Saurischia.
The bird-hipped group were all exclusively plant-eaters and included familiar creatures such as Stegosaurus and Triceratops.
The lizard-hipped group had two branches: the plant-eaters, such as Brontosaurus, and the meat-eaters, such as T. rex.
This organisation has been unchallenged until now.
In redrawing the relationships, researchers studied many more bones and included more species, quite a few of which have been discovered only in the past 30 years.
The team's analysis suggests that Seeley got it wrong.
The new approach argues for the meat-eaters, a group known as theropods, to be moved into the same classification as the bird-hipped dinosaurs.
Cambridge's Prof David Norman, who supervised the study, said it represented a major departure from past thinking.
"All the major textbooks covering the topic of the evolution of the vertebrates will now need to be re-written if this suggestion survives academic scrutiny and becomes accepted more widely," he explained.
"It seems that the dinosaur family tree is being shaken quite firmly. It will be interesting to see what drops from its branches in years to come."
The reason that the Northern Hemisphere, and the UK in particular, has become more likely to be the place for the emergence of the first dinosaurs is the fact that two crucial fossils were found in Scotland and England.
For decades they were dismissed as unimportant species, but following the redrawing of the dinosaur tree they are now placed close to its base.
The Scottish and English finds suggest that it is now more likely that the first dinosaurs emerged 245 million years ago in the northern part of the planet on a land mass called Laurasia, rather than 230 million years ago on a more southerly unit called Gondwana.
Matthew Baron said the results came as a "shock".
"A British scientist, Sir Richard Owen, gave the word dinosaur to the world. Now we may be looking at the possibility that the very earliest dinosaurs were roaming an area that has become Britain and the group itself could have originated on these shores."
The researchers involved cautioned, though, that the fossil record for early dinosaurs is so sparse that it would be difficult to make any firm claims at this stage for their origins. But the team hopes that its findings will spur palaeontologists to search for more fossil evidence to back up the new ideas.
A challenge to one of main theories of dinosaur evolution is bound to be controversial.
Prof Hans Sues of the Smithsonian Museum in Washington DC, US, said that the findings had to be tested and corroborated.
"I am sceptical as none of the other recent analyses obtained similar results - but I keep an open mind," he told BBC News.
Prof Paul Barrett of London's Natural History Museum, who was involved in the study and came up with the idea of testing Seeley's old theory, said that the new family tree seemed more logical than the previous one.
"Now we have our evolutionary tree, we can use it as a foundation to understand how dinosaur features evolved over time, and it is already beginning to help us explain some questions that have puzzled us," he added.
Among those questions is the fact that birds are thought to have evolved from meat-eating dinosaurs. But under the old scheme, they were not in the bird-hipped group.
The new tree fits more neatly with the observation that many meat-eating species and bird-hipped dinosaurs had feathers. The fact that previously they were in separate groups led some to speculate that all dinosaurs, including the long-necked sauropods were feathered. But there has been no fossil evidence for this, and it is a suggestion that never really took hold.
The latest work also indicates that dinosaurs evolved into meat-eaters on two separate occasions during their evolution and it even implies that the very first dinosaur was omnivorous and therefore ate both plants and meat.
There was, however, one potentially disastrous consequence of the new scheme.
It could have meant that the long-necked dinosaurs such Brontosaurus and Diplodocus would not strictly speaking be classed as dinosaurs. But anxious not to be known as the people who expelled the Natural History Museum's emblematic Dippy the Diplodocus skeleton from the status of dinosaur, Matt Baron and his fellow researchers carefully reworded the definition.
"I didn't want to make Dippy not a dinosaur. That would have created a lot of upset. They are a very well known group and everyone has recognised them to be dinosaurs. To be truthful, I didn't want to be chased out of every conference I went to for the rest of my career."
Mr Baron's new family tree has similarities to ideas developed by the biologist Thomas Henry Huxley in 1870. He believed, correctly as it turns out, that birds descended from meat-eating dinosaurs and he included them then with the bird-hipped dinosaurs in a group he named Ornithoscelida, or bird-limbed.
At the time Huxley's ideas were roundly dismissed and eclipsed by Seeley's.
As an acknowledgement of Huxley's contribution, the team has revived the name of Ornithoscelida for his new combined group.
As well as being a remarkable piece of research in itself, the work is a vignette of the scientific process itself - how challenging old, well-established ideas with a fresh eye is always worthwhile and can often bring new insights.
"We've proved Huxley right," said Mr Baron. "We didn't pay any attention to any of the dogma of the past 130 years. We tried to incorporate no prior assumption and so we have pulled apart the tree and reassembled it and have come up with solutions to questions that have been troubling scientists for a very long time."
Follow Pallab on Twitter.