E. coli swim in a fluid by using motors embedded in the cell wall to rotate bundles of slender hairs known as flagella (figure 1). If the motors all rotate counter-clockwise in a viscous liquid like water, the flagella bundle together and propel the cell forward in an approximately straight line. However, if one or more of the motors rotate in a clockwise direction, the flagella unbundle and the cell undergoes a tumbling motion.

When the cell is moving forward the thrust generated by the rotating bundle is opposed by the drag acting on the entire cell, while the counter-clockwise motion of the bundle is balanced by a clockwise rotation of the cell body. The swimming motion can therefore be described as force-free and torque-free. In the 1970s, however, Howard Berg, also of Harvard, discovered that E. coli does not swim in a straight line when it is close to a surface. Instead it moves in a clockwise direction, as viewed from above, tracing out a circle with a radius of the order of 25 microns.

The Harvard researchers model the bacteria by replacing the bundles with a single rigid helix and assuming that the body of the cell is spherical (figure 2). They go on to show that the circular motion is due to the hydrodynamic interactions of the swimming bacteria with the nearby surface. Moreover, their calculations demonstrate that the radius of the circle made by the bacteria increases with the length of its body. These results agree with experimental observations of E. coli (see movie).

"An immediate application of this work is to sort cells according to their size," says team member Eric Lauga, who carried out the modelling part of the paper with Howard Stone. The next challenge, he says, is to understand what controls the distance the cells and the solid surface.

Meanwhile, two of the experimentalists -- Willow DiLuzio and George Whitesides -- and co-workers have just published a paper in Nature in which they show that E. coli "drive on the right" when confined between two interfaces in a micron-sized channel.