A meteor stream is composed of dust particles that have been ejected from a parent comet at a variety of velocities. These particles follow the same orbit as the parent comet, but due to their differing velocities they slowly gain on or fall behind the disintegrating comet until a shroud of dust surrounds the entire cometary orbit. Astronomers have hypothesized that a meteor stream should broaden with time as the dust particles' individual orbits are perturbed by planetary gravitational fields. A recent computer-modeling experiment tested this hypothesis by tracking the influence of planetary gravitation over a projected 5,000-year period on the positions of a group of hypothetical dust particles. In the model, the particles were randomly distributed throughout a computer simulation of the orbit of an actual meteor stream, the Geminid. The researcher found, as expected, that the computer model stream broadened with time. Conventional theories, however, predicted that the distribution of particles would be increasingly dense toward the center of a meteor stream. Surprisingly, the computer model meteor stream gradually came to resemble a thick walled, hollow pipe.

Whenever the Earth passes through a meteor stream, a meteor shower occurs. Moving at a little over 1,500,000 miles per day around its orbit, the Earth would take, on average, just over a day to cross the hollow, computer-model Geminid stream if the stream were 5,000 years old. Two brief periods of peak meteor activity during the shower would be observed, one as the Earth entered the thick-walled "pipe" and one as it exited. There is no reason why the Earth should always pass through the stream's exact center, so the time interval between the two bursts of activity would vary from one year to the next.

Has the predicted twin-peaked activity been observed for the actual yearly Geminid meteor shower? The Geminid data between 1970 and 1979 show just such a bifurcation, a secondary burst of meteor activity being clearly visible at an average of 19 hours(1,200,000 miles)after the first burst. The time intervals between the bursts suggest the actual Geminid stream is about 3000 years old.

    Two opposing scenarios, the "arboreal" hypothesis and the "cursorial" hypothesis, have traditionally been put forward concerning the origins of bird flight. The "arboreal" hypothesis holds that bird ancestors began to fly by climbing frees and gliding down from branches with the help of incipient feathers: the height of trees provides a good starting place for launching flight, especially through gliding. As feathers became larger over time, flapping flight evolved and birds finally became fully air-borne. This hypothesis makes intuitive Sense, but certain aspects are Troubling. Archaeopteryx (the earliest known bird) and its maniraptoran dinosaur cousins have no obviously arboreal adaptations, such as feet fully adapted for perching. Perhaps some of them could climb trees, but no convincing analysis has demonstrated how Archaeopteryx would have both climbed and flown with its forelimbs, and there were no plants taller than a few meters in the environments where Archaeopteryx fossils have been found. Even if the animals could climb trees, this ability is not synonymous with gliding ability. (Many small animals, and even some goats and kangaroos, are capable of climbing trees but are not gliders.) Besides, Archaeopteryx shows no obvious features of gliders, such as a broad membrane connecting forelimbs and hind limbs.

    The "cursorial"(running) hypothesis holds that small dinosaurs ran along the ground and stretched out their arms for balance as they leaped into the air after insect prey or, perhaps, to avoid predators. Even rudimentary feathers on forelimbs could have expanded the arm's surface area to enhance lift slightly. Larger feathers could have increased lift incrementally, until sustained flight was gradually achieved. Of course, a leap into the air does not provide the acceleration produced by dropping out of a tree; an animal would have to run quite fast to take off. Still, some small terrestrial animals can achieve high speeds. The cursorial hypothesis is strengthened by the fact that the immediate theropod dinosaur ancestors of birds were terrestrial, and they had the traits needed for high lift off speeds: they were small, agile, lightly built, long-legged, and good runners. And because they were bipedal, their arms were free to evolve flapping flight, which cannot be said for other reptiles of their time.