Not surprisingly, the latest Archaeopteryx study has kicked up quite a stir within the media and scientific realms, considering the iconic status it has attained since discovery some 150 years past. This latest paper by Carney et al. and published in the journal Nature Communications claims to have resolved the plumage colour of a feather possibly from Archaeopteryx, a pretty neat little addition to the reconstruction of this critical species. They use methods employed during previous studies, namely the morphological or structural analysis of melanin-bearing organelles within feathers called melanosomes to infer that Archaeopteryx possessed an entirely black plumage.
There is no doubt that the structures are in fact melanosomes, and no doubt that melanosomes contribute to plumage colour. But the question is, how much do they contribute to the plumage colour..? Well, I don’t know. In fact, no-one knows, at least in extinct avian theropods. The authors don’t discuss this either. It has however been discussed elsewhere in similar studies, albeit only qualitatively and in passing.
One comment is from Zhang et al. 2010: “Melanosomes are lyosome-related organelles of pigment cells in which melanins are stored and are responsible in part for the colours exhibited by modern birds.”
And Li et al. 2010: “Other molecular pigments such as carotenoids and porphyrins also produce plumage colours but are not preserved morphologically, thus we cannot address their possible effects here.”
That seems like a pretty big caveat. It’s like saying if you mix green, yellow and red paint in unknown quantities, you get red every time. Or something similar, I suck at analogies. The point is, if in modern birds, there are other significant structures that dictate or contribute towards plumage colour, does it make sense to try and predict colour when these are absent?
UPDATE: Ryan has been kind enough to clarify this point in a comment below.
Furthermore, the 95% confidence intervals the authors use are practically useless. Look at the ordination provided in Figure 4 (not sure if I can copy it here, so won’t..). These 95% confidence ellipses mean nothing in the slightest, or at least nothing meaningful. What they should have shown is an envelope includes 95% of all points within a sample, so that when you insert data ‘blind’, if it falls within a completely discriminated envelope, you can be 95% certain that it belongs to that group (i.e., 95 times out of 100, a blind data point will be correct). The envelopes shown in figure 4 clearly do not show this (if you don’t have access to the paper, ask me for a copy, or take my word for it). As a result, the points calculated for the particular Archaeopteryx feather analysed could really be grey or black, or maybe brown at a push (the number of colour choices is simply overwhelming..).
So was Archaeopteryx lithographica black?
Probably, probably not.
Carney et al. (2012) New evidence on the colour and nature of the isolated Archaeopteryx feather, Nature, DOI: 10.1038/ncomms1642
Li et al. (2010) Plumage patterns of an extinct dinosaur, Science, 327, 1369-1372
Negro et al. (2009) Porphyrins and pheomelanins contribute to the reddish juvenal plumage of black-shouldered kites, Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology, 153(3), 296-299
Zhang et al. (2010) Fossilized melanosomes and the colour of Cretaceous dinosaurs and birds, Nature, 463, 1075-1078
Jon, to clarify:
1. While we provide evidence that the isolated feather was black with 95% probability, we do not claim that “Archaeopteryx possessed an entirely black plumage.”
2. We discuss melanins vis-à-vis other plumage pigments in our Discussion section: “It is improbable that either of the other major classes of feather pigments, carotenoids and porphyrins, were also present in the Archaeopteryx feather, as they are relatively uncommon (4) and generally do not occur in flight feathers, presumably for structural reasons (23,32). Additionally, even if there were codeposition of another pigment, the black melanin would likely mask its presence (4). Indeed, the melanic residue is ubiquitous throughout the feather vanes, and no banding patterns (15) are visible.”
3. Please refer to our cited references (Clarke et al 2010, Li et al 2010) for further information about the quadratic discriminant analysis we employed.
4. We provide information on the canonical axes in our Methods section: “Canonical axis 1 explained 71.32% of the variation and was most strongly associated with long axis; canonical axis 2 explained 17.23% of the variation and was most strongly associated with short axis.”
Hey Ryan, thanks for commenting and clarifying the above points!
Dean Lomax informed me as well that the feather might not actually be from Archaeopteryx too; I should have said that it was the media portraying the entirely black reproduction. That was more what I was frustrated with (will they ever actually get it right??)
Is it really to 95% probability though? If you construct a ‘confusion matrix’ where the percentage of specimens that are correctly and incorrectly assigned to their pre-defined groups is shown, I don’t think you’ll get a 95% success rate. You can see it qualitatively just by drawing envelopes that encapsulate 95% of the specimen points in discriminant space.
Cheers for the clarification on the pigment types – my fault entirely for missing that bit! Is it strictly just melanosomes that predict colour of pennaceous feathers/flight feathers in extant birds then?
Yes, the Archaeopteryx data really predicts black to 95% probability. What you’re referring to is the model’s cross-validation accuracy, an entirely different statistic which as we state in our Methods is 80%.
The coloration of flight feathers in extant birds is almost exclusively due to melanins. One notable exception to this is the unique metalloporphyrins in the family Musophagidae – although even in this case the pigments’ presence has been found to be masked by black melanin when there is codeposition (McGraw & Hill 2006).
Cool, thanks for the clarification on the stats – must have missed that bit too
Might have been a bit pissed off seeing all those images of a completely black Archaeopteryx popping up over the web, and it’s clearly made me make a few mistakes here in the assessment!
Cheers again for all the comments – look forward to seeing more coloured dinosaurs in the future!
And was it an Archaeopteryx feather?