Solids and liquids could hardly seem more different, one maintaining a rigid shape and the other flowing to fit the contours of whatever contains it. And of all the things that slosh and pour, superfluids seem to capture the quintessence of the liquid state--running through tiny channels with no resistance and even dribbling uphill to escape from a bowl.

A superfluid solid sounds like an oxymoron, but it is precisely what researchers at Pennsylvania State University have recently witnessed. Physicists Moses Chan and Eun-Seong Kim saw the behavior in helium 4 that was compressed into solidity and chilled to near absolute zero. Although the supersolid behavior had been suggested as a theoretical possibility as long ago as 1969, its demonstration poses deep mysteries.

Rotation is one way that superfluids reveal their peculiar properties. Take a bucket of ordinary liquid helium and rotate it slowly, then cool it down to about two kelvins, so that some of the helium becomes superfluid. The superfluid fraction will not rotate. Because part of the helium is motionless, the amount of force required to set the bucket and helium rotating is less than it would be otherwise. Technically, the helium's rotational inertia decreases.

Chan and Kim observed such a decrease of rotational inertia in a ring of solid helium. They applied about 26 atmospheres of pressure to liquid helium, forcing the atoms to lock in place and thereby form a fixed lattice. They observed the oscillations of the helium as it twisted back and forth on the end of a metal rod. The period of these torsional oscillations depended on the rotational inertia of the helium; the oscillations occurred more rapidly when the inertia went down, just as if the mass of the helium decreased. Amazingly, they found that about 1 percent of the helium ring remained motionless while the other 99 percent continued rotating as normal. One solid could somehow move effortlessly through another.

So how can a solid behave like a superfluid? All bulk liquid superfluids are caused by Bose-Einstein condensation, which is the quantum process whereby a large number of particles all enter the same quantum state. Chan and Kim's result therefore suggests that 1 percent of the atoms in the solid helium somehow form a Bose-Einstein condensate even while they remain at fixed lattice positions. That seems like a contradiction in terms, but the exchange of atoms between lattice sites might allow it. A characteristic of helium would tend to promote such an exchange--namely, its large zero-point motion, which is the inherent jiggling of atoms that represents a minimum amount of movement required by quantum uncertainty. (It is the reason helium ordinarily only occurs as a gas or a liquid: the extremely lightweight atoms jiggle about too much to form a solid.) Supporting the idea of condensation, the two researchers did not see superfluidity in solid helium 3, an isotope of helium that as a liquid undergoes a kind of condensation and becomes superfluid only at temperatures far below that needed by liquid helium 4.

Another possibility is that the crystal of helium contains numerous defects and lattice vacancies (yet another effect of the zero-point motion). These defects and vacancies could be what, in effect, undergo Bose-Einstein condensation.

But all those theories seem to imply that the superfluidity would vary with the pressure, yet Chan and Kim see roughly the same effect all the way from 26 to 66 atmospheres. Douglas D. Osheroff of Stanford University, the co-discoverer of superfluidity in helium 3, calls the lack of pressure dependence "more than a bit bewildering." He says that Chan and Kim have done "all the obvious experiments to search for some artifact." If they are correct, Osheroff adds, then "I don't understand how supersolids become super. I hope the theorists are thinking about it seriously."