Dromaeosaurs with Feathers: Some, or All?

Since fossils of feathered dinosaurs were first announced in the late 1990s, there has been a peculiar resistance to their implications for the appearance of previously known groups. Dromaeosaurs like Deinonychus and Velociraptor, convincingly restored as featherless animals in the 1993 movie Jurassic Park and its sequels, could require a completely new image. What generalizations can really be made from the feathered dinosaur specimens that are known? Some amatuer dinosaurology spectators have expressed the opinion that while some small dromaeosaurs in some environments apparently were feathered, this should not imply that others were similarly covered. Feathers have not been found in association with Deinonychus, so surely restorations of this dinosaur as feathered are still highly speculative, or even wrong, according to advocates of featherless dromaeosaurs. The lack of fossil feathers for Deinonychus is actually unremarkable, since bare skin and scales have also never been found. But is there any justification for restoring Deinonychus and other "raptors" with feathers, or without?

There is no need for a highly complicated theory when a simple one will do. Without having any evidence to the contrary, scientists prefer to theorize that the origin of feathers occurred only once in the history of dinosaurs. The last common ancestor of any two feathered dinosaurs should be a feathered dinosaur itself, unless feathers evolved convergently in each descendant line. Among the known feathered dinosaurs, Sinosauropteryx (the first genus found) still has the distinction of having the most remote common ancestry with birds, which the remaining feathered dinosaurs are more closely related to. The last common ancestor of birds and Sinosauropteryx, though not yet identified in the fossil record, was presumably feathered, and every known feathered dinosaur can trace their ancestry back to it. But some popular dinosaurs not yet known to be feathered are also descendants of this feathered ancestor.

For example, the feathered predator Sinornithosaurus has numerous dromaeosaurid features showing that it had a more recent common ancestor with fellow dromaeosaurid Deinonychus than with other unambiguously feathered dinosaurs like Caudipteryx. The last common ancestor of Sinornithosaurus and Caudipteryx would have branched into a group leading to Caudipteryx on one side, and a different group leading to both Sinornithosaurus and Deinonychus on the other side. All of these ancestral populations would have been feathered, assuming feathers did not evolve independently in Sinornithosaurus and Caudipteryx. Obviously, if Deinonychus evolved from feathered dinosaurs, the simplest inference is that it too was feathered.

Defenders of the old image of Deinonychus might ask what factor (insulation, display, flight?) would cause the presence of feathers to be selected for in this dinosaur, but feathers were already standard equipment for dromaeosaurs long before Deinonychus arrived. The truth is that to argue for featherless Deinonychus, one should have a good explanation of why natural selection would favour the absence of feathers. Inability to fly has never led to a total loss of feathers in modern birds, and I see no temptation among artists to restore extinct ground birds like Gastornis with no feathers! Nor is Deinonychus likely to have lost all of its feathers to prevent overheating as a result of its size, because it was less massive than the biggest living birds like the ostrich. Without any living analogy, a scenario that would produce featherless Deinonychus must require a good deal of creative thinking. And even the best story will still be speculation, not proof of feather loss.

Statements of common ancestry among dinosaurs can only be theoretical, and could turn out to be wrong. Birds may be more closely related to Sinornithosaurus, or to Caudipteryx. However, feathers have been found in a wide enough selection of theropods that I am confident to say Deinonychus and Velociraptor evolved from a feathered dinosaur of some sort. To suggest that Deinonychus never had feathered ancestors, one should be able to cite at least one derived anatomical trait uniting Sinosauropteryx, Scansoriopteryx, Yixianosaurus, Beipiaosaurus, Caudipteryx, Protarchaeopteryx, Archaeopteryx, Sinornithosaurus, Cryptovolans and Microraptor that definitely does not occur in Deinonychus or more primitive theropods. To really mean much, the number of traits supporting this arrangement should outweigh the number of traits supporting the current consensus on theropod interrelationships. This approach is likely impossible, so it's not surprising that no one has used it to justify featherless velociraptorines.

Alternatively, one could suppose that feathers actually did evolve more than once in the small theropods. This is an unlikely scenario, though not impossible. If Deinonychus and Velociraptor are to have never had a feathered ancestor, feathers probably evolved about five times, and possibly more (if we keep the current morphology-based ideas about theropod phylogeny). The problem with this is that for feathers to have evolved so many different times, there must have been a very strong selective pressure favouring the development of feathers in small flightless theropods---the opposite of what is usually claimed when arguing for featherless velociraptorines! It is difficult to imagine a reason why feathers would be selected for in theropods so dissimilar ecologically as Sinosauropteryx, Scansoriopteryx and Beipiaosaurus, yet not even possibly in Velociraptor. If it is possible that feathers evolved independently in each known group of feathered dinosaurs, it must also be possible that feathers evolved independently in some other dinosaurs for which the outer covering is not yet known.

Both feathered and featherless restorations of most dromaeosaurs are technically possible, and a decade ago the absence of evidence either way permitted both to be acceptable. However, the new evidence from a variety of feathered dinosaur fossils has shown the featherless model to be exceedingly less probable. In short, to support the image of featherless dromaeosaurs one must first disregard some conclusions from paleontology that are logical and obvious. Secondly, these must be replaced by other conclusions that seem highly unlikely and somewhat ridiculous. As feathered dromaeosaurs are both visually appealing and better supported by science, there is presently no reason to restore them otherwise.

Back