Tuesday, October 5, 2010

A Guide To Feather Colors

Above: Restoration of Changchengornis by Matt Martyniuk, all rights reserved. Even when the colors of a prehistoric feathered dinosaur haven't been revealed by studies of feather microstructure, there are ways to infer which colors were and were not likely.

The recent discovery of a fossil penguin from Peru, complete with preserved feathers and melanosomes that reveal their color, prompted me to dive a little deeper into this topic. Keep in mind this is an area I'm still learning about myself, so please feel free to add corrections or keep the discussion going in the comments.

For those of us interested in palaeontography,* the recent work by Jakob Vinther (see original post here) and others on reconstructing the life coloration of prehistoric animals has been some of the most exciting paleontology research of the decade. In the carefree halcyon days before fossil melanosomes were recognized, I and other artists were given free reign over the external appearance of our feathered dinosaurs. But since Vinther's paper, I've been inspired to look into exactly what biological factors go into bird coloration. Needless to say, this is something not a lot of others have probably looked at, as most paleoart follows the old philosophy that when it comes to color, anything goes. Apparently though, this was far from true even before Vinther and his colleagues came along.

* Since this is John Conway's term I reckon I'm stuck using the Queen's preferred spelling, sorta like "pycnofibres". Anyway, that's just a fancy way of saying "paleoart".

There are several processes that add color to the feathers of birds and, presumably, other feathered dinosaurs. At the most basic level, these can be divided into structural color and pigmentation (though sometimes it isn't hat simple, as I'll explain down the page).

Structural Colors
Structural colors come from the actual physical structure of the feather. At the microscopic level, feathers exhibiting structural color have a "foamy" texture of tiny spheres or channels which enclose minute air bubbles. Light scatters through these bubbles in various ways depending on the exact arrangement. The development of these extremely complex structures has recently been covered by Dufresne and colleagues (2009), so you can track down that paper for a technical treatment of structural feather colors.
(Right: Structurally colored blue feathers of a Blue-and-yellow Macaw Ara ararauna, by Jörg Groß, licensed.)

Basically, structural colors can do two things: produce colors not found among the various pigments, and enhance or change pigment colors. For example, among amniotes, there is no such thing as a blue biological pigment. The blue feathers of a bird are produced by scattering due to structural colors. Similarly, iridescence as seen in many birds comes from the feather structure. A bird with bright white or pitch black feathers likely uses structural colors to achieve this effect--without them, these colors would be flatter, duller, and less vivid. Structural coloration can act as a filter, combining with pigments to form new colors. In most birds that have them, green feathers are produced by layering yellow pigmentation nodules over a "blue" underlying structure.

Does it fossilize?: You bet! Iridescent feathers have been reported by Vinther, and it's often apparently to the naked eye alone. There are some stunning examples of iridescent insect fossils out there. Structurally colored feathers have been recognized by a distinct arrangement where a thin layer of densely aligned melanin overlies a looser conglomerate of melanosomes. This can be seen even if the overlying keratin scattering layer has degraded away (Vinther et al. 2008). This kind of structure-via-melanin is also found in the dazzlingly iridescent plumage of hummingbirds (Prum, 2006).

What it means for dinobirds: Blue, green, jet black and bright white can't be present in dinobirds that lack structural color in their feathers. I've said before that structural colors are impossible in the monofilament integument of primitive coelurosaurs. However, I'm not so sure that's true. The main difference between hair and feathers isn't the structure of the filaments, it's the structure of the underlying molecules. Hair is alpha-keratin, a helix-shaped molecule like DNA. beta-keratin, which makes up feathers, has a layered and pleated underlying molecular structure more conducive to structural scattering. So a blue-fuzzed Struthiomimus may be possible. However, in the iridescent fossil feathers studied by Vinther et al. (2008), the structural color was restricted to the barbules, which are not present in many primitive feathered dinosaurs.

Most bird colors are due in whole or in part to pigmentation, or lack thereof. There are several different kinds of pigments, with the two most common being melanins and carotenoids.

Melanins are what all the hubbub is about. Not only are these easily identified in fossil feathers, but their shape and concentration can tell you what color they gave their feathers. Melanins are responsible for black (though not deep, solid black, which require an extra push from structural color), gray, and a wide variety of browns through rufous orange colors. Melanins are the main pigment in mammalian hair, so think of the spectrum of mammal colors when imagining what shades are possible with melanin. A lack of melanin will produce white.

(Right: Different types of melanin in the feather of a Zebra Finch Taeniopygia guttata, from Not Exactly Rocket Science/Zhang et al. 2010).

Does it fossilize?: Of course! I've discussed the relevant melanosome papers in the past, posts linked below.

What it means for dinobirds: For carnivorous dinobirds, these are where the action is. Pure carnivores will usually lack the dietary requirements for carotenoids, so structural colors plus melanin are all they've got (and maybe porphyrins, see below). It seems odd that of the three described prehistoric dinobirds with color, they all seem to have the same color palate. Sinosauropteryx (rufous and white), Anchiornis (gray, black, white, brown, and rufous), and now Inkayacu (gray, white, and rufous-brown). These are all carnivorous/fish eating species, so it makes sense that they don't exhibit any more exciting colors. However, it's also possible that we're missing something: In their 2009 Anchiornis paper, Li and colleagues specifically noted that they didn't test for carotenoids. However, I would imagine that given their prior 2008 paper, they did look for structural color or at least iridescence in fossil feathers.

Carotenoids are, by and large, what give birds their characteristically bright colors. The trick is that carotenoids can't be directly synthesized by the body in animals (some can, but there need to be other types of carotenoids present to convert). Carotenoids come almost exclusive from a diet of plants or, secondarily, of things that sequester a lot of carotenoids in their body tissues (like plant-eating invertebrates and some fish). Gulls living near salmon farms have begun shown hints of pink in their feathers: this is because farm-raised salmon are fed artificial carotenoid sources to make their flesh pink, and these are transferred to the birds. The most unusual source of carotenoids, this time among a carnivorous species, is the Egyptian Vulture Neophron percnopterus, which gets its bright yellow facial skin by eating the droppings of ungulates, dropping which yield no significant nutritional value and appear to be eaten only for the carotenoids (McGraw, 2006)! Indeed, while carnivores aren't usually brightly colored, McGraw noted that there may be selective pressures in some species to add weird things to a diet in order to become more colorful.

(Right: Egyptian Vulture, photo by Dezidor, licensed.)

Does it fossilize? Yes, but it looks the same as melanin, and unlike melanin, you can't tell a carotenoid by its shape. According to Li et al. (2009), special chemical tests would have to be run to determine if a melanosome is really a carotenoid, and what color it was.

What it means for dinobirds: Even though we haven't yet identified carotenoids in fossils, we know that they can only be present in animals that are herbivores or feed on herbivorous insects. Scansoriopterygids, for example, could have been brightly colored by carotenoids, since they presumably ate tree-dwelling arthropods. Alvarezsaurids would probably lack carotenoids if they are mainly termites and other social, non-colorful bugs, as has been suggested in the lit. Jeholornis and Jinfengopteryx, two dinobirds with direct evidence of seed eating, are prime candidates for reds, oranges, and yellows (or even greens, with added structural color). Also, keep in mind that red carotenoids from crustaceans, when eaten by birds with otherwise melanin-free feathers, are what give pink wading birds like flamingos their distinctive colors. This is why many artists restore some ctenochasmatid pterosaurs, especially Pterodaustro, as pink (though how all this applies to pycnofibres is still anyone's guess).

(Right: Speculative restoration of Jeholornis prima with red and yellow carotenoid crown by Matt Martyniuk, licensed.)

Carotenoids are often used by modern birds as a sign of fitness when choosing a mate. Because carotenoids have to be eaten, a bird with a poor diet will be drabber than a bird that is very successful at finding food. A flamingo kept in a zoo will turn white if its diet isn't artificially supplemented with red carotenoids.

Carotenoids can also impact the eye color of a bird, as well as beak color and the color of the scales on its feet... even the yellow yolk of a chicken egg is due to carotenoids (some birds use Flavin for yolk color, see below). Keep in mind that adding orange, yellow or green feathers, or red, orange or yellow beaks, implies your dinosaur is eating a diet containing carotenoids.

Porphyrins are perhaps most famous for lending blood its red color and leaves their green (both heme and chlorophyl are porphyrins), but it can also color feathers, adding browns and reds (and green, but only in the turacoverdin pigments found in Turacos). Interestingly, porphyrins may have a role in temperature regulation. In addition to insulating eggs (see below), they are mainly found in the downy feathers of nocturnal birds like owls, and those that are active in colder temperatures.

The blue of American Robin eggs is created by porphyrins, as is most other egg coloration. In fact, some researchers note a correlation between porphyrin in eggshells and nesting behavior. Pure white eggs are only found in birds which nest in shelter like under foliage, and which constantly attend their eggs. Species which leave the eggs partly exposed to the elements have colorful porphyrin-containing shells, possibly because of the supposed temperature regulating effect. Paleoartists might want to consider this when drawing various dinosaur nests.

Does it fossilize?: I'm guessing no, as we're dealing at the molecular level here. However, I wonder if porphyrins could be detected via chemical analysis, like the one used to detect beta keratin in the feathers of Shuvuuia deserti.

(Right: The juvenile Black-shouldered Kite Elanus axillarus uses porphyrins to achieve a red-brown color not found in adults. Photo by Mdk572, licensed.)

What it means for dinobirds: This one is the big question mark. I've never seen references that describe a method to detect porphyrins in fossils. Luckily, they're mainly only brown and dull red, colors that could conceivably be found with melanin alone. If anything, porphyrins give us license to add some extra reddish splashes to purely carnivorous dinobirds, especially those that may have been active at night or in cold climates, like troodontids.

Uncommon pigments: There are a variety of minor pigments that can color a bird's feathers. Pterins are responsible for the yellow, red, white, and orange colors of some bird eyes (in humans, eye color is controlled by melanin; low melanin results in blue eyes, and some babies eyes darken as the melanin levels increase). Flavin pigments cause many egg yolks to be yellow.

(Right: The feathers of a Yellow-headed Amazon Amazona oratrix. Parrots are so stingy with their carotenoids they had to evolve an entirely new pigment to color their feathers. Photo by Rei, licensed.)

Psittacofulvins are found only in (you guessed it) parrots, and create yellows oranges and reds in place of carotenoids, which parrots have evolved to sequester, possibly for nutritional reasons. There are some undescribed pigments known only in penguins that add florescence to their yellow display feathers.

The take-home message:
When you add color to a feathered dinosaur restoration, you're presenting an implicit hypothesis about its diet, lifestyle, and soft tissue anatomy. when doing serious paleoart, keep these constraints in mind, and use them to make your art more interesting and your science more rigorous. Nobody can tell you not to draw a Utahraptor with a bright red face, but if you're trying to make paleontography and not just a bit of fun, why not depict it munching on a big, red fish or a pile of Sauroposeidon poop?

* Dufresne, Eric R., Heeso Noh, Vinodkumar Saranathan, Simon G. J. Mochrie, Hui Cao and Richard O. Prum (2009). "Self-assembly of amorphous biophotonic nanostructures by phase separation."
Soft Matter, 5: 1792-179
* McGraw, K.J. and Nogare, M.C. (2004). "Carotenoid pigments and the selectivity of psittacofulvin-based coloration systems in parrots." Comparative Biochemistry and Physiology B, 138: 229–233.
* Li, Q., Gao, K.-Q., Vinther, J., Shawkey, M.D., Clarke, J.A., D'Alba, L., Meng, Q., Briggs, D.E.G. and Prum, R.O. (2009). "Plumage color patterns of an extinct dinosaur." Science, 327(5971): 1369 - 1372
* McGraw, K.J. (2006). "Mechanics of Carotenoid-Based Coloration." Pp. 177-242 in Hill, G.E. and McGraw, K.J. (eds.), Bird coloration, Volume I: Mechanisms and Measurements. Harvard University Press.
* McGraw, K.J. (2006). "Mechanics of Uncommon Colors: Pterins, Porphyrins, and Psittacofluvins." Pp. 354-398 in Hill, G.E. and McGraw, K.J. (eds.), Bird coloration, Volume I: Mechanisms and Measurements. Harvard University Press.
* Prum, R.O. (2006). "Anatomy, Physics, and Evolution of Structural Colors." In Hill, G.E. and McGraw, K.J. (eds.), Bird coloration, Volume I: Mechanisms and Measurements. Harvard University Press.
* Vinther, Jakob, Derek E.G. Briggs, Julia Clarke, Gerald Mayr and Richard O. Prum (2008). "
Structural coloration in a fossil feather." Biology Letters, 6(1): 128-131.
* Zhang, F., Kearns, S.L., Orr, P.J., Benton, M.J., Zhou, Z., Johnson, D., Xu, X. and Wang, X. (2010). "Fossilized melanosomes and the colour of Cretaceous dinosaurs and birds." Nature, 463: 1075-1078.


  1. This is really informative! I was just thinking of doing a post on feather colors over at my own blog, and this really helped clear things up for me. I've read some really interesting things on this blog, but this one may top them all!

  2. Cool post. By the way, seems like the term paleontographer already has a different meaning: http://en.wikipedia.org/wiki/Palaeontography

  3. That's interesting, but what about the different colors in reptile scales? Some of those get bright and varied, many turn bright green or blue too, in patterns. Are those the same pigments and effects in scales or is there different chemistry and structure involved?

  4. Didn't mean any disrespect, this is pretty cool stuff especially since I'm just starting to learn how to do paleoart. It's a great article. I'm just wondering how to handle the nonfeathered areas on non-avian dinos.

  5. @robertsloan2
    Based on what I've researched, scales seem to follow the same basic 'rules' as feathers. Blues and greens are due to light refraction rather than pigment, bright reds and yellows are due to carotenoids, etc. This also seems to hold true for naked skin, as in my Egyptian vulture example (above) and even applies to mammals (blue skin in mandrills, etc.). There are exceptions, as some fish and frogs have independently evolved actual blue or green pigments, but this has not been found in other vertebrates as far as I know.

    Anyway, my conclusion here holds true for scaly dinosaurs too. A bright pink Lambeosaurus implies it's eating krill, and a bright yellow face on a T. rex suggests copious coprophagy ;)

    1. I'm not entirely convinced that diet influences color for all species.

  6. I'm a bit late to the party, but as someone fond of relatively conservative dino colours - compared to the clash of bright colours and patterns in some paleoart [Palaeontography? Palaeoillustration...?] that might be exemplified by Luis Rey - I enjoyed and appreciated this post quite a bit. And it's been a great education to my own occassional scribbles and doodles, too.

  7. @ Warren B.
    Thanks! I'm a big fan of Rey's, though I agree with you about his color choices. His work is fantastic artistically but also very highly stylized, and I've never been able to view it as a totally realistic or at least naturalistic portrayal of prehistoric life compared to that of some other artists.

  8. I posted that very late at night, and on second thoughts it seems a little harsh towards Mr. Rey's art. But yes: when the subject of brightly-coloured dinosaurs pops up, I can't help but think of some of his fleshy pink & lime green combinations.

  9. Thanks for posting this! I wrote down notes in my journal to refer to every time. :P

  10. This comment has been removed by the author.

  11. Does this mean that obligate carnivorous dinosaurs can't be green?

    1. Well, it means that most shades of green feathers are very unlikely for carnivores, though iridescent green is still possible. And of course green skin would be possible too.