How do you go about hunting for unicorns? More specifically, let’s say that you have found your unicorn, but then what? How would you know that it was a unicorn, rather than something else? How diagnostic, say, is that horn? Might there not be other creatures, as yet unimagined, that have unicorn-like horns but differ from unicorns (and everything else) in all kinds of unusual ways? Narwhals, for example, exist — and their unicorn-like horn is the only feature they have in common with unicorns. So what else might lurk out there in the great unknown? You know that unicorns don’t number among the domestic beasts, says Jorge Luis Borges in his essay ‘Kafka and his precursors’, but, without seeing one, how do you know what a unicorn is “like”?

Palaeontology is always about hunting for unicorns. Because fossils sample diversity from the past that might not exist today, there is always a chance that one will dig up something that defies categorization. Such is the case for the tiny dinosaur described online in Nature by Xing Xu and colleagues (see go.nature.com/jsxjxv), and discussed further by Kevin Padian (go.nature.com/s6g2aw).

The dinosaur is about as far from Brontosaurus and other behemoths familiar to the museum visitor as might be imagined. It belongs to a little-known group of creatures whose body size has hither­to scaled inversely with the length of their names. It is only the third known example of a (deep breath) scansoriopterygid, the other two bearing the unwieldy handles Epidendrosaurus and Epidexipteryx. Xu and colleagues buck the trend by calling their creature Yi. Or, in full, Yi qi. This must be the shortest dinosaur name ever, and it is commensurate with the tiny size of scansoriopterygids, which could have looked thrushes or starlings in the eye. Although scansoriopterygids had feathers and cluster phylogenetically round the ancestry of birds and other feathered dinosaurs, they do not seem to have had flight feathers on their disproportionately long forelimbs. Reconstructions make them out to be rather like feathered lemurs, scampering along branches, perhaps wheedling insects out of crevices with their long, clawed fingers.

Except that Yi qi is different. Attached to each wrist is a strut, made of bone or calcified cartilage, which cannot simply be homologized with regular wrist or hand bones. The strut seems to be a new structure made from an accessory wrist bone, possibly a sesamoid — the kind of bone usually embedded in a tendon or muscle. The ‘thumb’ of the giant panda is made of just such a bone. The sesamoid of Yi qi (if that’s what it is) is much larger, however, in relation to the animal as a whole — equal in length to the bones of the forearm. It had to be there for a reason, but what was it?

It is here that we enter unicorn territory — for no dinosaur, however unusual, has been found with anything like this feature. The authors are appropriately cautious, therefore, in their interpretation. They point to the hint of a suggestion that some soft tissue, preserved alongside these curious elements, represents what might have been a membrane that the sesamoid bone supported. From that, they suggest that Yi qi had membranous wings and might have glided from branch to branch, in much the same way as various tree-living mammals and reptiles do today. But it was probably not capable of powered flight as birds and bats are — and as were, presumably, the extinct pterosaurs, which were (one must stress) only distant relatives of dinosaurs and birds.

When, in the mid-1990s, the first dinosaurs with preserved feathers came to light, there was a great deal of celebration. However, evidence for the bird-like nature of dinosaurs had been accumulating for a while, so for many (though not all) people, feathered dinosaurs were a vindication rather than a challenge. Yi qi, by contrast, is something else. Here we have a feathered dinosaur and a close relative of birds that seems to have essayed an entirely different experiment in aerial locomotion. For a feathered dinosaur to have traded feathers for a membrane in an aerofoil is something nobody could have predicted. Whether or not it is a unicorn has yet to be determined.