The struggle to the skies
Flight is the defining feature of birds today, but how did animals closely related to the 7-tonne T. rex ever get into the air, and why did they bother?
One hundred and twenty million years ago, in the Early Cretaceous of north-eastern China, a dinosaur little bigger than a raven is using his sharp claws to scramble up the trunk of an ancient conifer. A shimmering golden sliver of morning sun crowns the distant horizon and bathes the treetops in orange light, illuminating the mist hanging between them. This miniature relative of Velociraptor is covered in long, glossy, blue–black feathers. He pauses on a mossy branch to preen them, carefully probing his plumage with tiny teeth and smoothing the feathers down again with his snout and claws.
Unlike the feathers of some of the larger dinosaurs down on the forest floor, his are not primarily for display or insulation; they are fully formed feathers with a central vane and an asymmetrical shape sculpted to aerodynamic perfection. These feathers are intended for flight.
And this little Microraptor, as his kind is known, doesn’t just have flight feathers on his forelimbs, but on his hind limbs and the end of his long tail, too. This early experiment in aviation has four wings and can use them to glide great distances artfully, on a winding path through the forest that takes him from tree to tree.
Preening concluded, a breakfast of fish from the nearby lake plays on his mind. He stands tall and shakes himself, opens his large forewings wide and launches from the branch, his hind wings hanging beneath and behind him. He isn’t capable of powered flight by flapping – or at least not very proficiently – but the extra surface area provided by the feathers on his legs and the fan of feathers around his tail makes him a very effective glider.
This Microraptor was one of the most common creatures in these Liaoning forests, and more than 300 fossils of his kind have been found so far. Although it was not a direct ancestor of modern birds, this unexpected little dinosaur, and a number of other four-winged flyers, are helping experts understand how dinosaurs first took to the air and made the transition to fully developed birds.
From gliding to powered flight
The first vertebrates to take to the air were probably gliders. Gliding is effectively a form of controlled falling, and is much easier to achieve than powered flight. The difference between gliding and powered flight is the ‘flight stroke’ of the wings, which generates lift. Rather than having a powerful flapping wing, gliders usually have a rigid gliding membrane, which makes them much less manoeuvrable than true fliers.
True flight not only requires flapping wings, but also strong musculature to power them, and – in birds – a breastbone or sternum to anchor the muscles. Wings (and feathers) are aerodynamically sculpted to provide the greatest lift. Wings tend to be curved on their upper surface and flatter underneath, and are rounded on their leading edge and tapered at the rear. This shape generates upward lift as the wing moves forward through the air, and this lift works against gravity to keep the animal airborne.
Aircraft wings work on exactly the same principle. But whereas aeroplanes use propellers or jet engines to generate thrust that moves them forward, the wings of birds, bats, dinosaurs and pterosaurs have to flap to provide thrust as well as lift. In gliding animals it’s falling that provides the thrust that moves them forward and generates some lift that in turn slows their rate of fall; they therefore have to glide at a slight downward angle, and they have to climb to a height to take to the air again, as gliding can only take them downwards.
Birds make flight look easy with their graceful swooping, hovering, turning and diving. The peregrine falcon, for example, can dive at speeds of more than 300 kilometres per hour. The spinetailed swift can fly in a straight line at 170 kilometres per hour, all the while executing sudden and seamless changes in direction. Hummingbirds can hover on the spot and even fly backwards, while albatrosses and swallows stay airborne for much of their lives, only rarely coming back to land.
Flight may look easy, but in truth it’s a marvel of mechanical engineering. Becoming a flyer, and evolving a complex suite of features that will make powered flight possible, is an almost unimaginably complex process. In birds these features include: hands fused into wings, flight feathers, small body size, pneumatisation (hollow bones), low body weight, high metabolism, a keeled sternum and wishbone to which powerful flight muscles attach, large heart, large brain, sharp eyes, powerful visual processing capability – the list goes on. In short, it’s so difficult that very few animals have had the evolutionary tenacity to pursue it over the past 350 million years: these were the insects, pterosaurs, birds and bats.
‘The transition to flight was a very tough evolutionary step, far greater than our own historical move to walking on two legs’, says Paul Sereno. ‘There was a great unpopulated niche-land to exploit, waiting for a toothy animal with a backbone that was more nimble than a pterosaur. That’s why evolution took them there, but the transition was very, very difficult, and achievable only at small body size.’
Though flight is difficult, the payback comes in some significant advantages over land-bound creatures. Flight allows animals to evade predators rapidly and remain out of their reach. It also allows access to previously inaccessible resources such as food and nesting sites high up in trees, and it allows animals to migrate over large distances in search of food and mates.
Flying may seem energetically expensive in that it uses more fuel than swimming or walking for the same period of time, but it uses less energy to cover the same distance, making it extremely efficient. Walking or running uses up to 10 times as much energy to cover the same distance as flight, and fliers move through the air 10–20 times faster on average than terrestrial animals can move on the ground.
Avoiding predation is a plausible reason that flight evolved in birds, an idea backed up by the fact that if you take predators away, many birds will give up flight altogether. Across the planet, birds have found their way to predator-free islands and, provided they have the resources they need, lost the ability to fly. Most of the examples seem to be in the Pacific – the takahe, kakapo, kiwi and moa of New Zealand come to mind, as well as numerous species of Pacific rail – but there’s also the great auk of the North Atlantic, the dodo of Mauritius, the elephant bird of Madagascar and the Galapagos cormorant.
Flight certainly seems to be linked to success, in terms of numbers of species in the groups that have achieved flight. Living birds number around 10 000 species compared to 5500 mammal species. Of those mammals, around 1240 (23 per cent) species are bats. In terms of species, bats are the most numerous of all mammals apart from rodents, although we’re rarely aware of it, because bats are active when most of us are asleep. In total about a third of all species of land vertebrate can fly.
‘Flight is clearly a factor in the diversity of these animals and how they’ve been able to exploit the environment’, says Luis Chiappe. Insects, which make up the fourth group in which flight evolved, are the most numerous of all animals, numbering perhaps a million species or more. The majority of insects fly during at least one stage of their development, which means that the majority of all animals can fly. There’s clearly a strong correlation between flight and evolutionary success. Flight is difficult but very useful. So how did dinosaurs achieve it?
Flight and flightlessness
The sad thing is that once you go flightless, there’s probably little chance of evolving flight again. And since flight is a powerful tool to help animals avoid becoming somebody’s lunch, when humans spread out across the globe – often with other predatory species, such as rats, cats, dogs and pigs in tow – these flightless wonders, with no innate fear of predators, stood little chance of survival. A study from the University of Canberra has shown that as Polynesians colonised the 269 larger islands of the eastern Pacific around 700–900 years ago, they likely drove around 1000 species of birds extinct, many of which were flightless ground-dwellers.
From tiny to Titanic
What may be the largest dinosaur, Argentinosaurus, weighed 100 tonnes and was more than 30 metres long. The smallest member of the dinosaur family is the bee hummingbird, found on Cuba today. Its body is scarcely bigger than an insect’s, measuring just 5 centimetres from beak tip to tail tip, and it weighs less than 2 grams. This hummingbird beats its wings 80 times a second, and the energy expended doing this means it has to consume half its body weight in nectar a day, which it has to gather from around 1500 flowers. A single Argentinosaurus weighed as much as 50 million bee hummingbirds. This staggering figure illustrates the amazing flexibility of the dinosaur body plan, which has adopted more shapes, styles and sizes than just about any other group of animals, but it also suggests that miniaturisation was an essential prerequisite for flight. All animals that have achieved flight started out small, and dinosaurs were no exception. Some of the smallest known dinosaurs, such as Microraptor and Anchiornis, are those closest to the origin of birds.
‘Palaeontologists thought that miniaturisation occurred in the earliest birds, which then facilitated the origin of flight’, said Alan Turner of the AMNH. ‘Now the evidence shows that this decrease in body size occurred well before the origin of birds and that the dinosaur ancestors of birds were, in a sense, pre-adapted for flight.’ Turner was first author of a 2007 Science paper showing that the dinosaur lineages leading to birds began to downsize fairly early in their history.
Turner’s research revealed an 80-million-year-old dinosaur in Mongolia’s Gobi Desert that was an early offshoot of the dromaeosaur lineage (which includes Velociraptor and Microraptor). Mahakala is thought to have weighed 700 grams and measured about 60 centimetres in length. Although the species dates to a period long after the evolution of birds, it is thought to have split off from other members of the group much earlier in the Mesozoic (the era comprising the Triassic, Jurassic and Cretaceous), before the first birds appeared. The experts behind the find said it was evidence that dinosaur lineages had progressively decreased in size as they got nearer to birds. ‘Flight isn’t an easy thing, because you are, in effect, countering the force of gravity. Being really small appears to be a necessary first step’, Turner wrote. ‘Other groups that evolved flight, such as pterosaurs and bats, all evolved from small ancestors. With the discovery of Mahakala we were able to show that this miniaturisation occurred much earlier.’
There are two competing theories about how the transition to flight was eventually made once dinosaurs were small enough to achieve it: ‘ground-up’ and ‘trees-down’. The issue is still the subject of debate among palaeontologists. The ground-up idea (more widely accepted until recently) suggests that the ancestors of birds flapped their feathered forearms to increase running speed. ‘This scenario, to some extent duplicated by some living birds, would explain the origin of flight from the ground up, as these flapping and running dinosaurs became smaller and their wings became larger’, Chiappe says.
Some speculate that feathers helped swift predatory species become more streamlined. One theory holds that the sharp retractable claws of the raptors were a specialisation that allowed them to hook onto – and scale the back of – large herbivorous species. In this context, it’s easy to see how branched flight feathers could have helped give these animals an extra bit of lift to jump and glide when stalking prey. Another theory suggests that the feathered forearms of dromaeosaurs allowed ‘stability flapping’ to maintain balance while gripping onto struggling prey, a behaviour that can be observed in hawks and eagles today.
Not everyone agrees, though. David Varricchio at the University of Montana argues that the ground-up hypothesis doesn’t really add up. ‘Personally, I feel the [theory] is absurd since we have no examples of “running flyers” today that could represent an intermediate stage’, he says. ‘There are more examples of flying fish than flying runners. But there are many examples of trees-down gliders: gliding frogs, marsupials, various placental mammals, snakes and lizards.’ But the problem with the trees-down theory is that although there are hundreds of gliding species alive today, none of them appears to have developed anything like a flight stroke or wing beat that might make them an intermediate stage between gliding and powered flight.
This said, bats may have evolved from gliders and seem most likely to have come from the trees down. Modern bats need a short vertical drop in order to take to the air and are helpless if they find themselves on the ground, needing to crawl awkwardly to a tree or other vertical surface and climb up it to get airborne once more. In outback Western Australia I’ve seen zoologists throw bats into the air after they’ve trapped, weighed and measured them, and the bats seamlessly begin to fly once more. In this context it seems likely that bats evolved from gliding ancestors.
Pterosaurs (built on a somewhat similar body plan to bats, with sheets of skin supported by elongate forelimbs) may have evolved from gliders too, but their fossils show no evidence of adaptations to living in treetop environments as bats do. ‘If time and again the evolutionary answer to trees-down flight is a membrane, then bird feathers are an exception that suggests a different history’, argues biologist Thor Hanson in his book Feathers: The evolution of a natural miracle. ‘With their unique follicles and helical growth, their complex structure and diversity of forms, feathers seem grossly overqualified for the job. Why go to all that trouble when a flap of skin would do?’
Our old friend Thomas Henry Huxley, Darwin’s bulldog, was among the first to posit the ground-up idea, after his studies of Archaeopteryx in the 1860s. He suggested it had evolved from a swift bipedal runner, something like Compsognathus, which had a very similar-looking skeleton. For the carnivorous theropods, perhaps, the beginnings of flight would have made hunting easier and more effective, allowing them to leap onto prey and run them down, or flap into the air to snatch at insects. Another tick in favour of the ground-up theory is that birds today use their forelimbs to fly, and in order for this to have evolved, it seems they must have been freed up for the purpose in advance. This is exactly the case in a bipedal runner.
In 2003 Liaoning offered up something totally unexpected – and it lent powerful support to the trees-down school of thought.
Microraptor gui was a four-winged species with modern flight feathers on both its forelimbs and its hind limbs, in addition to a fan of feathers on its tail. Xu Xing led a team that described the species from six specimens found in Liaoning, all around 125 million years old. The 80-centimetre-long animal probably held its hind limbs behind it at a 45-degree angle to the plane of its wings, and then glided, like a flying squirrel, between the treetops. Xu’s original idea was that Microraptor splayed its hind wings out to either side directly behind it, but anatomists have pointed out that this would have dislocated its hip joints. Another idea was that it tucked its hind wings up under its forewings to create a biplane configuration. This has also been discounted, and the best evidence yet, from models flown in wind tunnels, seems to suggest its legs would have hung beneath it but angled out to the sides a little, while its wings (forearms) would have been held out to the side in the position we’re familiar with for birds.
Resolving the ‘temporal paradox’
One confusing problem in unravelling the evolution of flight has been that most of these feathered dinosaurs are significantly younger than Archaeopteryx itself, which is about 150 million years old. This is because not many of the older, Jurassic deposits are as good as the Cretaceous deposits of Liaoning. Although Liaoning’s species can give us clues about the transition, they cannot be direct ancestors of birds, as they were contemporaries of a diverse fauna of early birds, which are also found in the deposits. The idea instead is that these species are close parallels of much older animals that sequentially split off on different branches of the family tree on the way to birds. It can be a bit perplexing to think about, but when new types of species evolve, older types often remain alongside them. For example, our own ancestors are known to have included fish and amphibians, but these still exist alongside us today.
Nevertheless, several new dinosaurs with fully developed feathers on their hind limbs and even feet have helped settle this ‘temporal paradox’. Anchiornis huxleyi, described in Nature in September 2009 by Xu Xing and colleagues, is a troodontid species that lived between 151 and 161 million years ago in Jianchang, China. It therefore pre-dates Archaeopteryx.
‘Anchiornis is really amazing. It shows that four-winged dinosaurs, from different groups, lived across different time periods’, Xu says. ‘We now have quite strong evidence supporting the idea that the first bird was a four-winged animal like Anchiornis or Micro- raptor, and it probably originated around 160 million years ago.’ Evidence is building that feathered hind limbs were the ancestral condition in a species that pre-dated the split between dromaeosaurs such as Microraptor, troodontids such as Anchiornis, and birds.
Described in 2011 by researchers including Xu, Xiaotingia zhengi is another very bird-like species with feathered hind limbs. It is of a similar age to Anchiornis and Archaeopteryx, and at the time Xu claimed it knocked Archaeopteryx off its perch as the earliest bird. Subsequent studies have disputed this, however, suggesting instead that Xiaotingia is an early dromaeosaur.
A study of 11 kinds of early bird – all of which lived around 130 years ago in Liaoning and had long, vaned feathers on their hind limbs – shows that the four-winged flight model continued to be common among early birds after the split from non-avian dinosaurs. This is quite unlike the condition in modern birds that have leg feathers – some varieties of chicken and pigeon, for example. In these species the feathers are small, downy feathers good for insulation but useless for flight.
Using specimens in the vast collections of the Shandong Tianyu Museum, a Chinese team revealed that feathered hind limbs were surprisingly common and may have been an important step that allowed the evolution of full flight. The authors of a paper in Science, including Zheng and Xu, noted that these hind-limb feathers were ‘aerodynamic in function, providing lift, creating drag and/or enhancing maneuverability, and thus played a role in flight’.
Some of these early bird species – including Sapeornis, Yanornis and Confuciusornis – appeared to be in the process of reducing their hind-wing feathers and developing the kind of scaly, bird-like feet we are familiar with today. This would have allowed them to move more swiftly on the ground, as the hind-limb feathers of four-winged dinosaurs such as Microraptor and Anchiornis must have been a significant impediment to a speedy escape from predators, particularly if they had to gain height before they could take to the air once more.
Running up that hill
A famous series of studies carried out by Ken Dial, director of the Flight Laboratory at the University of Montana, has offered up some unexpected clues that are instructive in the ground-up versus trees-down debate. Although we may not have found any species that show an intermediate stage between gliding and powered flight, Dial’s work has demonstrated what an intermediary stage between running and flying might look like.
Dial had been interested in how flight might have evolved in birds for some time, but found the lack of evidence from the fossil record frustrating. Instead he wondered if he might be able to glean anything from living animals. The species he decided to focus on was the chukar partridge, which is similar in some ways to theropod dinosaurs in that it’s a bipedal ground-dwelling runner. It has some flight ability, but chooses to run instead 95 per cent of the time.
Chicks are born in nests on the ground and learn to run rapidly in order to avoid being eaten. To make an efficient escape from predators they have developed something Dial calls wing-assisted incline running or WAIR. Working with his teenage son, Terry, Ken revealed that escaping partridge chicks flap their small wings frantically in such a way as to help create a downwards force as they run up the side of steep inclines or vertical obstacles. This is similar to the way a spoiler pushes a racing car towards the ground to give it traction as it turns a corner. So is it possible that the ground-dwelling feathered dinosaurs first began to evolve the flapping strokes that would later lead to flight as a method to help them rapidly negotiate obstacles in the terrain? It certainly seems plausible.
Since Darwin’s day it has seemed idiotic to think that a halfformed wing could be useful. But as evolutionary biologist Richard Dawkins says: ‘Half a wing is indeed not as good as a whole wing, but it’s certainly better than no wing at all.’ What Dial’s research suggests is that the stumpy wings of partridge chicks do serve a useful purpose, and that similar proto-wings in dinosaurs may have allowed them to flutter back down to the ground from the trees whose trunks they’d just used their wings to scramble up.
Dial likes to think of his idea as a ‘complex marriage’ of the ground-up and trees-down ideas. ‘We have taken the beautiful sage elements from each one, and I feel we integrated them perfectly to say you never needed to go strictly from the ground up or tree down’, he said in 2008:
The eons-long evolution of flight is revealed to us in the development of baby birds … Our thesis came out from the demonstration of what living animals actually do. And now we have fossils that we never imagined being discovered in China, South America and Africa that look exactly like we expected – dinosaurs with feathers; dinosaurs with half a wing.
Most particularly he points to Caudipteryx, a species discovered in Liaoning in 1998. It has a small fan of feathers around its stunted forelimbs, which may have been used for display but could also have been used for wing-assisted incline running.
David Varricchio argues that Dial’s research is really interesting, but because partridges evolved from earlier birds that already had full flight capacity, they may not be a fair example to look at when trying to work out how flight evolved in the first place. ‘You’re basically looking at an animal where its ancestors had full powered flight’, he says. ‘How do you translate that model back to say this is how a dinosaur evolved in flight?’
There are some other clues that suggest large vaned feathers and wings could have offered benefits for running, leaping and climbing animals long before they became useful for flight. African ostriches have large wings they use as sophisticated air rudders and braking aids to increase their manoeuvrability as they run. This means they can perform rapid zigzag manoeuvres to escape danger, even while running at 70 kilometres per hour. Ostriches are part of an ancient lineage of early birds, and the fact that they are a large flightless species with powerful leg muscles means they can act as a useful analogy for bipedal species of dinosaur. Research on ostrich running mechanics led by Nina Schaller at the Senckenberg Research Institute in Frankfurt, Germany, hints that very large species of feathered dinosaur, such as the 8-metre Gigantoraptor, may have used feathered forelimbs to counteract their great bulk as they ran, allowing them to move much more nimbly than we might expect for an animal of their size.
Successful flight
What Ken Dial’s research tells us is that we don’t have to choose between the two theories: ground-up or trees-down. Perhaps the first fliers developed feathers as an advantage in running and leaping, which also allowed them to run right up into the trees and then glide back out of them to the ground again. Feathers had begun as insulation, gone on to a role in camouflage and display and, finally, much later been co-opted as aerofoils for flight.
We can also see now that the distinction between dinosaurs and birds is utterly blurred. There is a tight knot of dinosaur species somewhere near the origin of birds: dromaeosaurs and troodontids that had evolved small size, long flight feathers on their arms and legs, and at least some limited flying ability. In some sense what we now define as a dinosaur and as a bird is arbitrary, but many experts draw the line at full powered flight. Archaeopteryx is believed to have had this – just – but it still eluded close relatives, such as Xiaotingia and Anchiornis.
By all accounts, early birds such as Archaeopteryx and Confuciusornis would have been abysmal fliers, with their heavy skeletons, poor flight strokes and, in the case of Archaeopteryx, long bony tail. These traits may have made it more difficult for them to get airborne and rendered them cumbersome operators once in the air. Aerodynamic agility would come later. Nevertheless, these feathered dinosaurs had finally made it to the skies.
This opened up a whole new world of opportunity for them – one that would carry them through the mass extinction that wiped out their land-bound relatives at the end of the Cretaceous, and allow them to prosper and diversify into the most numerous and successful group of vertebrates that has ever lived. As you read this, an estimated 400 billion individual feathered dinosaurs, of 10 000 species, can be found on earth, in almost every habitable environment. You need only step outside and look up into the trees and the wide blue skies to find them.