by Brian Switek
A mix of fossil and molecular evidence hints at how feathers could have evolved. Birds are the living descendants of the coelurosaurs, and crocodylians are the closest living relatives to dinosaurs as a whole, so features shared between birds and crocodylians might have been present in the last common ancestor of both lineages (and therefore also present in dinosaurs). Both birds and alligators, for example, share the regulatory proteins sonic hedgehog (abbreviated Shh, and named for the video game character) and bone morphogenetic protein 2 (BMP2-), both of which underlie the formation of both the scales of alligators and the feathers of birds. Hence it is likely that, during the evolution of dinosaurs, these proteins were co-opted from their roles in forming the tough hides of dinosaurs into the creation of feathers.
The diversity of feather types among coelurosaurs suggests how feathers were modified once they had begun to evolve. As seen in Sinosauropteryx, the earliest feathers were simply tubes that grew from the skin. Once these structures evolved there would have been enough variation for them to split and become branched, something that has been observed in the downy covering of baby chickens, with each feather providing greater coverage on the animal. From there, the branching filaments could be organized along a central vane, like what is seen in Caudipteryx and Sinornithosaurus. After this point, little barbs branched off from each filament along the shaft, locking them together and stiffening the feather. This was the kind of feather needed for flight, and it is what is seen in most modern birds. That these structures are feathers and not just degraded collagen or some other quirk of fossilization is beyond reasonable doubt.
The majority of dinosaur fossils are just bones and teeth, and even fossilized skin impressions only preserve patterns, not colors. But scientists have recently discovered that there is a way to detect some colors in the fossil record. While studying an exceptionally preserved fossil, squid paleontologist Jakob Vinther saw that its ink sac was packed with the same kind of microscopic spheres that give the ink of living squid their color. These bodies are called melanosomes, and once Vinther realized that they could be preserved in the fossil record he began to wonder what other prehistoric remains might contain them.
One of the first tests was on the forty-seven-million-year-old feather of an extinct bird from Messel, Germany (home of “Ida” and not far from the final resting place of Archaeopteryx). Since the feather seemed to show light and dark bands, it was a good test case to see whether the bodies were truly pigment-carrying melanosomes (in which case they would only be found in the dark bands) or were just bacterial remnants scattered all over the feather. The results were better than could have been expected. In 2009 the researchers behind the study announced that not only did the feather most certainly contain melanosomes in the dark bands, but their arrangement corresponded to a pattern seen in living birds that gives feathers a glossy sheen. This was better than just an isolated discovery. It presented paleontologists with a new technique and two teams, working independently of each other, turned to the fossilized feathers of dinosaurs to see if they, too, contained remnants of color.
The first team, lead by Fucheng Zhang, published their results in the journal Nature on January 27, 2010. They had turned their attention to two of the first feathered dinosaurs to be found, Sinosauropteryx and Sinornithosaurus. Feather samples from both contained two different types of melanosomes; those that created dark shades (eumelanosomes) and those associated with reddish hues (phaeomelanosomes). This allowed the scientists to speculate that Sinosauropteryx had a garish red-and-white-striped tail, which might have been used to signal to other members of its species.
Vinther and his team published their own findings in Science the following week. Building on the previous research on the fossil bird feather, they attempted to present a specimen of the recently discovered dinosaur Anchiornis in Technicolor. After determining the pattern of melanosome distribution throughout the feathers they compared the arrangements to what is seen in living birds to restore the long-lost pigments. As it turned out most of the feathers of Anchiornis were black, but they were set off by white accents on its wings and a plume of rufous feathers on top of its head. Even though the study did not look for chemical traces of color in the fossil that would have marked the presence of other shades, for the first time the researchers were able to produce an image of an entire living dinosaur.
FIGURE 38 - Restorations (not to scale) of Anchiornis and Microraptor, based upon exceptional specimens that also preserved feathers. The discovery of such fossils has overwhelmingly confirmed that birds evolved from dinosaurs.
The question of just what a feather is has become more complicated, however. Very early on in dinosaur evolution there was a split in the dinosaur family tree that resulted in the evolution of the ornithischians (containing an array of herbivorous dinosaurs such as the ankylosaurs, hadrosaurs, and ceratopsians) and the saurischians (comprised of the predatory theropods and the forebears of the gigantic, long-necked sauropods).42 The presence of feathers in coelurosaurs alone suggested that fuzzy body coverings had evolved only once among dinosaurs within the saurischian side of the split, but at the beginning of the twenty-first century scientists found similar structures among ornithischian dinosaurs. In 2002 Gerlad Mayr and colleagues announced that they had discovered a specimen of the ceratopsian Psittacosaurus with long, bristlelike structures growing out of its tail and it was joined in 2009 by Tianyulong, another bristle-covered ornithsichian described by a team of researchers led by Zheng Xiao-Ting.
These animals were about as far removed from bird ancestry as it was possible to be while still remaining a dinosaur, yet they were covered in structures similar to the proto-feathers of Sinosauropteryx. Either the filamentous body covering evolved twice in two different groups of dinosaurs, or, even more spectacularly, was a common di-nosaur trait later lost in some groups. Regardless of how many times “dino fuzz” evolved, however, these structures were only adapted into true feathers among the coelurosaurs, but how flight evolved is another evolutionary mystery.
FIGURE 39 - A Styracosaurus, covered in bristles, scavenges the body of a dead tyrannosaur. The discovery that ornithischian dinosaurs like Tianyulong and Psittacosaurus had bristlelike structures growing out of their skin suggests that it is possible that many other ornithischian dinosaurs did, as well.
John Ostrom presented one hypothetical scenario in 1979. Inspired by his work on Deinonychus and Archaeopteryx, he proposed that the ancestors of the first bird were small coelurosaurs covered in rudimentary feathers. With their grasping hands, these tiny predators would have been adept hunters of flying insects, and their simple feathers would have provided an unexpected advantage. The feathers along their arms would have helped trap insects, and so longer feathers would have been selected for over time. Eventually these “proto-wings” would have allowed the dinosaurs a little bit of extra lift while jumping after their prey, and this shift in selection would precipitate the origin of the first flying birds.
Ostrom’s “insect-net hypothesis” never truly took off, as it was marred by functional problems surrounding how feathers might be used as a net, but it did reignite an old debate about whether flight evolved from the “trees down” or the “ground up.” According to the advocates of the arboreal hypothesis, small feathered dinosaurs climbed up into trees and launched themselves into the air to glide a short distance, and eventually they would be adapted to beat their wings to truly fly. The four-winged dinosaur Microraptor, a relative of Deinonychus, has most recently been taken to throw support to this idea, as it may have launched itself out of trees to glide, if not truly fly, through the forest.
Other paleontologists have preferred one version or another of the cursorial hypothesis. In this view, feathered dinosaurs ran along the ground, perhaps hopping into the air after insects or other prey, until by some mechanism they developed the ability to actually fly. In fact, feathered arms may have even made some dinosaurs better runners. A key piece of evidence for this hypothesis co
mes from chukar partridges. These birds are capable of flight, but if they need to escape into a nearby tree or over a natural obstacle they often run rather than fly, flapping their wings as they do so. As discovered by scientist Kenneth Dial this technique gives the birds better traction while running, so much so that they can run right up vertical inclines. As hypothesized by Dial, feathered dinosaurs could have gained a functional advantage by flapping their arms while running (be it after prey or to avoid becoming prey), and this behavior could then be co-opted to allow them to start flying.
As recognized by most working paleontologists today, however, the old arboreal versus cursorial dichotomy is no longer helpful. Much like Williston, Nopsca, and Beebe, we can create numerous plausible scenarios but, without knowing which feathered dinosaurs were the root stock from which birds evolved, any origin-of-flight hypothesis must be regarded as provisional right from the start. Even as the numerous feathered fossils have confirmed that birds evolved from dinosaurs, they have also made the relationships between those fossils and birds much more complex. At one time it seemed that Velociraptor and its relatives were the closest relatives of early birds, but a little-known group of recently discovered forms may be even closer.
Described in 2002, the small feathered dinosaur Scansoriopteryx was one of the most bizarre coelurosaurs ever found. With large eyes, a short snout, and a very long third finger, this sparrow-sized dinosaur was unlike many of its coelurosaur cousins. Its description was followed in 2008 by the announcement of a close relative named Epidexipteryx, a pigeon-sized dinosaur, covered in fuzz, that also sported two pairs of ribbonlike feathers on its shortened rump and a mouth full of forward-oriented teeth. Given that they may be older than the earliest birds, they could represent the kind of dinosaur birds evolved from, in which case the Velociraptor and its relatives would be further removed from the origin of birds than had been previously supposed.
For over a century Archaeopteryx was the key to understanding bird origins, as it was the oldest bird ever discovered, but as more feathered dinosaurs have been found the connection between Archaeopteryx and other fossil birds has become looser. As the delineation between non-avian dinosaur and bird has become increasingly blurred it has become difficult to tell what side Archaeopteryx falls on. As research continues, it may turn out that Archaeopteryx was, like Microraptor, a feathered dinosaur and not a true bird.
FIGURE 40 -- A drawing of the skeleton of Epidexipteryx, denoting the “halo” of feathers around the skeleton and the pairs of elongated feathers coming out of its tail. It may be one of the closest relatives of early birds.
The unstable relationships of some of the feathered dinosaurs was exemplified by the redescription of Anchiornis huxleyi in 2009. The fossil, named in honor of T. H. Huxley’s work on bird origins, had been announced the year before as the closest dinosaurian relative of birds and, at thirty million years older than Archaeopteryx, was especially significant. When a better-preserved specimen was found, however, the scientists realized that their initial hypothesis was wrong. Anchiornis was actually a troodontid, or a member of a group of coelurosaurs closely related to the famous “raptors,” yet it was very similar in form to Archaeopteryx.
Even if Archaeopteryx is dethroned from the vaunted position of “earliest known bird” that Richard Owen bestowed upon it, the fact remains that birds evolved from dinosaurs, and much more than fossilized feathers supports this hypothesis. During the 1920s the explorer Roy Chapman Andrews led a string of expeditions for the American Museum of Natural History into Mongolia’s Gobi Desert to search for the evolutionary center of origin for all mammals (including humans). No evidence of a mammalian Eden was found, but the excursions did return with the ghostly white bones of the Cretaceous dinosaurs Velociraptor, Protoceratops, and Oviraptor, the latter of which was especially fascinating because it was found in the act of robbing a Protoceratops nest.
But in 1994 it was announced that the wrong dinosaur was inside the supposed Protoceratops eggs. Instead of an embryonic horned dinosaur, there was the miniscule skeleton of a developing theropod much like Oviraptor. The specimen the Andrews expedition had found was probably caring for its own eggs, not robbing the eggs of others. The discovery of several skeletons of a crested Oviraptor relative named Citipati which were found sitting atop nests of the same kind of eggs, supported this hypothesis. Their arms encompassed the sides of the nest in a position only seen in birds, and the close relationship between Citipati and the feathered Caudipteryx opened the possibility that these dinosaurs, too, were covered in feathers that they used to regulate the temperature of their nests. This discovery of fossilized behavior dovetailed beautifully with the numerous feathered coelurosaurs, and the description of the tiny troodontid Mei long in 2004 also surprised paleontologists. Like the skeletons of Citipati on their nests, several of these dinosaurs were suddenly killed and buried while sleeping, perfectly preserved in the position in which they died. They were curled up just like slumbering birds.
FIGURE 41 - The skeleton of the troodontid dinosaur Mei long, with a line drawing identifying the visible bones on the right. Abbreviations: cev, cervical vertebrae; cv, caudal vertebrae; dv, dorsal vertebrae; lh, left humerus; lr, left radius; lu, left ulna; pg, pelvic girdle; rh, right humerus; rr, right radius; ru, right ulna; sk, skull.
Even the unique breathing system seen in modern birds first appeared long before their ancestors first took to the air. As you relax reading this book you go through a breathing cycle of inhaling and exhaling. When you inhale, air enters your lungs (where oxygen is absorbed), and when you exhale the carbon dioxide-rich, oxygen-depleted air is forced out. Unlike you, however, birds lack a diaphragm and cannot inflate or deflate their lungs. Instead birds have a “one way” breathing system in which fresh air moves through their respiratory system both when the bird inhales and exhales. This is made possible by a series of anterior and posterior air sacs that can expand and contract. This is a more efficient way of getting oxygen from the air, but these air sacs also have a structural benefit. They arise from the lungs and invade the surrounding bones, thus making birds lighter. This infiltration into the bone leaves telltale hollows and indentations on bones, which have been seen in dinosaurs for over one hundred and fifty years.
It might not come as a surprise that coelurosaurs have evidence of air sacs on their bones, but what is interesting is that other saurischian dinosaurs shared the same feature. This makes sense given the evolutionary history of these dinosaurs. There is no sign that air sacs were present in the ornithischian dinosaurs, but the evidence for air sacs in saurischian dinosaurs goes all the way back to one of the earliest presently known. Called Eoraptor, this small bipedal dinosaur was not unlike Compsognathus, and it may be a fair approximation of what some of the earliest saurischian dinosaurs were like. Its bones were marked by indentations that indicate that it had at least some rudimentary air sacs, and later predatory dinosaurs from the coelurosaurs to the knobbly-headed abelisaur Majungasaurus and the Allosaurus-relative Aerosteon had even better developed air sacs.
FIGURE 42 - A diagram of the air sacs inside a bird. Abbreviations: atas, anterior thoracic air sac; cas, cervical air sac; clas, clavicular air sac; hd, humeral diverticulum of the clavicular air sac; lu, lung; pns, paranasal sinus; ptas, posterior thoracic air sac; pts, paratympanic sinus; t, trachea.
FIGURE 43 - A reconstruction of the skeleton of Majungasaurus, showing the placement of air sacs within the body inferred from pockets in its bones. Although Majungasaurus was not closely related to birds, the presence of these structures in its skeleton shows that these features were widespread among saurischian dinosaurs.
The other great saurischian dinosaur group, the sauropods, also had bones infiltrated by air sacs. If you tried to design an animal like a 100-foot-long sauropod with thick, heavy bones in its neck, it would have been unable to lift its head. Much like a bridge, their skeletons reflect the selective pressures for strength and lightness, and air sacs allowed them to
achieve this. They probably inherited this trait from their last common ancestor with the theropod dinosaurs.
While not exactly like those seen in living birds, the air sacs in many of these saurischian dinosaurs may have also provided them physiological benefits. Air sacs may have initially been selected because they lightened the skeleton, but if they provided dinosaurs with more efficient breathing (allowing them to be more active, for example) there would have been additional benefits for natural selection to act upon. Research into this area is still new, but it is clear that rudimentary air sacs appeared in dinosaurs seventy-five million years before Archaeopteryx, long preceding the first birds.
Some dinosaurs were even plagued by parasites that now infest the mouths of living birds. From healed wounds on skulls paleontologists have known for years that large predatory dinosaurs bit each other on the face during combat. Tyrannosaurs, especially, showed scars from such conflicts, but many Tyrannosaurus jaws often had holes in the lower jaw not apparently caused by the teeth of a rival. When paleontologists Ewan Wolff, Steven Salisbury, Jack Horner, and David Varricchio took another look at the jaws of tyrannosaurs that had these holes they did not find any sign of infection, inflammation, or healing that would be expected if the dinosaurs had been bitten. Bone, after all, is living tissue, and would slowly remodel itself in the wake of an injury. Instead, the holes were smooth, as if the bone was being slowly eaten away.
It seemed more likely that the holes were the result of some kind of pathology, and the researchers found that the sores were consistent with damage done by a single-celled protozoan called Trichomonas gallinae that infests modern birds. When inside living birds this microscopic creature causes ulcers to form in the upper digestive tract and mouth of the host, virtually identical to the damage seen in the Tyrannosaurus jaws. The species of protozoan that afflicted Tyrannosaurus might have been only a close relative of the living kind, but this was the first evidence of an avian disease afflicting dinosaurs.