The evolution of feathers
The rise and rise of one of evolution’s niftiest inventions.
If you could step back to the Late Jurassic, 160 million years ago, and conceal yourself in the prehistoric foliage of Mongolia, you’d see something remarkable. Between the tree ferns and cycads, an unusual-looking bird would appear. It tidies up a clearing – removing fronds, sticks and other debris. Then, with a dramatic flourish, the pigeon-sized creature stands on tiptoe, puffs up its strikingly coloured plumage, and starts to dance jerkily from side to side, all the while producing clicks and shrill little calls. Most conspicuous are its four long tail feathers, which flick and waft as it shimmies to an internal beat.
This is actually a courtship ritual, very much like the kind played out by birds of paradise today in New Guinea. But on closer inspection, the performer isn’t a bird at all. It doesn’t have wings, but lightly feathered forelimbs with sharp little claws; and instead of a beak, it has a full set of pointy teeth. What we are spying on is actually a small dinosaur named Epidexipteryx, Greek for ‘display feather’. A delicately preserved fossil of Epidexipteryx hui, featuring impressions of four 20-centimetre-long, ribbon-like feathers, was unearthed in Inner Mongolia in 2007 and described in Nature the following year. It was the first clue that feathers found a use in display long before they ever helped a creature become airborne. The scenario described above is fanciful, but the paper’s authors, from the Chinese Academy of Sciences in Beijing, are convinced the feathers were used for seduction. ‘Ornamental plumage is used to send signals essential to a wide range of avian behaviour patterns, particularly relating to courtship’, they wrote. ‘It is highly probable that the [tail feathers] of Epidexipteryx similarly had display as their primary function.’
There is now good evidence that many carnivorous theropod dinosaurs, even fearsome and well-known species – such as Allosaurus and Tyrannosaurus – had feathers, and that they used them for a variety of functions. ‘The most startling revelation about Velociraptor and its kin is that many are now known to have possessed feathers’, writes John Long:
This fact has made us think again not only about the transition to birds, but also about how they might have used their feathers. Did they use feathers in complex mating rituals? Did they use them to brood their young? Or did feathers act primarily as a stepping-stone in the evolution of flight? We know from fossil evidence that some of these scenarios, and possibly all of them, were true.
Feathers are so intimately entwined in our minds with flight that this idea takes some getting used to. Nevertheless, animals with flight feathers can’t have appeared from nowhere, so it makes sense that the earliest feathers had another purpose entirely.
The first feathers
Sinosauropteryx was the first dinosaur to be discovered with feathers, in 1996, but since then a great flock of feathered species has flapped or scurried to the fore. The fossils of about 40 species have so far been found either with feather impressions, or with circumstantial evidence in the form of either quill knobs (pits in the bones where the ligaments of feathers attach) or a pygostyle (the shortened tail structure to which a fan of feathers attaches). Nearly all of these feathered dinosaurs are carnivorous theropods, and the vast majority of the new finds come from the Liaoning region of north-eastern China, although some hail from Mongolia and a smattering of finds have been made in Europe, North America and Madagascar (see ‘An A–Z of feathered dinosaurs’ for a full list).
In the years following the discovery of Sinosauropteryx, a handful of other feathered species were found, none of which appeared to have been able to fly – they didn’t have fully formed wings or the wings weren’t right kind of shape to provide lift. It started to become clear to palaeontologists that feathers may have evolved for another purpose entirely and were only later co-opted for flight.
‘The protofeathery fringe of Sinosauropteryx, now known to extend to its flanks as well as along its midline, was obviously not made for flight, and it is questionable whether it could have served any kind of function in insulation’, mused Kevin Padian in Nature in 1998:
Camouflage, display and species recognition come to mind as other possibilities. Caudipteryx and Protarchaeopteryx go it one better, evolving long feathers with a central rachis (shaft). Were these feathers airworthy? Their vanes are symmetrical and very even, suggesting interlocking barbs, although most flying birds have asymmetrical feathers. However, the arms of Caudipteryx are only 60 per cent as long as the legs … Evidently, the arms and feathers were not large enough for flight.
The feathers of many of these animals were simpler in structure – more what the experts endearingly call ‘dinofuzz’ than anything we’d recognise as feathers today. The spread of these species across the family tree does suggest, however, that having feathers was a common trait in the theropod dinosaurs, even among the many species for which no feathers have been recorded.
Despite the fact that feathers were undoubtedly widespread, it took a long time to find dinosaur fossils that preserved any trace of them, since 90 per cent of fossil sites preserve bones alone and no so-called soft tissues such as skin, muscles, internal organs or feathers. For a long time, the only dinosaur-era fossils with feathers were a handful of specimens of Archaeopteryx, but as John Long points out, in recent years ‘a whole series of great fossils have been discovered of both feathered dinosaurs from a variety of theropod families, right through to many examples of primitive birds. The overall body of evidence from all different directions – growth rates, physiology, bone structure, feathers – points to these dinosaurs being the close relatives of modern birds’.
It’s very likely that the earliest feathers didn’t function in flight or locomotion, but instead were used like the downy fuzz of chicks for insulation or otherwise for display. ‘To start with, feather structures are not all that complicated – they are a coat of simple filaments’, says Paul Barrett. ‘These animals are small and quite active, they have elevated metabolic rates compared to reptiles, and this is a way of retaining heat in small animals.’ The next group of dinosaurs that play with more feather-like structures tend to have big pennaceous feathers (that is, feathers the typical modern shape, with a central vane running the length and interlocking barbs running off to either side), which are more obviously used for showing off. These might be a tail fan or a bunch of feathers on each arm they could have waved at one another or fanned out across the nest to insulate their young.
Some unusual research into the tail muscles of a bunch of oviraptorid dinosaurs, which were common in Mongolia and China during the Cretaceous, provides perhaps the best evidence yet that dinosaurs used their feathers for elaborate displays. Oviraptorids are a group of parrot-beaked omnivorous theropods, including the Gobi Desert animals that have been found fanned out protectively over their clutches of eggs. Despite not being in the part of the theropod family tree most directly related to birds, they have such a perplexing number of very bird-like features that some people have even suggested they were birds that had lost the ability to fly rather than bird-like dinosaurs. These dinosaurs have a pygostyle, with the final few vertebrae fused to form a ridged, blade-like structure.
A 2013 study by experts including Phil Currie and Scott Persons at the University of Alberta, and Mark Norell at the AMNH revealed something completely new about the display purpose of this structure. In the fossils of five species of oviraptorid – including Similicaudipteryx and Nomingia – the team found marks on the bones suggestive of a tight group of large muscles that would have allowed the stumpy tail to be flexed and posed in a number of ways. The researchers had found evidence that these dinosaurs had flexible tails they could flick, waft and shake as desired. They argued that the males of these species likely indulged in tail- shaking mating displays to attract females, much as turkeys and peacocks do today.
The fossil of Similicaudipteryx in particular has structures on the vertebrae of the pygostyle showing feather attachment and suggestive of a tail fan. The other oviraptorids they studied don’t have direct evidence of tail feathers, but as they were close relatives of Similicaudipteryx and evolved later, they are likely to have shared the same features. The vertebrae at the base of the tails were short and numerous, which would have made them highly flexible. These animals also had large muscles with many connection points extending along the tail, which could have been used literally to (in the words of the famous Motown track) ‘shake a tail feather’ – allowing them to flick the tail vigorously both vertically and horizontally. Epidexipteryx, which we talked about at the start of this chapter, isn’t in the same group of dinosaurs, but there’s no reason why it couldn’t have used its four long tail filaments (see image section) for much the same kind of mating display.
Moulting
Another benefit of using feathers for display is that, through moulting, animals can rapidly perform a costume change, switching in and out of brightly coloured mating plumage (in the manner of Australia’s splendid fairy wrens, the males of which turn a brilliant blue during the breeding season). Other birds, such as those that live in snowy winter environments, can change their plumage for the purposes of camouflage.
Studies of 125-million-year-old Similicaudipteryx fossils have also been instructive in this regard; research by Xu Xing and Zheng Xiaoting, reported in Nature in 2010, shows a difference in plumage between adults and young. Feathers on the tail and wing of the adult fossil specimens they studied appear to be standard pennaceous quills, but the feathers on the fossil of a juvenile Similicaudipteryx show something never seen before in a living bird: while the feather tip has the typical shaft and barbules, the base of the feathers is reduced to a flat, ribbon-like shaft with no barbules.
‘This baby dinosaur has bizarre flight feathers, which are strikingly different from those of adults’, Xu told Nature. For the first time researchers had shown that young dinosaurs had a different kind of plumage from adults, which means that moulting was used to change their appearance. In modern birds, babies would normally simply switch from a downy covering of fluffy feathers to adult feathers without this extra step in between, so something unusual had been discovered.
Not everybody was convinced, though. Ornithologist Richard Prum, an authority on feathers at Yale University, argued that the fossil may simply have captured the feathers in the process of emerging from their ‘feather sheaths’, which happens during moulting. The jury is still out on that, but the research did show for the first time that dinosaurs underwent plumage changes through moulting.
Important research of this type, which requires many specimens of the same species, has been made possible by the incredible fossil collection built by Zheng Xiaoting, whose Shandong Tianyu Museum of Natural History houses the largest assortment of complete dinosaur fossils – and by far the largest array of feathered dinosaurs and early birds – anywhere in the world.
Across an ocean and a continent, remarkable Canadian fossils, discovered by the University of Calgary’s Darla Zelenitsky and her team, provided the second piece of evidence for moulting and plumage changes in dinosaurs. Detailed in Science in 2012, these fossils of the speedy, ostrich-like Ornithomimus from the badlands of Alberta were also the first feathered specimens ever found in North America, and proved that ornithomimid (‘bird-mimic’) dinosaurs sported a comprehensive plumage.
In the Jurassic Park movie, a herd of ostrich-like, speedy Ornithomimus were shown escaping from a marauding T. rex. Zelenitsky’s work has shown that these animals should have been depicted with feathers rather than the naked, scaly covering they had in the film. The fossils of an adult and a juvenile Ornithomimus revealed fluffy filament-like feathers in the juvenile, but long pennaceous feathers on the forelimbs of the adult that wouldn’t have been any good for flight. Instead, the researchers think they were used for mating displays similar to the tail fans of the oviraptorids. ‘This is a really exciting discovery as it represents the first feathered dinosaur specimens found in the Western Hemisphere’, Zelenitsky told reporters. ‘This dinosaur was covered in down-like feathers throughout life, but only older individuals developed larger feathers on the arms, forming wing-like structures … This pattern differs from that seen in birds, where the wings generally develop very young, soon after hatching.’
So there is good evidence of feathers being used initially for insulation and display, but how did they come to be co-opted for flight? Eventually, the extra surface area of feathers on the tail and forearms used for display would have offered some lift when jumping or gliding. Then evolution would have started to select for the running or flying functions of feathers rather than simply keeping warm or showing off. (For more on the origins of flight, see chapter 7.)
Feathered tyrants
Most of the known feathered dinosaurs are close relatives of birds in a carnivorous group known as the maniraptors (‘hand snatchers’), which includes dinosaurs such as the dromaeosaurs and troodontids, therizinosaurs and oviraptorids. Nearly all the rest are within in the coelurosaurs, a wider grouping that also includes ornithomimids, tyrannosaurs and compsognathids (see ‘Relationships of the theropod dinosaurs’). Although feathers have only been found in a smattering of species across the whole group, the fact that some of them are early and ancestral members (Sinosauropteryx, for example), and that known feathered species occur on nearly every branch of the family tree, is strong circumstantial evidence that all members of the group were feathered. The reason that, for the most part, we’ve only seen feathers on the species from China is that the level of preservation of fossils in Liaoning is exceptional, retaining feathers, skin and internal organs that rarely leave any trace in fossils from other parts of the world.
Until recently, the general consensus among experts seemed to be that T. rex and other very large theropod dinosaurs probably only had feathers as juveniles, if at all. The idea was that large animals have no need for feathers for insulation as their surface area is much smaller relative to their body size than that of small animals, and they lose heat to the environment much more slowly (the largest land animals alive today – elephants, rhinos, hippos – are all hairless precisely because they have trouble keeping cool). But the discovery of two feathered relatives of Tyrannosaurus has turned this idea on its head.
The first, Dilong paradoxus (‘paradoxical emperor dragon’), was discovered by Xu’s team in Liaoning in 2004. This lightly built predator was an early relative of T. rex that stalked the fauna of Early Cretaceous China around 125 million years ago (T. rex itself hails from North America and dates to the Late Cretaceous, around 66 million years ago). As it was a relatively small animal, 2 metres long, the downy covering of dinofuzz preserved in the fossil was not wholly unexpected. Dilong is among a number of small tyrannosaurs discovered in China in recent years. Others such as Guanlong wucaii are also inferred to have had feathers, even though their fossils don’t retain direct evidence of this. Guanlong (an illustration of which graces the cover of this book) is the earliest known of all tyrannosaurs, at about 160 million years old.
Much more surprising was Xu’s discovery of the bus-sized, 9-metre-long Yutyrannus huali (‘beautiful feathered tyrant’) in 2012. Also from the early Cretaceous of Liaoning but around 5 million years younger than Dilong, this shaggy predator was closer in size to T. rex itself. It showed that simple downy feathers were probably much more widespread among dinosaurs than anyone had expected. Yutyrannus is not only the largest feathered dinosaur discovered to date, it is also the largest feathered animal known to have lived; at an estimated 1.4 tonnes, it tipped the scales at about 5000 times the weight of an average bird, such as a pigeon.
Three specimens of Yutyrannus were found, two of them with a largely complete set of bones, and they indicated downy feathers on the legs, arms and neck. (The specimens were also found in the same quarry, hinting that Yutyrannus huali may have been a social species that lived in groups, and adding credence to Phil Currie’s theory that some tyrannosaurs were pack hunters.) A great illustration of Yutyrannus, released along with the Nature paper that described it, showed a family group in a snowy-looking environment with puffs of breath misting in front of them (see image section). The fact that feathers were only found in a number of patches on the body may just be an accident of preservation, or they could indicate that that they were used only for display purposes rather than as insulation.
In The Bird: A natural history of who birds are, where they came from, and how they live, Colin Tudge argues against the idea that only juvenile T. rex were feathered, pointing out that if feathers had a sexual display purpose then adults would be even more likely to be extravagantly feathered than their young. ‘It is not possible, and perhaps never will be, to state definitively whether T. rex did have feathers – or to confirm or disprove the idea that grown-up T. rex’s were turned out like burlesque queens. But I reject that school of biology that feels it is virtuous simply to be dour’, he writes. ‘In reconstructions, T. rex is generally dressed in scaly leather, like a Gucci handbag. But it might have been decked in gaudy feathers and, in due season, crested and plumed for good measure.’
Although Yutyrannus was still just a fifth the weight of T. rex, its discovery certainly increased the chance that T. rex might also have feathers, as Tom Holtz, a University of Maryland palaeontologist, told National Geographic. And even with a fluffy covering of down it would have been just as scary, he added: ‘Underneath the fluff, it’s still the same gigantic crushing teeth and powerful jaws and softball-sized eyes staring at you … [feathers] might make it a little more amusing, but only until the point right before it tears you to shreds.’
Filaments, bristles, scales and fluff
A few of the new Chinese fossils hint that feathers have their origins much deeper in the dinosaur family tree, not close to the species that evolved into birds. There’s even the tantalising possibility that feathers originated in the ancestors of animals that gave rise to dinosaurs and their sister group of flying reptiles, the pterosaurs.
Among the dinosaurs that have been discovered in recent decades, there have been some pretty weird creatures, from shaggy pot-bellied beasts with metre-long, scythe-like claws to hump-backed predators with fully feathered ‘wings’. One of the strangest came not from China but from Germany – from the very same Bavarian Jurassic limestone that offered up Archaeopteryx 150 years ago. Sciurumimus, the ‘squirrel mimic’, a small megalosaur with a great big bushy tail of downy feathers, was found in 2012. Head to tail, the fossilised juvenile was about 70 centimetres long, but adults would have been much larger. The really interesting thing about this feathered species is that it is among a very early group of carnivorous theropods and isn’t a coelurosaur, the group that contains the majority of the other feathered species we’ve discovered.
‘All of the feathered predatory dinosaurs known so far represent close relatives of birds’, said Oliver Rauhut of the Bavarian State Collection for Palaeontology and Geology, one of the scientists behind the find. ‘Sciurumimus is much more basal within the dinosaur family tree and thus indicates that all predatory dinosaurs had feathers.’
There have been other remarkable finds – of seemingly feathered dinosaurs that are even more distantly related to birds. Tianyulong confuciusi was a small bipedal herbivore with a fuzzy covering of primitive feathers. Nothing unusual in that, except this animal was in the ornithischian group of dinosaurs thought to have diverged from the line that led to birds 220 million years ago (the ornithischians are one of the two great branches of the dinosaur family, the other being the saurischians, which included the theropods that gave rise to birds and the giant sauropods).
Tianyulong is not the only ornithischian to have been found with structures bearing a similarity to feathers. Psittacosaurus is a very early member of the parrot-beaked ceratopsian lineage that eventually led to Triceratops near the end of the dinosaur era in the late Cretaceous. Psittacosaurus was originally described from Mongolian rock deposits, but more recently discovered Chinese specimens reveal plumes of bristles along the tail. It’s therefore possible that all groups of dinosaurs could have had simple feathers. ‘There appears to be evidence suggesting that even the filaments of ptero- saurs are likely to be a kind of primitive feather’, says Xu.
Paul Sereno, veteran dinosaur hunter at Chicago’s Field Museum, isn’t so sure. He says there’s evidence of scaly skin impressions without feathers for large dinosaurs, such as sauropods, and for many groups other than the theropods. There’s no evidence of feathers in other ornithischians, agrees Paul Barrett who published a study on this in early 2014 along with David Evans of the Royal Ontario Museum in Toronto. ‘We have lots of skin impressions from duck-billed and horned dinosaurs, and none of them show anything that looks like feathers. So it could be that these dinosaurs started off with feathers and lost them, or it may just be some underlying genetic mechanism whereby dinosaurs do stuff with their skin – because they also have lots of armour and spikes that form in the skin too.’
Experts have known for some years now that many pterosaurs had a fur-like covering. ‘That’s been observed in fossils – there are quite a number of exceptionally preserved fossil pterosaurs, different ages, and under the microscope it seems that they have hair-like structures covering the body’, says Mike Benton. ‘This was because pterosaurs, like birds, had to have a very high metabolic rate [for flight] and the covering would have been used for insulation.’
The question now is did all dinosaurs and pterosaurs inherit feathers from the same common ancestor, or is it just that the group had a remarkable plasticity that allowed its members to play around with dermal structures of different kinds – such things as bristles, quills, fuzz, fluff, ribbons and, eventually, elegant, complex and beautiful feathers, aerodynamically sculpted for the purpose of flight? ‘It could be that all dinosaurs had the propensity to produce feathers or whiskers, but whether we can say that the ones in Psittacosaurus are actually feathers, or whether they are some other kind of dermal structures, we don’t yet know’, says Benton. ‘It could be that all small dinosaurs had feathers – it could be lost in the large ones – or it could be that dinosaurs manifested a variety of dermal [skin] growths and dermal structures of many kinds.’
‘There might just be a propensity to experiment with skin structures in dinosaurs – and ornithischian dinosaurs just experimented with feathers once or twice’, agrees Barrett. ‘Maybe they had the underlying genetic machinery but never really went into it for one reason or another. That’s one of the big mysteries. We don’t know if other groups will have feathers, but it’s not looking likely. It’s looking like it’s part of the theropod dinosaur story, but not part of the story of the other groups.’
Nevertheless, if one of the early ancestors of all dinosaurs – or dinosaurs and pterosaurs – did have simple feathers, then it opens up the possibility of fluffy sauropods as well as fuzzy duck-bills, armoured dinosaurs and their ornithischian kin.
‘Some artists have already taken the opportunity and there are stegosaurs and sauropods appearing with hints of fluff’, writes British palaeontologist Dave Hone on his blog ‘Lost worlds’:
At the moment, it’s probably best considered not much stronger than informed speculation, but it’s certainly not unreasonable as a hypothesis or improbable … It was thought unlikely to the point of impossibility that ornithischians would have any kind of covering beyond scales and armour, but two different species were found to have filaments less than a decade apart, and it has even been suggested that the famous Triceratops had some bristles as part of its skin.
The only skin-impression fossil found for Triceratops appears to reveal it had bristles, not purely scaly skin as has been long supposed. The remarkable fossil was found by palaeontologist Bob Bakker and is held at the Houston Museum of Natural Sciences in Texas, where he is a curator (although at the time of writing, nothing had yet been published on it). We know that Psittacosaurus had bristles, so given it is an early member of the lineage that led to Triceratops, it seems reasonable to assume that these were passed down.
What good is half a wing?
‘The sight of a feather in a peacock’s tail, whenever I gaze at it, makes me sick!’ said Charles Darwin in a letter to his colleague Asa Gray in 1860. The issue troubling Darwin was that a peacock’s plumage seemed so exuberantly gaudy and flashy, and so counterintuitive in terms of preventing predation, that it made a mockery of his recently published theory of natural selection. But if he’d spent more time thinking about the structural complexity of feathers, and how they could possibly have evolved through natural selection, that might have led to a sense of nausea instead. Indeed, the evolution of flight and feathers presented such a problem to evolutionary biologists for so long that creationists would routinely cite them as evidence against the process of natural selection. ‘What good is half a wing?’ has often been asked (and we’ll come to the answer in chapter 7).
Darwin’s contemporary and friend Alfred Russel Wallace, who came up with the theory of natural selection at roughly the same time and found much evidence for it in his travels through South-East Asia, was as perplexed as many other evolutionary biologists by the problem. ‘Evolution can explain a great deal; but the origin of a feather, and its growth, this is beyond our comprehension, certainly beyond the power of accident to achieve’, he told London’s Daily Chronicle in 1910:
A feather is the masterpiece of nature. No man in the world could make such a thing, or anything in the very slightest degree resembling it … Watch a bird sailing high above the earth in a gale of wind, and then remind yourself of the lightness of its feathers. And those feathers are airtight and waterproof, the perfectest venture imaginable!
The idea of feathers and flight are so intimately entwined in our minds that we find it hard to imagine the former might ever have served any other purpose. But it’s clear that fully flight-capable animals can’t have emerged with a complete set of flight feathers ready to go. Feathers are complex structures and must have developed over a significant period of time.
The feather impressions on the fossils of the ‘first bird’, Archaeopteryx, show that it had flight feathers similar to those found on modern birds, which – something like the wing of an aircraft – are asymmetrical to make them aerodynamic and provide added lift. Despite this, most experts think Archaeopteryx must have been a pretty poor flyer that struggled to get airborne and most likely had to make ungainly crash landings. Its lack of a breastbone to which strong flight muscles could attach, in addition to its long bony tail and dense skeleton, denied it the kind of aerial grace modern birds take for granted. In any case, Archaeopteryx tells us frustratingly little about the evolution of feathers, even though, at 150 million years of age, it’s significantly older than many of the Cretaceous-era feathered dinosaurs discovered in China (a confusing problem that has been labelled the ‘temporal paradox’ – more on that in chapter 7).
There are several ways of looking back at feather evolution over time – or indeed the evolution of any other trait, including: examining the hard evidence of feathers in fossils; studying the differences between feathers in different groups of animals and using that to make inferences about how they evolved; and looking at the development of feathers in the embryos of modern birds. A slick new field called ‘evo-devo’ (which is short for evolutionary developmental biology) looks at the changes that happen in the developing embryo and relates them to genetics. The 19th- century German biologist and comparative anatomist Ernst Häckel pointed out that ‘ontogeny recapitulates phylogeny’. Ontogeny is the course of development of an organism and phylogeny is evolutionary history. In simple terms, we can, in a sense, replay the history of a feature, such as feathers or teeth, by watching how it develops in the embryo.
In the early 1990s, Richard Prum came up with an idea, based on development in the embryo, about how feathers might have evolved and the steps they went through to arrive at the beautiful and intricate complexity of a modern flight feather. Never did he believe he might one day be able to follow these steps of feather evolution in a series of exquisitely preserved dinosaur fossils, but this is exactly what he has been able to do.
How did feathers evolve?
Prum is the world’s pre-eminent expert on feathers. In his early career he travelled the globe from South America to Madagascar observing bird courting and mating behaviours and listening to their many and varied calls. Of particular interest to him were the manakins of the South and Central American tropics, small songbirds with an unusual syrinx or voice box that can produce a diverse repertoire of trills, whistles and buzzes. As a graduate student he began to unpick their complex relationships and produced the first family tree detailing their evolutionary history.
A promising career as a field ornithologist lay ahead of him, or so it seemed, but then in the early 1990s, when Prum was in his mid-20s, calamity struck. Through illness, he began to lose his hearing in first one ear and then the other. Eventually he lost the ability to hear high notes at all – which was disastrous, as it meant he became effectively deaf to most birdsong.
Prum often recounts the story of when he first realised his hearing loss was significant enough to affect his work as an ornithologist. In 1998 he was in Madagascar with a group of University of Kansas colleagues, taking them to see the mating dances of a bird called the velvet asity. The group followed forest trails to find birds Prum had banded four years earlier and quickly spotted one of the males in the same patch he’d seen it the last time he was there. ‘I lift up the binoculars, and there he is, and I’m showing all these guys their first velvet asity’, Prum told the Yale Alumni Magazine. ‘He opens up his mouth [to sing] and … I couldn’t hear a damned thing. I knew that my hearing loss had started, but I didn’t realise this bird, whose song I had described to science, was now inaudible to me.’
Though disastrous to him personally, this hearing loss was a win to scientific enquiry, as Prum focused the fierce beam of his intellect on another ornithological problem that had never been adequately answered. First he began to look at the colours of birds and the way they use them to attract mates, but he soon began to look more specifically at the mystery of feathers.
Though they are more structurally complex, feathers are similar to hairs, nails, claws and scales in both chemistry and origin. All are made of varieties of the protein keratin and are dermal structures, formed as outgrowths of cells in the outer layer of skin or epidermis.
A typical feather – perhaps much the same kind fabled to have been dropped by Galileo from the Leaning Tower of Pisa to prove that the time taken for objects to fall is independent of their mass, or the kind used as a quill by Shakespeare to scribble down his sonnets and soliloquies – has a central rachis or vane, from either side of which protrude a series of branched filaments or barbs. These barbs themselves also have barbed branches, with tiny hooks that allow them to attach to and align with one another, creating the elegant and neat design that is the modern pennaceous bird feather. Birds also have downy or fluffy plumulaceous feathers, which have no central vane and are a messy jumble of filaments. These are the kind of feathers we use to fill our pillows, quilts and winter jackets, precisely for the reason that they are wonderfully insulating. In fact, a great variety of feathers is built around these models – even down to the eyelashes and whisker- like sensory structures of some species of bird – and all of this diversity comes from a simple hollow tube of protein produced by the skin.
Prum’s developmental theory of the origin of feathers looked not to the fossil record and why feathers evolved, but instead to the development of the embryos of modern birds and how feathers first appear and grow in complexity.
The reason scant progress was made on understanding the origin of feathers before the 1990s is that experts assumed feathers evolved for the purpose of flight, and on that basis struggled to come up with any plausible scenario of how this might have occurred. ‘Only highly evolved feather shapes … could have been used for flight’, wrote Prum in a Scientific American article he co-wrote with Alan Brush in 2003. ‘Proposing that feathers evolved for flight now appears to be like hypothesizing that fingers evolved to play the piano.’
There was also an idea that feathers had evolved from scales, but this seemed all wrong to Prum, as feathers are tubes and scales are flat. Before 1996, scientific discussions about the origin of feathers were entirely speculative, because there were no modern examples of early designs of feathers. ‘Archaeopteryx has feathers that are identical in structure to a modern bird’s, so there were no data’, Prum says. ‘The literature consisted of trying to extrapolate backwards from modern complexity to some simpler, primitive, antecedent structures. Most researchers tried to do that by imagining functional scenarios for the origin of feathers – usually organised around the idea that feathers evolved for flight. These failed completely.’
Instead, Prum proposed a new theory of the evolutionary origin of feathers based on the details of how feathers grow. He developed the theory in ornithology classes he ran at the University of Kansas, but wasn’t motivated to publish it until the spate of feathered and fluffy dinosaur discoveries in the late 1990s.
Prum’s idea about how feathers evolved is elegant in its simplicity. He argues that the steps via which a single feather develops in a bird today closely follow the evolutionary steps by which the modern feather came about, and therefore offer a window into the past. In the first stage of growth of a feather, the epidermis thickens to form a little bump on the surface of the skin called a placode, which then elongates to create a hollow tube or unbranched quill (stage 1). This then divides into a fluffy tuft of feather filaments with no central vane (stage 2). Following this the feather can develop further branching in one of two ways: it can form a central vane with simple filaments branching off it (stage 3a), or it can form tufts of feathered filaments with no central vane but further branching of each of the filaments or barbs to form barbules (stage 3b). The next step sees both of the former stages combine to create a modern pennaceous feather, and also hooks and grooves form in the barbules, allowing them to grip into and align with one another neatly, a little like a zipper (stage 4). In the final stage of development, the length of the barbs or filaments on one side of the rachis increases to create a modern asymmetrical aerodynamic flight feather (stage 5).
The beauty of this theory is that it’s now testable through examination of the fossil record. Sinosauropteryx, discovered in 1996, appeared to have had a fluffy covering of simple feathers not unlike the stage 2 filaments; Caudipteryx, found a few years later, appears to have this same dinofuzz, but also stage 3 feathers on the forelimbs and tip of the tail with a central vane and symmetrical filaments branching off it. Stages 4 and 5 were found in 2000 in Microraptor, a crow-sized theropod with some flight capability, now known to have had blue–black feathers and an iridescent sheen.
Feathers entombed in amber
Although the fossils from China have been a great source with which to test these ideas about feather evolution, even unusually fine shale fossils can preserve only so much detail, and what we can learn from them about feathers is limited. That’s why it was so exciting when a team of researchers from the University of Alberta, including Phil Currie, announced in Science in 2011 that they had found a veritable treasure trove of tiny feathers in Canadian amber from the Late Cretaceous, 70–85 million years ago.
In a paper illustrated with a beautiful series of colour photos of these feathers and filaments, the authors reported that they had found 11 different types, and that they were likely to be a mixture of bird and dinosaur feathers. You can imagine these long-gone beasts brushing past sticky patches on trees and leaving the feathers behind, ready to be entombed as the sap became amber, preserving the physical structure of the feathers almost perfectly. This was yet another example of something totally unexpected from the fossil record. The researchers had used museum collections to scour 3000 pieces of amber known to have interesting ‘inclusions’ embedded within them.
The paper’s first author, Ryan McKellar, told The Atlantic that the discovery was significant for a number of reasons. ‘It supports a model for the evolution of feathers that has previously relied on compression fossils that are difficult to interpret and have been hotly debated’, he said. ‘Also, the amber-entombed feathers show that some of the most primitive feather types, also known as protofeathers, were still around just before the dinosaurs went extinct. They existed alongside feathers that are nearly identical to those of modern birds. Given what we know about the animals that were alive in the area at the time, it is reasonable to suggest that the protofeather-like specimens are attributable to dinosaurs.’
McKellar, Currie and their co-workers reported that the feathers represent a number of distinct stages of feather evolution, from stage 1 protofeathers of the kind sported by Sinosauropteryx right through to much more complex types of feather, such as those that help diving and swimming birds repel water and those that birds use for flight. As an added bonus, some of the specimens retained colour, suggesting mottled patterns of brown and black.
In an accompanying commentary published in the same issue of Science, Mark Norell noted that it was only a very short while since most people had considered dinosaurs to be scaly and dull-looking animals, and our only peek at feathers from the deep past were at those of Archaeopteryx, which appeared so frustratingly modern, offering little clues as to how they had evolved. ‘How things have changed – now it would take a warehouse to store all the feathered Mesozoic stem birds and non-avian dinosaurs that have been collected from global deposits’, he wrote. ‘Feathered animals abound and extend deep into non-avian history – even, perhaps, to the base of dinosaurs themselves. Now, instead of scaly animals portrayed as usually drab creatures, we have solid evidence for a fluffy coloured past.’ It may not have been dinosaur blood or DNA, but the discovery of bits of 80-million-year-old feathers preserved in tree resin showed that the idea of dinosaur genes coming from mosquitoes trapped in amber isn’t as far fetched as it might first appear.