But our data on the minerals that cover Mars' surface, though still fragmentary, seems to point in the other direction. We seem to be seeing relatively few of the minerals that would be expected from Martian rocks coming into contact with large amounts of liquid water, and instead we are finding evidence of some that CANNOT survive such contact for more than a very short period without being changed into other minerals.

Later spacecraft -- starting with the three American and European Mars landers and the European "Mars Express" orbiter that are all scheduled to arrive over the next two or three months -- should provide much more data on this. (NASA announced recently that one such instrument -- a "Mossbauer spectrometer", on the first of its two Mars rovers, that had been partially malfunctioning -- has now been successfully fixed by radio command). But up now we've have to depend on "near-IR" spectra of reflected sunlight on the borderline between red and infrared light taken by ground-based telescopes, spectra obtained by MGS's "TES" spectrometer of the longer-wavelength "thermal" infrared actually given off by the surface's warmth, and the data from the thermal-IR THEMIS camera on Odyssey.

TES and THEMIS work well together. TES can obtain detailed spectra, but can only make surface maps with a fuzzy 3-km resolution. But a comparison of its spectra with THEMIS' more detailed photos of the same areas -- which are made in only nine different spectral bands, but cover the surface with a resolution of only 100 meters -- helps scientists interpret THEMIS' pictures to look for small outcrops of interesting minerals. Christiansen's talk discussed THEMIS' results up to now.

THEMIS has also provided a lot of data on the physical consistency of Mars' surface. Since large pieces of solid rock take much longer to cool off after sunset than loose dirt, its nighttime infrared photos of Mars' surface warmth have provided very detailed maps -- much more varied and informative than expected, in fact -- of the amount of solid material on the surface. They reveal the average size of rock fragments or soil grains in a region, and thus data on the amount of erosion and the way in which the winds have been moving and depositing Mars' thick blanket of surface dust.

(Some small spots on Mars are covered with bedrock stripped bare of soil by the winds -- and THEMIS has also discovered that most of Mars' steep slopes tend to be bare of soil and covered largely with exposed rocks. Other areas, however, have had meters-thick layers of fine dust gradually dropped onto them by the planet's wind patterns.)

And THEMIS' pictures also reveal layers of solid material made out of soil grains cemented together by various processes to form a hardpan, which is very widespread on Mars -- all of which reveals a lot about the geological history of various regions in itself. (THEMIS has also been looking for any small spots that might be geothermally warmed by Mars' remaining volcanic activity -- thought not to have completely died out yet -- which would be very promising places to look for both fossilized and still-living microbes. But it has yet to find any.)

Its chemical information is harder to analyze, but results are now starting to come in. What's been looked for most urgently is carbonate minerals. These are produced when carbon dioxide dissolves in liquid water and then reacts with the existing silicate surface rocks. So their presence in fair-sized quantities would be a dead giveaway that early Mars did indeed have a lot of surface liquid water, which might actually gradually have removed much of early Mars' dense CO2 atmosphere by turning it into carbonates -- helping reduce Mars to its current near-airless state.

TES and THEMIS, however, have come up frustratingly empty in that search. TES is now thought to have confirmed a few percent of carbonates in the fine windblown dust that blankets most of Mars, as had been suspected from earlier near-IR spectra of the dust by Earth-based telescopes. These carbonate traces are likely to have just been slowly manufactured over the eons by the dust's contact with the traces of water vapor remaining in Mars' air, and perhaps also by its repeated contact with tiny traces of liquid water that may appear briefly on Mars' surface during the warmest parts of its summer days, before they quickly boil away into vapor in the planet's low air pressure. (These processes by themselves could have removed a good deal of early Mars' CO2. More on them in the next part of my report.)

But TES has found no sign of the much more concentrated outcrops of carbonate rocks that would have been formed had early Mars had a lot of surface liquid water -- and now THEMIS, despite its much sharper spatial resolution, also cannot find even any signs of even tiny outcrops of such minerals.

It's been proposed that thermal-IR spectra can be insensitive to carbonate rocks with rough textures, but the fact that TES is detecting carbonates in the fine dust shows that this can't be the explanation for their absence in its data.

Alternatively, Benton Clark has suggested that early Mars had far more sulfurous gases in its early atmosphere from its stronger volcanic activity then, and these may have eaten away the initial carbonate rocks on Mars' surface, before the volcanic activity faded away and allowed the more recently created carbonates in the surface dust to survive. This would mean that there may still be large deposits of undestroyed carbonate rock buried underground.

But Christensen tells "SpaceDaily" that in that case TES, and especially THEMIS, should still be seeing some outcrops of carbonates on cliffs and slopes that have been freshly exposed by recent landslides of the sort that often occur on Mars. Martian rocks do have small amounts of carbonates -- they've been found in some of the 29 Martian meteorites discovered on Earth, apparently produced by relatively small amounts of liquid water -- but Christensen now thinks it fairly certain that Mars just does not have large deposits of them.

The data from TES and THEMIS also indicate that the rugged, high-altitude highlands that cover Mars' southern hemisphere are largely covered with boulders made of volcanic basalt that has never undergone much if any weathering by water.

The northern lowland plains are partially covered with a different type of rock that may be water-weathered basalt -- but they may instead be a more silica-rich type of rock called andesite, which would have been formed when pockets of molten volcanic rock deep in Mars' interior encountered water, but which has been completely dry ever since the volcanoes vomited it onto the surface.

Both TES' spectra and shorter-wavelength near-IR spectra of Mars taken by Earth-based telescopes have also consistently failed to show any large amounts of clay minerals, which should also be formed if liquid water is weathering rocks there.

And TES and THEMIS have found scattered but important surface outcrops of a greenish mineral called olivine -- a silicate of iron and magnesium which is often erupted as lava, but which breaks down completely into other minerals after only a few thousand years of exposure to even ice-cold liquid water (a mere moment in Martian terms).

TES first found evidence of olivine in small surface outcrops sprinkled across Mars, and THEMIS' maps have allowed us to locate them far more precisely. There is one 100-meter thick layer of it clearly exposed in the wall of Ganges Canyon -- a side branch of Mars' gigantic Marineris Valley -- which was apparently laid down as an ancient lava flow, and which has been unweathered ever since its exposure to the surface. A larger layer of olivine has been exposed in the Nili Fossae area, on the slopes of the great ancient shield volcano in the Syrtis Major area.

Some of these olivine outcrops must have been formed at least 3.5 billion years ago, and perhaps much earlier. They cannot have been exposed to underground liquid water during all those eons, either, which also makes it unlikely that they have been exposed to the surface by more recent landslides and just haven't happened to come into contact with surface liquid water during that shorter time. They provide strong additional evidence that the surface and near-surface of Mars, except perhaps for a few areas of long-lasting volcanism, has been frozen solid since at least the end of the Noachian Age -- and had very little nonfrozen liquid water on it even before then.