Give a riderless bike a push, and chances are it will cruise along for a bit without tipping over. Scientists have long thought they understood the physics of what keeps the bike upright. But a new study shows that much of what they believed is wrong.
Researchers assumed that two mechanisms kept a riderless bike upright. First is the "caster effect," which relies on the position of the front wheel relative to the steering axis—the imaginary line that extends from the front forks downward. On most bikes, the front wheel meets the ground just behind this axis, so it tends to center itself like the casters on a shopping cart. The other mechanism is known as gyroscopic precession: Because the front wheel is spinning forward, any tilt to the left or right creates a force that will steer the bike in that direction, pulling it out of the fall.
To show that neither of these mechanisms is in fact crucial, engineer Andy Ruina of Cornell University and colleagues used a computer to design a bicycle that excludes both mechanisms. Unlike a normal bike, the front wheel of their vehicle lies in front of the steering axis, so the caster effect can't operate. On top of that, both front and back wheels are connected to duplicate wheels spinning in the opposite direction, so any gyroscopic precession is canceled out.
The contraption doesn't resemble a normal bike. The wheels are small, and it sports heavy, metal weights close to the ground in front and atop a tall pole jutting up where the seat would normally be. "It looks like a funny scooter," Ruina says.
Once the researchers had built the bike—described  online today in Science—they gave it a push and discovered that it stayed upright almost until it came to a halt, correcting any leaning just as a normal bike does. Ruina's conclusion: "Gyro effects are important contributors to self-steering [and] so are caster effects. It's just that they are not essential."
So what does keep the bicycle from falling over? The trick is its odd mass distribution: low in front and high in the rear. As a result, the front will try to fall faster than the rear, just as a short pencil topples faster than a tall broomstick. Because front and rear are connected, this lopsided falling makes the front steer into the fall and in doing so pulls the bike out of it.
Joel Fajans, a physicist at the University of California, Berkeley, who has studied bike stability, calls the work a "very thorough, well-done study." The most surprising result, he says, is that the caster effect isn't necessary, as previous experiments had already suggested that gyroscopic precession isn't as important as many people think. "It's possible that these results might influence bicycle design," he says.
Ruina and his colleagues agree that their findings could help professional bike engineers to explore other strange-looking designs that are just as stable. "With these changes of appearance, one might have bicycles that are easier for old people to ride, bicycles that are easier to maneuver but still feel stable when riding straight, bicycles that have a more comfortable sitting arrangement, or bicycles that have more convenient places for storage or carrying kids," Ruina says.