What this discovery seems to do is to further confirm a conclusion that more and more Mars scientists were already leaning towards -- namely, that before we dispatch any large number of new Mars landers, we badly need to do very extensive and detailed aerial reconnaissance of the planet's surface in order to find the best and most interesting possible landing sites.

Dr. John Kerridge of the University of California at San Diego has been particularly forceful on this subject. As he pointed out even before the failures of the 1998 Mars probes, unmanned Mars sample return missions will always be expensive and infrequent -- and each of them will return at most a few kilograms of samples.

And since it's fairly difficult to find ancient microbial fossils even on Earth, clearly such Mars missions will have to be very carefully targeted toward the best possible landing sites in order to have any reasonable chance of returning such fossils from Mars.

He thought that NASA's existing plan of launching sample-return vehicles as early as 2003 was wildly premature -- and, indeed, the failure of the 1998 Mars probes may to some extent have been a blessing in disguise, by forcing NASA to completely revise that strategy.

Moreover, MGS' super-detailed orbital photos of Mars' surface have shown that it is even more varied and complex than anyone had anticipated -- and that there are a wide variety of physical processes taking place on its surface that we don't yet properly understand.

The first step in such aerial reconnaissance will be the next Mars Surveyor orbiter scheduled for launch next April and arrival in October.

The Orbiter carries two main instruments, both of major use in this field. One is a reflight of the gamma-ray and neutron spectrometer lost in 1993 on Mars Observer.

Not only can it identify some key elements that will allow the overall rock types on Mars' surface to be identified, but it can also detect hydrogen in the upper soil -- which is directly proportionate to the amount of water ice there.

Its gamma-ray portion can detect ground ice at a depth of about 15 centimeters, while its neutron portion can detect it at a depth of 1 meter -- allowing some gauging of just how close to the surface such ice is, as well as a better overall understanding of how much groundwater Mars does have. But its spatial resolution is very poor: only about 300 km.

The other instrument is "THEMIS", a thermal-emission infrared camera that will take pictures of most of Mars' surface in 10 different spectral bands simultaneously, with a resolution of only 100 meters.

The thermal-emission spectrometer already studying Mars on MGS has the ability to sense many minerals that would be produced by ancient Martian liquid water in contact with rocks (such as carbonates and sulfates) -- but its resolution is a fuzzy 3 km, so that it simply is not sharp-eyed enough to detect small local outcrops of such minerals such as are likely to exist on Mars.

And, indeed, MGS so far has not detected any sign of surface carbonates or sulfates, although they were expected on theoretical grounds to exist on or near the surface.

But THEMIS, with its sharper eyesight, should have a far better ability to detect such minerals in small patches. A full-fledged mapping thermal-IR spectrometer (like the high-resolution shorter-wavelength near-IR mapping spectrometers on the Galileo and Cassini spacecraft) is still too heavy and bulky an instrument to fly easily on a small spacecraft -- but THEMIS, with its ability to take pictures in 10 different spectral bands, should still be able to clearly identify most of the scientifically interesting minerals. And it will be used to observe the identified runoff sites to see whether there are indeed minerals there produced by liquid water.

Moreover, THEMIS will map Mars' nighttime surface with equal sharpness to look for local geo thermally warmed spots, with a sensitivity of only 1 deg C. If any of the possible runoff locations are still warmer than normal -- such as would be the case if liquid water was still very close to the surface -- it can confirm it.