Harry McSween of the University of Tennessee and colleagues have studied samples from the Shergotty meteorite -- volcanic rock from Mars that fell to Earth nearly 150 years ago. They conclude that, when Mars was young, a lot of water may have been spewed out as steam from the molten rock (magma) beneath the planet's surface.

They estimate that martian magma contained far more water than has previously been thought. The results, published in this week's Nature1, suggest that as it started its ascent from the red planet's interior to the surface, the magma that was to become the Shergotty meteorite was probably as much as 1.8% water.

But the loss of water during the magma ascent has left the meteorite's mineral grains with a deceptive 'water-free' coat that masks the true picture, the researchers say. They found that the mineral grains in a fragment of the meteorite contain a higher proportion of water-soluble elements in their interiors compared with their outer crusts.

This indicates that the rising magma could have released a significant amount of water at the planet's surface, the team says.

Human exploration of the red planet may be as close as 2020. The bait is the prospect that Mars once harboured life -- a hundred-year-old fantasy revitalized by the discovery in the 1970s of dried-up river-beds on the planet.

The photos of these sinuous valleys, taken in 1972 by the Mariner 9 spacecraft, were eclipsed by the stunningly detailed images returned by the orbiting Mars Global Surveyor satellite in the late 1990s.

Most of these valleys are believed to have formed and dried up billions of years ago. But last June the Global Surveyor turned up gullies apparently carved by water in terrain no more than a million years old, implying that water flowed on Mars even in the geologically 'recent' past. Perhaps there could still be a vast amount of water locked away beneath the dusty surface.

The channels in the planet's surface imply that Mars might have once supported entire oceans. But where this water came from not such a simple tale.

Some of it could have been delivered, as on Earth, by icy comets striking the planet in the early days of the Solar System. But it is thought that much of it may have emerged as steam from the planet's once-fiery interior.

This process of 'outgassing' is still happening on Earth. Volatile substances such as water were dissolved in the magma of the hot young planet. As the magma rises towards the planet's surface in volcanic regions, pressure is released and the 'volatiles' come bubbling out like the carbon dioxide from an uncorked champagne bottle. In the early days, this outgassing would have been much more intense, supplying gases to the primeval atmosphere.

Mars no longer has active volcanoes -- it is too cold. But the immense remnants of former volcanoes are still evident, and these too would have served as conduits for outgassing. The trouble is that until now samples of volcanically formed rocks from the planet have shown little sign of having solidified from magma once rich in water. But if McSween's group is right, this evidence of a low water content in martian magma may be misleading.