Over the past decade so many
It takes a little bit of background knowledge to understand why this is so shocking. Dinosaurs can be divided into two large groups: the saurischia (theropods and sauropods) and the ornithischia (armored dinosaurs, horned dinosaurs, hadrosaurs, etc.). To date nearly all the dinosaurs with feathers have been coelurosaurs, a group of theropods to which birds also belong, but there is one exception.
In 2002 paleontologists announced that they had discovered a specimen of the small ceratopsian dinosaur Psittacosaurus with a plume of bristles jutting from its tail. Since Psittacosaurus was an ornithischian dinosaur it was about as far removed from bird ancestry as a dinosaur could get, so why did have hollow bristly tubes on its tail that resembled early feathers? Careful examination confirmed that the structures were not some strange artifact of preservation like collagen fibers from the decomposing body, but just why this dinosaur had bristles was a puzzle.
Enter Tianyulong. This dinosaur from the Early Cretaceous rock of China was a heterodontosaurid, a group of small dinosaurs placed near the base of the ornithischian family tree. What is so special about the first specimen of Tianyulong, though, is that it exhibits three patches of hair-like structures very similar to the "bristles" on the tail of Psittacosaurus. These structures did not branch like feathers and appear to have been more rigid than the feathery "dinofuzz" of coelurosaurs. The structures possessed by Tianyulong were not feathers, but just what we should call them is now open for suggestion.
The big question is whether the bristles on Tianyulong and Psittacosaurus represent the independant evolution of a body covering among ornithischians or whether the bristles were derived from an earlier body covering shared by the common ancestor of ornithischians and saurichians. The feathers of coelurosaurs and the bristles of the ornithischians differ significantly but it is possible that they represent different derivations from a more ancient kind of body covering. The presence of a feathery or hairy body covering in dinosaurs, then, would have been lost in some groups and retained in other groups. The other alternative is that some ornithischian dinosaurs independently evolved a different sort of body covering, perhaps more than once. Which is the correct hypothesis? More research is required to know for sure, but what Tianyulong and Psittacosaurus show is that dinosaurs expressed a wider range of body coverings than we previously appreciated and hint at more amazing discoveries yet to be made.