For the interested, this is from a program called Radiolab, which you can listen to online. It's really worth a listen.

Oh, good, now I have another unsolvable question that will rattle around in my brain until I die.

Neat stuff.

Serious question: does the same apply for blind people?

If you're a hunter-gatherer, off in the jungle or forest or savannah, with no maps, and you set off hunting and gathering, you'd like to end up at the end of your trip back roughly where you started (or at least somewhere you're familiar with), so you don't have to backtrack. That way you can hunt & gather for your entire trip rather than just the first half.

Alternatively, staying in a perfectly straight course is _hard_. Going in a straight line over a significant distance (say, 1 mile) requires consistent accuracy for each step over about 2000 steps, or the ability to correct for past inaccuracy, which you have no way of knowing without visual feedback. Small things like the mentioned difference in leg strength, or minor variations lead to a small offset in natural bias that eventually overwhelms the normal random error vector in each step.

Last alternative: Coriolis effect.

It's only when any reference points are taken away that humans can't walk straight. If you can see the sun, or a mountaintop, it's ok.

Bees and such rely heavily on the earth's magnetic field to keep their direction. I imagine if you were to mess that up they would have the same problem (I believe studies have been done with this in regards to bees and power lines.)

Radiolab=automatic 5☆