Just when you thought the picture of feather evolution was getting clearer... enter Tianyulong, the feathered heterodontosaurid.
You heard me. Read about the new find at Ed Young's blog.
[Image: Tianyulong confuciusi, credit Li-Da Xing]
So what does it all mean? I doubt there'll be any consensus, but the authors do a pretty good job of showing that, like the feathers of Sinosauropteryx and Beipiaosaurus, these structures are hollow and not collagen. So were all dinosaurs (or ornithodirans as a whole, dinosaurs + pterosaurs) feathered, and many lost feathers and re-grew scales? The distribution is now making it look like all these structures (ceratopsian quills, ptero-fuzz, feather,s protofeathers and now ornithischian fibres) are homologous--that is, they all derive from a common, ancestral structure. However, I think it's more likely that these started out at bristle-like quills, and evolved into softer, more fuzzy pelts in a few lineages independently, the small, active ones that would be pre-disposed to adapting these inherited quills.
But what do I know? I'll keep you posted on what ideas start getting inevitably tossed around.