Deprived of any external cues, people go round: feedback errors accumulate as walkers try to go straight.

In one scene of the 1999 movie The Blair Witch Project, three film students searching for a legendary creature hike for hours only to end up at the spot where they had started.

Their misfortune is not just a suspenseful twist in a fictional world, says Jan Souman of the Max Planck Institute for Biological Cybernetics in Tubingen, Germany. Given no external cues to direction, people trying to walk straight over unfamiliar terrain end up doing intermittent loop-de-loops, Souman and colleagues report in a paper published online August 20 in Current Biology.

Circular walking occurs when people have to rely solely on bodily cues, such as rotational shifts and joint movements, to estimate the location of "straight ahead," Souman hypothesizes. As random errors in bodily feedback accumulate, a person drifts to one side or the other. A walker dependent on bodily cues may first make a circle to the right, drift back to a straight-ahead direction, start to zigzag and then make a circle to the left.

"You may think that you're walking in a straight line, but in fact the direction you're walking in is drifting more and more away from straight ahead, making you walk in circles," Souman says.

That's "a simple but elegant" proposal for how walking without any external directional signs can lead people in circles, says Roberta Klatzky of Carnegie Mellon University in Pittsburgh.

In these situations, people start to circle after traveling only about the length of a football field, Souman's team finds. Researchers are now trying to mathematically model how random errors in direction can yield systematic movements.

Klatzky led an earlier study in which blindfolded adults veered toward the right or left, without circling, while traversing almost half the length of a football field. She initially attributed this finding to a biological tendency to turn in one direction, perhaps because one leg is slightly longer or stronger than the other. In the new study, most individuals circled to both the left and right within the same trial, undermining Klatzky's proposal.

Souman's team first instructed three men to walk straight in part of the Sahara desert. The two men who walked during the day, with the sun visible, veered off course but did not go in circles. After clouds hid the moon, the man walking at night made several sharp turns in the same direction, nearly turning to face the direction from which he came.

In a second experiment, six college students walked for about four hours in a dense German forest where the landscape provides no clear cues to direction. The four volunteers walking on an overcast day traveled in a series of circles. Both students walking on a sunny day followed an almost straight course, except when clouds blocked the sun.

People, like bees and pigeons, may compensate for changes in the sun's position as they move, Souman speculates.

In a third experiment, 15 blindfolded participants tried to stride straight forward in a large, flat field. In a series of five-and 10-minute trials, participants walked in circles often no more than 20 meters wide. Only three veered consistently to the right or left. Souman plans to study people walking on a treadmill in a virtual forest so that he can control visual cues.

In emergency situations, cinematic monster hunters and others may become panicked, disregarding external heading cues and unintentionally ending up where they started, Souman suggests.

[ILLUSTRATION OMITTED]

Forest loop-de-loops In a new study, six students walked for four hours in a dense German forest. The paths of the four who walked on an overcast day (blue lines in this satellite view) made circles. The two who walked on a sunny day (tan lines) didn't make loops, possibly because the sun provided an external directional cue. Red dots indicate starting positions, and the tan dot indicates where one of these students (M J) left the study area and had to be redirected.