For years, some astronomers have been pushing the idea that there are large patches of ice at the Moon's poles. If confirmed, this ice could be useful for future lunar exploration efforts, acting as a resource for a permanent base, for example. But the paper on page 835 has dashed those hopes. It presents high-resolution radar images of the Moon's south pole — and finds no sign of ice. “Our results go against the hopes of many people,” says Donald Campbell of Cornell University in Ithaca, New York, who led the study.

The idea that there might be ice on the Moon is perfectly reasonable. Comets or meteoroids could have brought water with them as they smashed into the lunar landscape. The slight tilt in the Moon's axis of rotation means that there are some areas at the poles — mainly craters — that are in constant shadow from the Sun. Here the temperature would be less than −173 °C, more than low enough for any water to be frozen solid.

The concept was bolstered in 1992, when radar maps of Mercury indicated that there were large sheets of ice at its poles. And in 1994, the lunar orbiter Clementine sent back fairly low-resolution radar signals that suggested similar deposits of ice were present near the Moon's south pole.

The data gathered by Campbell and his team were obtained using the same principles as previous efforts. Radar waves are transmitted to the Moon's surface and the reflected wave is detected and analysed. The waves are scattered by rocks and soil, but are reflected back in a coherent pattern by ice. In addition, if the outgoing waves are polarized, ice will preserve the direction of polarization.

Campbell's team set out to image the Moon's south pole, and covered a larger area at higher resolution than ever before. The outgoing radar signals were transmitted from the Arecibo telescope in Puerto Rico, and the reflected waves were picked up by a telescope in West Virginia seconds later.

“We had to resolve a lot of issues relating to the amount of available memory in computers” to deal with the data collected, says Bruce Campbell (no relation) of the Smithsonian Institution in Washington DC, who worked on the project. The team collected data in April and October 2005, and constructed the first images six months later. “Suddenly an image pops up. It is a very satisfying feeling,” says Donald Campbell.

The team did find some features on the Moon that had radar-reflecting properties similar to the icy regions on Mercury. But unlike on Mercury, where these regions are all in shadow and so could be ice, the spots on the Moon were scattered randomly among sunny and shadowy regions. “These properties are not as unique to icy surfaces as we thought. That was a surprise,” says Donald Campbell. If there is ice on the Moon, he adds, it is probably in the form of grains mixed in with the soil, and not in large patches.