As little as 50 years ago, the theory that birds descended from dinosaurs seemed completely ridiculous--after all, everyone knows that most birds are small, light, fluttery creatures, while most dinosaurs were huge, plodding, and distinctly unaerodynamic. But as the evidence--small dinosaurs possessing feathers, beaks, and other birdlike characteristics--began to mount, the connection between dinosaurs and birds became apparent to scientists, and then to the general public. Today, it's the rare paleontologist who disputes the descent of birds from dinosaurs, though there are some outliers who try, and we're left to explain why birds aren't dinosaur-sized.
This doesn't mean, however, that all of the technical aspects of the dinosaur/bird transition have been settled once and for all. Researchers still disagree about which families of dinosaurs were most closely related to modern birds, whether the feathers of these dinosaurs were aerodynamic or ornamental, and--perhaps most contentiously of all--how these reptilian proto-birds managed to achieve the huge evolutionary leap into powered flight.
The Origin of Feathered Dinosaurs
Why, and how, did the small theropod dinosaurs of the Jurassic and Cretaceous periods evolve feathers? It's a common mistake among those unversed in evolutionary theory to assume that feathers evolved specifically for the purpose of flight. Evolution, however, is a blind process--it doesn't "know" where it's going until it gets there. For this reason, the most widely accepted explanation today is that dinosaurs evolved feathers as a means of insulating themselves in cold climates (and, possibly, as a way to puff themselves up in the eyes of the opposite sex with coats of garish plumage).
If this sounds unlikely, bear in mind that even birds that have been flightless for millions of years, like ostriches and emus, still retain their feathers, an expensive accessory in terms of energy usage. If the purpose of feathers was solely to power flight, there would be no reason, from an evolutionary perspective, for penguins to keep these appendages: in fact, they might be better off completely naked, or sporting thick coats of fur! (For more on this subject, see Why Did Dinosaurs Have Feathers?)
The first indisputably feathered dinosaurs--like Archaeopteryx and Epidendrosaurus--appeared on earth during the late Jurassic period, anywhere from 160 to 150 million years ago. As the eons ground on, the primitive (that is, short and hairlike) feathers of these early dino-birds gradually evolved into the broad, flat feathers we're familiar with today, which are better suited to trapping air (and thus insulating the underlying skin). At this point the question asks itself: how did these feathered dinosaurs make the transition to flight?
Theory #1: Feathered Dinosaurs Took a Running Leap Into Flight
Extrapolating backward from the behavior of some modern birds, it's reasonable to infer that the small- to medium-sized, two-legged theropods of the Cretaceous period (notably the ornithomimids, or "bird mimics," but also raptors and possibly even small tyrannosaurs) could attain top running speeds of 30 or 40 miles per hour. As these theropods ran (either in the act of chasing down prey or trying to escape being eaten themselves), their coat of insulating feathers gave them a slight aerodynamic "bounce," helping them land their next meal or live to see another day. Since well-fed dinosaurs, and those that avoided predation, produced more offspring, the evolutionary trend was toward larger feathers, which provided more "lift."
From there, the theory goes, it would only have been a matter of time before a feathered dinosaur achieved actual flight, at least for a brief period of time. But at this point, it's important to understand what "a short time" means in an evolutionary context. There wasn't a single defining moment when a small, feathered theropod accidentally ran straight off the side of a cliff and magically took flight like a modern bird. Rather, you have to picture this process happening incrementally, over the course of millions of years--leaps of four feet, five feet, ten feet, until something resembling powered flight gradually emerged.
In the excellent Nova episode The Four-Winged Dinosaur (about a specimen of Microraptor that had recently been discovered in China), a paleontologist is quoted as saying that the hatchlings of modern birds tend to recapitulate their evolutionary heritage. That is, even though these newly hatched chicks are unable to fly, they can jump for farther distances, and more easily scuttle up inclined surfaces, with the aerodynamic lift provided by their feathers--the same advantages as may have been enjoyed by the feathered dinosaurs of the Jurassic and Cretaceous periods.
Theory #2: Feathered Dinosaurs Achieved Flight by Falling Out of Trees
The trouble with Theory #1 is that birds aren't the only animals alive today whose behavior can be extrapolated back to extinct dinosaurs. Flying squirrels, for instance, glide across forest canopies by leaping off the tall branches of trees and spreading the flaps of skin attached to their arms and legs. They're not capable of powered flight, of course, but they can glide for impressive distances, up to two-thirds of the length of a football field for some species. (Another family of gliding and flying animals is the pterosaurs, which were only distantly related to dinosaurs and not directly ancestral to modern birds.)
Conceivably, some types of feathered dinosaurs might have lived high up in trees (which would entail their being a relatively small size and having the ability to climb). These theropods, the reasoning goes, might then have followed the same evolutionary path as flying squirrels, gliding for longer and longer distances from branch to branch, or from tree to tree, as their feathers slowly evolved to the optimum shape and configuration. Eventually, they could leap off a high branch and take to the air for indefinite periods of time, and voila - the first prehistoric birds!
The main problem with this "arboreal" theory of flight, as it's called, is that it's easier to imagine powered flight evolving in the ground-up scenario (picture a terrified dinosaur desperately flapping its vestigial wings while trying to escape a ravenous Allosaurus) than as a result of tree-to-tree gliding. We also have indirect evidence against this scenario, which is that, despite millions of years of evolution, no flying squirrel (with the exception of Bullwinkle's pal Rocky) has managed to achieve powered flight--although, to be fair, bats certainly have. More to the point, though, paleontologists have adduced absolutely no fossil evidence for tree-dwelling dinosaurs.
Current Thinking About Feathered Dinosaurs and Birds
New genera of small, feathered dinosaurs are constantly being discovered, many of them in China. Since these dinosaurs date back to different geologic times ranging from the Jurassic to the Cretaceous, separated by tens of millions of years, it can be difficult for paleontologists to reconstruct the exact evolutionary line that led from dinosaurs to birds. For example, the weird, four-winged Microraptor has provoked intense debate: some researchers see it as an evolutionary dead end, others as an "intermediate" form between dinosaurs and birds, and yet others as not technically a dinosaur at all, but an offshoot of the archosaur family tree that predated the rise of the dinosaurs.
Further complicating matters, it's possible that birds evolved not once, but multiple times during the Mesozoic Era. (This type of "convergent evolution" is fairly common; it's why, for instance, modern giraffes mimic the body shape of hundred-million-year-old sauropods). Some of these birds may have achieved flight runway-fashion, others by falling out of trees, and still others by some bizarre combination of the two. All we can say for sure is that all modern birds derive from one common ancestor; that is, if birds did indeed evolve multiple times during the age of dinosaurs, only one of these lines managed to survive into the Cenozoic Era.