The Origin of Flight

Evolutionary innovations such as the ability to fly do not appear from the first in their finished form; they must progress through intermediate developmental states that are themselves adaptive. If a design does not work for some purpose, it is unlikely to last long enough to be modified further by evolutionary influences. Nevertheless, transition innovations do not necessarily serve the same function as their ultimate forms. For example, feathers evolved from the scales of birds' reptilian ancestors. But long before the complex structure of flight feathers came about, simpler designs functioned perfectly well to insulate small, active protobirds (the earliest form of birds) from the cold. A feathered body covering likely evolved along with endothermy; maintaining a high body temperature requires a lot of food, so there is considerable benefit to reducing heat loss. A myriad of small feathered dinosaurs, including Sinocalliopteryx and Sinosauropteryx, have recently been unearthed in fossil beds in China. These reptiles did not use their feathers for flight because their feathers lacked the rigid structure of vaned feathers (which make up modern birds' outer coat) and more closely resembled the down feathers that modern birds wear under their outer coat. Hence it follows that protobirds did not just leap into the air one day to become flying vertebrates. So how did it happen?

The truth may ultimately be explained by one of two proposed hypotheses, or more likely a combination of both. The arboreal hypothesis, first championed by the paleontologist Othniel C. Marsh, suggests that birds started out as tree-climbing animals, then later evolved wings to assist in controlled glides to the ground or to lower branches. The premise that flight originated in tree-dwelling animals has the advantage that gravity would initiate takeoff, which is perhaps the most difficult part of flying. However, proponents of this hypothesis fail to explain how complex feathered wings could evolve from structures used primarily for gliding. Other gliding vertebrates-for example, flying squirrels and sugar gliders-use sail-like skin folds extended between their outstretched limbs to glide between trees. Draco lizards extend fanlike structures that are supported by elongated ribs. Birds, however, fly by flapping their front limbs; gliding membranes along the length of the body are simply not conducive to the evolution of feathered wings.

The cursorial, or running, hypothesis (first developed by Samuel Wendell Williston) suggests that birds began as small, two-legged running dinosaurs that used their front limbs to catch insects or other small prey and that flight began as a series of short jumps into the air. Any increase in the surface area of the forelimbs, perhaps by elongation of the feathers already covering them, would increase the height of the jump and overall stability by providing a little more upward force, or lift, on the bird. Eventually arms would become wings and jumps would lead to flight. Research simulations suggest that this running-and-jumping mode of foraging could indeed yield more food captures. It is generally accepted that birds' ancestors were small, agile dinosaurs-characteristically two-legged species such as Velociraptor-so this hypothesis makes sense. However, because of the added resistance when sweeping them through the air, some paleontologists doubt that oversized feathered front limbs would offer an advantage for catching prey. They also feel that the ground speed required for takeoff would exceed the maximum running speeds of existing lizards and running birds by a factor of three. It is possible, however, that protobirds used wind speed to increase lift during takeoff by moving into the wind as it blew against them, as aircraft do.

More recently, paleornithologists have proposed a third model-the pouncing proavis hypothesis-which combines some elements of the previous two. This model describes protobirds as small, active ambush predators that hid on elevated perches such as shrubs or boulders, then jumped or swooped down to capture prey. Additional lift would allow longer and more controlled swoops, particularly if feather and wing development focused on the hands, the location of primary flight feathers responsible for forward thrust in modern birds. This hypothesis implies a gentler, more gradual approach to the evolution of flight; even the intermediate stages of feather and wing development would provide some adaptive advantage to the animal sporting them.