
Feathered non-avian dinosaurs are nothing new.
Sinosauropteryx was the first non-avian feathered dinosaur found, in 1996. As a compsognathid,
Sinosauropteryx pushes the origin of avian feathers back near the base of the Coelurosauria, a fact which predicts that certain other theropod lineages will also have feathers. And for the most part, these predictions have come true.
Dilong is usually thought to be a basal tyrannosauroid*,
Beipiaosaurus is a basal therizinosaur,
Caudipteryx and
Protarchaeopteryx are basal oviraptorosaurs,
Jinfengopteryx is a troodontid, and there are a wealth of feathered dromaeosaurs, including
Microraptor and
Sinornithosaurus. I don’t really know how scansoriopterygians shake out, but they had feathers, too. As more fossils were unearthed, it became clear that feathers come in a wide range of shapes and types. Most non-avian theropods had what I call “baby chicken” feathers—simple shafts which mat together and probably had an insulatory function (as they do in baby chickens). Then, there are “tuft” feathers, in which a multitude of filaments arise from a single point. Think of these as feathers without quills. The next step seems to be quilled feathers lacked individual barbules, so the feathers are frayed and fluffy, like the downy feathers of many waterbirds. Finally, you get a true “modern” feather—one with a quill and branches with barbules that slide together to give rigidity the structure. Only oviraptorosaurs and paravians have true feathers, and only paravians have asymmetrical feathers—those used for flight. So feather evolutions seems to have followed a fairly straight path, from insulating structures (basal coelurosaurs) to display structures (oviraptorosaurs, paravians, scansoriopterygians) to aerodynamic assists (paravians).
Two fairly recent finds have cast doubt on this linear progression of feather evolution. First, a few years ago, a Chinese specimen of
Psittacosaurus turned up with absolutely wonderful soft-tissue details preserved, including incredible skin impressions, an “outline” detailing the body’s overall shape, and—most importantly—long, bristly quills erupting from along the dorsal surface of the tail. These quills are circular in cross-section and appear to be hollow. They are not as tightly packed-together as theropod integument, but they are definitely numerous. I was instantly reminded of porcupine quills, although those of the parrot lizard curve gently back. The second find was a new specimen of
Beipiaosaurus published earlier this year. It's "feathers" are roughly similar to the parrot lizard in being long, unbranched "quills," although while those of
Psittacosaurus are hollow, cyndrical shafts,
Beipiaosaurus' quills seem to be flatter. At any rate,
Beipiaosaurus' quills look, to me, like very elongate versions of those on
Sinosauropteryx. 
Just a week ago, we could say with a certain confidence that the quills of
Psittacosaurus and those of
Beipiaosaurus arose convergently. The great phylogenetic distance between the two taxa, with nary a feathered form in between them argued strongly in favor of an independant acquisition of "quills." In fact, they probably had different functions. In
Beipiaosaurus, the feathers thickly cover the entire body and were probably used for thermoregulation. Because
Psittacosaurus' quills were only present on the dorsal side of the tail, it's likely they had a defensive or display role (or both). But all it takes is one fossil to throw everything paleontologists think they know into question, and one of those fossils was published in this week's issue of the journal
Nature.

Behold!
Tianyulong confuciusi, a feathered
heterodontosaur from China. Now, just to remind you all where heterodontosaurs sit on the dinosaur family tree, a recent reappraisal of basal ornithischians by Butler et al. (2007) suggested that heterodontosaurs represent the basalmost formal group of ornithischians (no more "fabrosaurs" here!). So we've got a ridiculously basal ornithischian with feathers, an incredibly derived ornithischian with "quills," and a whole bunch of coelurosaurs with feathers. But then you've got a whole mess of dinosaurs in between with preserved skin impressions showing scaley, reptilian skin. Does this mean the integument of ornithischians is non-homologous to the integument of saurischians? Well, not necessarily. There may be several factors at play here. "Feathery" integument could be basal to all of the Dinosauria, but may be a juvenile character. The peach fuzz is present on hatchlings but lost upon reaching adulthood. In some dinosaur groups, coelurosaurs and heterodontosaurs, for example, that peach fuzz is retained into adulthood (ontogeny). Or perhaps only small dinosaurs had feathers, or at least those below a certain maximum mass. Large-bodied mammals often have extreme short, fine hair or have gone "bald" entirely. Elephants, hippos and rhinos have all convergently developed hairless bodies, probably in response to the hot temperatures they operate at. Even naked, these animals need to cool off in mud or waterholes. Many dinosaurs became rhino or elephant-sized (or much, much larger), so you wouldn't expect to see feathers on an animal the size of
Triceratops or
Sauroposeidon. Maybe decorative patches here and there, but certainly not the extent of
Sinornithosaurus. Ornithischians could've just used feathers in a different way--
Psittacosaurus only retained them on the dorsal surface of the tail, after all, and probably used them for defensive display. The parrot lizard certainly isn't at the theoretical maximum size to still have a full coat of feathers. It's hardly bigger than
Beipiaosaurus, which itself is fully fluffed up.
On the other hand, there's a strong preservation bias against feathers. You only get good evidence of feathers on a slab fossil like
Archaeopteryx or
Tianyulong. You won't find fuzz on a 3D fossil, and that's how larger dinosaurs come out of the ground. Even so, the size assumption does seem to pan out. Skin impressions from a wealth of giant dinosaurs have shown that their skin is pebbled, not feathered. So maybe feathers are basal to all of the Dinosauria, but many groups, including thyreophorans, hadrosaurs, and sauropods secondarily lost their plumage entirely. One thing does seem clear--of all the dinosaur groups where integument of any kind is known, only one small branch of the Theropoda developed
true feathers--the kind that gets you up in the air. If the feathers on
Psittacosaurus,
Tianyulong, and
Baipiaosaurus really are homologous, then feather structure remained relatively conservative across the spectrum, and only got the "next stage" in the theropods.
Tianyulong's authors note the differences (mostly in cross-sectional shape) between the feathers of the heterodontosaur and the therizinosaur, but these are fairly minor quips. Aside from one being circular and the other being flatter, they look startlingly similar. I would be genuinely surprised if "feathery" integument was
not basal to the Dinosauria at this point.
But wait--there's more! Check out those pterosaurs. They all had some kind of fuzz, too. It's usually thought to have arose separately, but what if, just
what if, the integument of
Sordes 'n' Pals is homologous to the integument of
Tianyulong and
Baipiaosaurus? That would mean that all ornithodirans, basically everyone from pterosaurs to dinosaurs, was fluffy to some degree.
Marasuchus,
Silesaurus, nobody's safe. They might've all had feathers.

The fact that all of the Ornithodira could be fluffy has
enormous consequences for paleontology. First of all, it probably means that ornithodirans were endothermic. Ectothermic animals, today anyway, don't have hairy or feathery integument. No hairy lizards or feathery turtles, I'm afraid. Interestingly, the one mammal that has reverted to ectothermy, the naked mole-rat, has
lost its furry coats--without the ability to generate heat internally, it has no need to try and retain that heat, so the hair disappeared. Naked skin seems to equal ectothermy, and fuzzy skin seems to equal endothermy. That doesn't mean dinosaurs that secondarily lose their feathers reverted to ectothermy. Far from it! The larger an animal gets, the more dangerous a coat of hair or feathers becomes. African elephants don't have mammoth coats of hair because if they did, they'd overheat. They have enough problems maintaining their body temperature at a reasonable constant with all that surface area, thank you very much. I imagine the same problem plagued sauropods, hadrosaurs, and ceratopsians. For those guys, a coat of feathers would've been suicide.
But wait--there's more! Suddenly, dinosaurs who live in Alaska, Russia, and Australia have a way to beat the cold! Now granted, temperatures in Cretaceous Alaska never reached the horrible lows we see today, but ice would've formed in the winters, and snow may have fallen. Australia faced much the same situation, and Siberia probably did, too. Alaska and Siberia were farther north, and Australia was farther south. All this time, paleontologists have been suggesting migration or hibernation for high-latitude dinosaurs, but now that fuzz has come along on ornithischians, little
Leaellynasaura suddenly has a way to insulate against those chilly twilight nights. Now, up in Alaska, I'm not saying that larger dinosaurs (like
Edmontosaurus and
Pachyrhinosaurus) did NOT migrate south, but maybe they adapted to the cold in the same way that Pleistocene megafauna did: grow a thick, lusterus coat of fluff in the winter, and shed most of it in the summer. Of course, skin impressions from Alaskan specimens would help, as would publication of the material coming out of the Colville. Our
Pachyrhinosaurus might be a distinct species, but nothing substantial has been published on it.
So there you have it, folks. Feathered heterodontosaurs are ridiculously exciting for a number of reasons. Now, this initial find should be treated with
some caution. It's not uncommon for Chinese fossils to be faked somehow, and I'd like to see more ornithischian families discovered with evidence of feathers before the idea of feathers being plesiomorphic for Dinosauria is really cemented. Still, it's one helluva step forward.
*There is increasing doubt, however, that Dilong and Guanlong are ancestral tyrannosaurs. Instead, they may simply represent forms which fall between Ornitholestes and compsognathids. This would actually be helpful, as it would mean the origins of feathers are pushed to the very base of the Coelurosauria.