Wednesday, August 26, 2009

The Colour and the Shape

Above: Iridescent Peacock feather, from Flickr. A new study shows that patterns, iridescence and even color can be preserved in fossil feathers.

Every dinosaur picture book aimed at kids comes with a disclaimer /slash/ incentive: "We don't really know what colors dinosaurs were." They were often depicted as green and drab, camouflage suited to their 1930s-era stint as lethargic reptilian swamp dwellers. But, the kid's books tantalizingly continue, "they could have been any color, with any pattern, even bright fuscha with purple polka dots!" (I'm guessing these books are to blame for Barney...).

Well, that's mostly true. We now know dinosaurs are more closely related to birds than modern reptiles (and lets not sell those short, many are very brightly colored). Birds have excellent color vision, and often employ bright colors and striking patterns to attract mates. Some artists have taken this concept to the extreme: See the almost day-glo colors employed by Luis Rey. All of this falls within artistic license, though even this bastion of dinosaur mysteries may disappear for some species. Actually, color can be preserved in fossils, and several recent studies have applied this to fossilized feathers, like those found on some dinosaurs.

Above: One of the iridescent feathers studied by Prum and colleagues in the new paper, discussed below. From National Geographic.

Fossils preserving life body patterns or even color are not new. Many fossilized fish and insects have preserved patterning or even hints of iridescence in their wings, shells, and scales. At least one ammonite fossil is even said to preserve the original, blood red coloration, possibly indicating a deep-sea habitat due to similar color in modern animals from that environment.

A new technique has been developed to help determine if color patterns are real or due to preservation. A 2008 paper by Jakob Vinther and colleagues described looking at apparently patterned fossil feathers under a microscope, looking for (and finding) melanosomes, the remnants of pigment present in life. The traditional view held that most fossil feathers which are preserved as dark 'stains' in a halo around the fossil are caused by bacteria, which cover the feathers as they decay. Vinther and colleagues found fault with that interpretation, though. For example, some fossil feathers show both dark and light bands in regular patterns. Why would bacteria only be present in such regular segments of fossil feathers? Vinther reckoned that rather than bacteria, the small nodules associated with these bands and dark spots (seen under a microscope) were pigment cells, and represented the actual dark/light patterning (though not the specific color) that would have been present in the living animal. Vinther and colleagues concluded that in fact almost all fossil feathers are preserved this way.

Above: Figure 1 from Vinther et al. 2008, showing melanosomes preserved in a banded fossil feather.

That's not to say that feathers lacking a dark carbon film lacked melanin in life--further examination is needed. For example, most specimens of Archaeopteryx do not preserve dark, stained feather impressions. Does this mean the feathers were white? No--in fact, close examination shows impressions where melanosomes would have been, but have since been lost or decayed away.

Still, this could have implications for the famous feathered dinosaur fossils from China. Many of these do preserve feathers as dark carbon stains (presumably caused in part by melanin), and do sometimes show a banded pattern. Such banding can be seen in the tail feathers of Caudipteryx. The "ring tailed" appearance of the holotype Sinosauropteryx tail is probably real, given an unpublished study by Nick Longrich and confirmation (via Dinoforum, of course) that the bands match up across both fossil slabs and so can't be an artifact of splitting the rock as originally suggested. Longrich also suggested that the apparent absence of feathers on the underside of Sinosauropteryx doesn't mean they weren't there, but rather that these feathers were white and did not leave a stain in the rock. Whether or not these feather patterns contain melanin is still an open question that, to my knowledge, nobody has really studied.
Above: Sinosauropteryx may have been counter-shaded, with a white belly and ringed tail. Image from a work in progress digital painting by Matt Martyniuk, all rights reserved.

So, we might actually know what kinds of patterns a few dinosaurs had in life, thanks to preservation of feathers and melanin. But what about actual colors? Well, a new study of melanin in fossil feathers (this time from the Eocene Messel deposits of Germany, home of Darwinius) has found that even traces of iridescence and life color can be recovered. In the future, more detailed study of dinosaur feathers might reveal something similar. A few barely-published specimens of Microraptor do have a very beautiful blue sheen to them (see photo above). Preservation artifact, or dromaeosaurid Blue Jays?

Luckily, formal evaluation of the feathered dinosaurs using this technique is in the pipeline. As co-author of the study Richard Prum told NatGeo, "We are eagerly hoping to be able to work on some of the Chinese dinosaur feathers to try to reconstruct the colors of the feathered dinosaurs." So, the days of drawing these dinosaurs with whatever day-glo magenta stripes you like may be over.
Above: Tail feathers of Microraptor specimen TNP0099624, from figure 2 in Xu et al. 2003. Could these show the color in life or iridescence? Future studies will find out.

And does anybody else remember that very old bit of goss from the DML about a Velociraptor specimen with associated feather proteins that could reveal the color...?


  1. Nice sum-up -- looking forward to the results.

  2. Cool stuff. There's also a paleocene turtle that preserves some original colorisation on it's carapace. I believe it was specimen of Neurankylus.

  3. Amazing! Even the mystery of (prehistoric) dinosaur coloration (or at least that of a few species) is beginning to be solved.