Instead, it would hurtle nonstop over Mars' surface at 40 km altitude and scoop up a sample of both the atmosphere and the very fine dust particles floating in it, and return them to Earth.

This dust wouldn't include any meaningful biological evidence - but it would provide us with much more knowledge of the mineralogy of Mars' surface in general, as opposed to the current Mars meteorites that seem to have come from only a handful of locations on the surface.

We could, for instance, understand more clearly what kind of weathering processes have affected Mars' surface minerals - and whether the minerals found in the dust indicate that Mars had large amounts of liquid water on its surface in the past.

And the accompanying air sample would be invaluable in itself - for instance, by analyzing the percentages of trace isotopes in it, we could estimate how much of Mars' original dense atmosphere was actually swept away into outer space by the flow of solar wind past the planet, as opposed to chemically reacting with Mars' surface rocks or being blasted into space by giant meteorite impacts.

SCIM would take off in Sept. 2007 and make its first Mars flyby in late 2008 - but that would be a high-altitude flyby, allowing Mars' gravity to redirect it into a different solar orbit.

Then, in July 2009, SCIM would return to Mars and hurtle over its southern hemisphere at 40 km altitude, where using two small pads of shock-absorbent "aerogel", similar to those employed by the Stardust mission, will collect micron-sized dust samples.

The extremely fluffy aerogel can capture these "dust" samples without causing serious scientific damage to them - even at speeds of 20,000 km/hour. Simultaneously, SCIM would scoop up a separate sample of Martian air.

To survive such a low pass through Mars' atmosphere (unlike the accidental low pass that doommed Mars Climate Orbiter in 1999), the SCIM spacecraft would be mostly wrapped in a heat shield - but, unlike the shield on any earlier spacecraft, it would be long and pointed to minimize the extent to which air friction slowed down the craft during its fiery dive.

Immediately after skimming over Mars' edge and then out again into space, it would fire a rocket motor to compensate for the 2200 km/hr of speed it had lost during the pass, so that it would then be on course again for a return to Earth in May 2010, at which point it would drop off its collected samples in a small Earth-return capsule - again like the one being employed for Stardust. During the entire trip, it would be powered by a single long solar panel trailing behind it, which would be safely located in its nonheated wake during its dive through Mars' air.

The dust - unlike material returned directly from Mars' surface - would have been sterilized by solar UV light during their long period floating in the Martian atmosphere, and so wouldn't need to be quarantined. And - given the extreme sensitivity of modern, big Earth-based analytical instruments - individual dust grains could be inspected and analyzed in detail far beyond anything that in-situ Mars instruments could possibly do, which of course is one the central arguments for sample return missions.

During a recent competition held by the European Space Agency's for the possible use in 2005 of a copy of the Mars Express bus, France proposed its own Mars atmospheric sample-return mission. But it was rejected as too expensive for the time being, and in any case it would have dipped no lower than 120 km altitude into Mars' atmosphere - too low to collect any dust, and with even its air sample being atypical of the planet's lower-atmospheric gases.

Another mission - "Mars Scout Radar", proposed by Bruce Campbell of the Smithsonian Institution - is just what its name indicates: an orbiter (again based on the MCO bus), equipped with an umbrella-like radar antenna fully 5 meters across.

Mars Express and the 2005 U.S. orbiter are scheduled to use very low-frequency radar to probe several kilometers below Mars' surface, looking for subsurface ice and pockets of liquid water. But MSR's radar - its only instrument - would use a much shorter wavelength, for a completely different purpose.

Mars Global Surveyor's high-resolution photos and IR spectra of Mars' surface minerals have confirmed something long suspected: most of Mars' surface is blanketed with a layer of wind-blown dust and sand - sometimes meters thick - which is concealing most of its ancient surface features, and making analyses of local minerals almost impossible, as the dust has been mixed to an even composition all over the planet.

MSR would use three shorter wavelengths - 3, 25 and 75 cm - to construct high-resolution "synthetic-aperture radar" pictures of Mars' surface similar to those that Magellan used to map Venus, which would look like real photographs.

But the two longer wavelengths would punch through as much as 5 meters of dust or sand to reveal the features of the underlying bedrock, making a planetary map with a resolution of half a km.

This would allow revelations of a vast variety of features of the surface of ancient Mars that are still unknown today - just as similar radar on the Space Shuttle has punched through the desert sands of the Sahara to reveal networks of ancient, buried river valleys.

MGS has found that most of Mars' ancient "valley networks" are partly filled with windblown sediments, making it impossible to tell from the cross-section shape of their beds whether they were carved by fluid (presumably water) running along early Mars' surface, or tunneled out by slower underground streams of fluid until the tunnel's roof caved in - in which case ancient Mars' surface might have been below freezing.

MSR could settle that - and it could also pierce sediments along the edge of Mars' great northern lowland plains to find if they were, as some think, the shores of a vast northern ocean.

By looking for such ancient river valleys, lakebeds and ocean shorelines MSR could provide critical new data for choosing good landing sites of later missions. And it could also provide a much better count of craters under sand-shrouded terrain to find out how old it was when it when it was first sediment-covered.

Moreover, the longer-wavelength radar could pierce as much as 70 meters through the ice of Mars' polar caps, providing insight into the changes in their deposition and of local weather over the millennia. Meanwhile, the shortest, 3-cm radar could provide an altitude map of Mars' features with a horizontal resolution of only 100 meters - far better than MGS' laser altimeter.

MSR would be able to map Mars completely in only about 6 months. The scientific usefulness of this mission is so clear that NASA's current Mars plan - if the extra money from Bush comes through - tentatively calls for such a radar-mapping orbiter to be launched anyway in 2009, probably in collaboration with Italy. But Campbell's proposal would provide a 2-year jump on this.