Paleobiology and the origins of avian flight

When interpreting the paleobiology of long extinct taxa, new fossils, and reinterpretations of well-known fossils, sharply at odds with conventional wisdom never seem to cease popping up. Given the vagaries of the fossil record, current notions of near resolution of many of the most basic questions about long-extinct forms should probably be regarded with caution. Even major aspects of the paleobiology of intensely studied, recently extinct taxa (<10,000 yrs.) remain unresolved [e.g., the cause(s) of late Pleistocene large-mammal disappearances (1) and the lifestyles of specialized forms (sabre-tooths) (2)]. Little wonder then that so fascinating a subject as the origins of birds and bird flight, both of which almost surely occurred more than 150 million years ago, have stirred such publicly visible and intense, nearly century-long, controversies (3). In this context, perhaps the most hotly debated and fundamental issue relates to the origin of avian flight: Did proto-birds take to the air as arboreal gliders where gravity was the primary source of flying energy (from the “trees-down”), or as swift, cursorial ground dwellers (from the “ground-upward”)? In PNAS, Alexander et al. (4) describe results of a novel “paleobiological” study: three-dimensional model-based, empirical data that provide important insight into this old question.

Although an affinity between theropod dinosaurs and birds has long been recognized, for most of the 20th century, paleontologists generally accepted that an ancestor–descendant relationship between these groups was unlikely (3). Among the traits distinguishing the two taxa, the earliest known bird, Archaeopteryx (Late Jurassic Period, ≈150 m.y.), seemed to have been primarily arboreal; known theropods were exclusively cursorial. Avian flight was widely thought to have originated in gliding, arboreal proto-birds (i.e., the trees-down scenario). That accepted wisdom was turned on its head in the 1970s when Yale’s John Ostrom reanalyzed the skeletons of Archaeopteryx and the cursorial dromeosaurine theropods (“raptors”) (5). He concluded that, in fact, birds were closely related to theropod dinosaurs and were more or less directly derived from the raptors. Thereafter, Ostrom’s scenario for avian origins was bolstered by cladistic analyses consistent with a dromaeosaur-bird lineage (6) and with the discovery of apparent feathers, or feather-like structures, on a number of late Mesozoic Era theropods (7).

Concomitant with shifting conventional wisdom that birds were derived from cursorial theropods, many late 20th century authors, including John Ostrom, advocated a cursorial, or ground-upward (as opposed to a trees–down), origin for powered flight in birds (8). Throughout this period, skeptics pointed to substantial aerodynamic difficulties associated with the cursorial, ground-upward hypothesis (9). Nevertheless, by the beginning of this century, extant birds were described, in many scientific and lay communities, as “living dinosaurs,” and the ground-upward scenario had, in many circles, become accepted wisdom (10).

Not so fast. In 2003, Xu et al. (11) described a fully feathered specimen of the small basal dromaeosaur Microraptor that bore four wings of fully modern, asymmetrical flight feathers on its forelimbs and legs. As Xu stated: The metatarsus [hindlimb] feathers are inconsistent with the suggestions that basal dromaeosaurs are cursorial animals because such long feathers on the feet would be a hindrance for a small cursorial animal. It is unlikely that a small dromaeosaur could run fast with such an unusual integument and this provides negative evidence for the ground-up hypothesis for the origin of avian flight. Xu et al. reconstructed limbs of four-winged Microraptor as tandem wings like those of insects and gliding fish where, during flight, all wings are spread horizontally in a “tetrapteryx” fashion. They concluded that Microraptor was probably a tree-dweller, thus supporting a trees-down origin for avian flight. Given the nature of the controversy, perhaps it is not overly surprising that the recently discovered Microraptor, as well as Xu et al.’s interpretation of its tetrapteryx flight posture, is eerily similar to naturalist William Beebe’s 1915 hypothetical tetrapteryx stage in the origin of avian flight (Fig. 1) (12).

Fig. 1.

Naturalist William Beebe’s hypothetical, intermediate gliding stage in the origin of birds and bird flight (modified from ref. 12). Beebe’s proposal appeared in 1915 (12) but was almost prescient in its morphological similarity to the recently discovered Microraptor gui (11) and that animal’s likely tetrapteryx gliding posture [see Alexander et al. (4)].

Nevertheless, advocates for a cursorial, ground-upward origin (“aerial cursors”) for avian flight remain. They assert that hindlimb flight feathers in Microraptor were somehow oriented in such a manner that interference with running would have been minimized. Supposedly, flight in a cursorial Microraptor would necessarily have been biplane-like, wherein the hindlimb wings/feathers (but not the skeleton) would have been horizontally situated, but well below and somewhat behind those of the forelimb (13).

Clearly, these two interpretations of Microraptor are broadly at odds with one another and imply profoundly different scenarios for the origin of avian flight. Given that direct observation of living Microraptor is not on the horizon, Alexander et al. (4) present results from what is probably the best available approach to provide some empirical resolution to this question: Glide tests were performed by using three-dimensional life-size scale models on the skeleton of one individual that had been prepared “in the round.” Glide tests of the model were conducted with hindwings abducted and extended laterally (tetrapteryx), as well as with the previously described biplane configuration. Although the biplane model glided almost as well as the other model, it required an anatomically unlikely, especially massive, and ungainly head for aerodynamic stability.

So, is the answer, after all, a hybrid of the two old theories, i.e., avian origins from an arboreal, gliding theropod dinosaur? Perhaps, but then this is paleobiology—very recent data suggest that many clearly cursorial theropods previously thought to have been feathered may not have been so (14) and that dromaeosaurs, the group that birds are assumed to have been derived from, may not even have been dinosaurs (15). What pops up next is anyone’s guess.