By Dr David Whitehouse
BBC News Online science editor
New data suggests that if there is ice at the Moon's poles then it is probably in the form of scattered grains rather than thick sheets, say scientists.
There are places on the Moon where the Sun never shines
Radar data from the permanently dark regions of the Moon lack the tell-tale signature of thick ice deposits.
The observations, from the Arecibo radio telescope in Puerto Rico, do not rule out ice, which could be a valuable resource for a future base on the Moon.
The results of the latest study are published in the journal Nature.
Bruce Campbell and his colleagues used the Arecibo radio telescope to look at the Moon's shadowed poles, choosing a previously unused radar wavelength of 70 centimetres. This can penetrate several metres of dust but would be reflected strongly by thick ice.
They did not detect any strong reflections from the lunar poles which may mean that any ice is likely to be present only as grains or thin layers embedded in rock.
Some researchers say this is only to be expected, as the data from Lunar Prospector spacecraft in 1998-99 suggested that the ice was sparse, occupying less than 2% of the volume of the lunar dirt.
Fuel and water
The "discovery" of ice in dark polar craters has caused a sensation and, if true, a re-evaluation of the Moon's place in space exploration.
Water-ice, it is argued, could provide drinking water for a colony as well as rocket fuel in the form of liquid hydrogen and oxygen.
Arecibo bounced radar beams off the Moon
But interpreting the evidence of water-ice on the Moon has always been open to differing views, depending upon whether the observations were made from the Earth or from the vicinity of the Moon.
In 1996, radar data from the Clementine lunar mission suggested ice was present at the south pole of the Moon.
However, in 1997, researchers published a paper in the journal Science reporting that radar imaging of the lunar poles by the Arecibo telescope showed no evidence of ice.
At the time, researchers speculated that the Arecibo radar would only have detected ice if it had been in the form of large chunks or slabs. The absence of an Arecibo radar return signal did not preclude ice being present in small chunks or crystals mixed in with the lunar soil, they added.
In 1998 another chapter in the story of ice on the Moon was written.
The neutron spectrometer aboard the Lunar Prospector orbiter, launched in January 1998, detected significant deposits of hydrogen as it flew over the Moon's north and south poles.
This was interpreted as indicating the presence of water-ice, since hydrogen in water molecules is thought to be the most likely source of the element at the lunar poles.
Clementine mapped the Moon for two months in 1996, and Lunar Prospector was deliberately crashed on to the Moon's south pole in 1999.
This means that until the European Space Agency's Smart 1 mission arrives at the Moon in 2004, the only investigations of lunar ice that can be made are Earth-based radar ones.