This was a major shock, since the general belief was that modern-day Mars is far too cold for near-surface liquid water to exist.

And the gullies are even more puzzling because they aren't located in Mars' warmer equatorial latitudes, but only between 30 and 60 degrees latitude, where the planet is even colder -- many of them in areas that spend half of each Martian year at around the freezing temperature of CO2, and none of them in any region where the near surface ever gets warmer than about minus 18 deg C even in mid-summer.

Nor -- save for a few located in the Dao Valley -- are any of them located near any region that looks as though it may have been volcanically warmed in the geologically recent past.

On top of that, they're about 2.5 times more likely to flow down the colder slopes facing toward the poles than they are to flow down slopes facing toward the Equator! What's going on here?

Well, these are the latitude zones on Mars where near-surface ground ice is mostly likely to exist -- the somewhat warmer climate closer to the equator is thought likely to have caused any near-surface ground ice to "sublimate" (boil directly from ice into vapor in Mars' thin atmosphere) long ago.

There are strong indications that Mars' surface is considerably smoother on a small scale at higher latitudes, and big craters there have oddly blurred features -- all of which may be due to the existence of near-surface permafrost which is slowly "creeping" sideways over periods of millions of years, even at Mars' low temperatures.

Malin and Edgett suggest that, if liquid water could somehow get within 100 meters of the surface in these regions, it would tend to boil directly into vapor from the exposed surfaces of strata on the warmer equator-facing cliffs -- but on the colder pole-facing cliffs, it might instead form a plug of ice there which would sometimes suddenly burst open under the pressure of liquid water building up behind it, allowing the liquid water to gush out and down the slope for hundreds of meters before it froze and evaporated in the thin Martian air.

But that still leaves us with the problem of explaining how liquid water could get so close to the surface there in the first place.

One possible explanation has been that Mars' ground ice might be mixed with salts of the type thought likely to exist in large amounts in Mars' upper surface, which could lower its freezing point to an amazing degree.

Magnesium chloride could lower it to minus 33 deg C.; calcium chloride to minus 53 deg; and a mixture of them could conceivably lower it all the way down to about minus 63 deg -- low enough for such briny ice to thaw out during the summer even in the gullies' high latitudes.

At the latest LPSC, Paul Hess and John Longhi suggested that the temperature at which even plain water freezes would be greatly lowered if it was mixed with fine-particle soil.

As for the source of heat capable of melting such ice, William Hartmann has suggested that the evidence that volcanic eruptions occasionally occur even on modern-day Mars also indicates that the planet's overall level of subsurface geothermal warmth -- even in areas far away from surface signs of volcanic activity -- could melt such subsurface ices.

M.T. Mellon elaborated on this by saying that a surface layer of Mars' dry, powdery soil would be an efficient insulator to keep a near-surface layer of geothermally heated water from losing its heat to the cold surface.

However, Victoria C. Gulick -- a fan of the idea that hot springs have played a major role in Mars' past -- nevertheless concluded from her own analysis that such geothermal warmth is "insufficient to produce substantial fluid discharge from the surface" unless a small "dike" of subsurface magma is fairly nearby.

Another possibility has to do with the fact -- pretty universally accepted -- that Mars, without a large moon to stabilize it, has its "obliquity" (the tilt of its spin axis) slowly rock back and forth between 0 and 45 degrees over a cycle of about a hundred thousand years -- with less frequent periods during which it may tile all the way over to 60 degrees, leaving the planet "lying on its side" the way Uranus and Pluto do.

Very strange things happen to the weather of such a planet; because one pole or the other is continually exposed to sunlight for months at a time while equatorial regions still go through a day-night cycle, both poles actually get warmer year-round than the equator.

Such a climate cycle could cause water ice near the poles to thaw out and vaporize in Mars' thin air, migrate to somewhat lower latitudes and then refreeze -- after which, as Mars' tilt reduced again and its lower-latitude regions warmed up, the water ice stored there would again thaw out, erupt from the gully sites as liquid and then evaporate into the thin air, and then migrate back to the poles and refreeze.

John F. Mustard pointed out at the Conference that MGS has shown that this same 30 to 60-degree latitude zone on Mars is widely covered by "stippled" terrain covered by pits just a few meters across.

He proposes that the 100,000-year climate cycle alternately causes Mars' windblown dust in such regions to be fused together into a smooth permafrost surface when it lands, and then causes the surface to become rough again during the next stage of the cycle when the ice in the upper 1 to 10 meters of soil warms up and sublimates back into the atmosphere to migrate back to the polar regions -- which, despite their periodic warming, never get warm enough below the soil surface for their ground ice to sublimate, so that they remain smoother-surfaced.

In some areas, the buildup of a thick surface layer of "duracrust" -- soil fused together into a hardpan when water sublimates out of it and leaves behind crusted sulfate salts, which is definitely known to cover a great deal of Mars' surface -- could seal away melting subsurface ground ice from sublimating back into the air, so that a near-surface layer of liquid water would build up and could occasionally break through and gush out onto slopes to form the isolated gully sites.

Pascal Lee suggested a variation on the climate-cycle theory.

The Martian gullies all tend to start from pocket-like "alcoves" on the slopes, suggesting that these may be places where subsurface liquid erupted.

But based on his observations of very similar-looking gullies on slopes in Greenland, Lee thinks that the Martian gullies may have been formed not by thawing subsurface water, but by local snowbanks that naturally accumulate in such alcoves (carved by the wind or landslides) during the cold part of the 100,000-year climate cycle and then simply thaw out and run down the surface when the slope warms up again.

All these theories, however, assume that the substance erupting really is liquid water -- and Nick Hoffmann, once again, suggests that this interpretation is radically wrong, and that the erupting subsurface substance is actually liquid CO2, stored in near-surface pockets.

(He was joined in this by D.S. Musselwhite.) The pressure 100 meters below the surface is quite high enough to maintain CO2 in liquid form, provided it's sealed off from the thin surface air by a layer of ice -- and Mars' meteor-pulverized upper surface ("regolith")is porous enough for atmospheric CO2 gas to seep down that far below the surface when such an ice layer doesn't exist.

Musselwhite's model says that when the pole-facing slope surfaces where the gullies are formed chill down to particularly low temperatures in the cold part of the climate cycle, some of that underground CO2 gas freezes directly into solid "dry ice" around the regolith fragments, until the spaces between them are completely packed with solid dry ice.

Then, when the cliffs warm up again during the next part of the obliquity cycle, the top surface of the dry-ice layer, exposed directly to Mars' low-pressure air, starts to boil directly back into gas -- but its lower and inner parts, still sealed off from that thin surface air pressure, instead melt into an underground deposit of liquid CO2, which finally bursts out of the cliff side when the solid dry-ice barrier finally erodes away completely, flashing explosively into gas and producing a miniature version of the titanic "cryoclastic" gas outbursts that Hoffmann thinks carved Mars' immense catastrophic-outflow channels during the Hesperian era.

The high-pressure gas jet produces a cloud of dense gas, soil, loose rocks and still-sublimating chunks of CO2 ice, which rolls down the slope for hundreds of meters before the gas disperses into the surrounding thin air, producing a surprisingly thin gully that looks just as if it was carved by fluid.

One piece of evidence for this is that the gullies all seem to have erupted from places around 100 meters from the top of cliff sides -- just the right depth and pressure for underground liquid CO2 to exist on Mars.

By contrast, if the gullies were carved by eruptions of briny water, it's hard to explain why some of the eruptions wouldn't have occurred higher up on the cliff faces, or perhaps even lower down on them.

At the Conference, Hoffmann proposed another piece of evidence -- the very recent discovery of a set of the gullies at fully 71 degrees South latitude, actually within the area where some of Mars' CO2 atmosphere freezes into a temporary dry-ice polar cap every winter and then thaws out in spring.

This area is so cold that it's even harder to conceive of even briny water existing near its surface -- and, to quote Hoffmann, "the freshness of this gully system despite annual burial [by the CO2 ice polar cap] suggests that this gully system is active at the present day and literally renews itself with fresh flows each Martian spring."

In short, while some gully eruptions may only be unleashed every hundred thousand years or so by Mars' slow obliquity climate cycle, others may actually be produced every spring -- which means that MGS or future Mars orbiters will see it happen at some point.

In another LPSC paper, Sarah T. Steward and Francis Nimmo objected to Hoffman's overall geological model, on the grounds that "there are considerable thermodynamic difficulties with storing such large quantities of CO2 in the near surface on Mars."

That is, the huge amounts of subsurface liquid CO2 necessary to fuel the gigantic gas eruptions that produced the catastrophic outflow channels in the Hesperian could never have gotten down there in the first place -- nor could enough liquid CO2 be concentrated in any local area to explain the small recent slope gullies.

But Musselwhite's seasonal cycle theory could explain the latter, since it calls for the near-surface regolith in the gully areas to be solidly packed anew with solid frozen CO2 every 100,000 years (or even every Martian winter), which then melts back into liquid CO2 when the area warms up again.

As for the catastrophic outflows: even if Hoffmann is wrong in believing that Noachian Mars was frozen solid from the very beginning -- and that it instead actually had a CO2 atmosphere dense enough, and producing a strong enough greenhouse effect, to keep itself from freezing -- it may very well be that part of that early dense atmosphere was gradually stripped away by giant impacts that splashed it into space, by erosion by the passing solar wind, and/or by conversion into subsurface carbonate deposits by geothermally heated subsurface liquid water (as is happening right now in Iceland).

Then when the air had thinned enough that its greenhouse effect could no longer keep Mars' poles above the freezing point of CO2, the remaining atmosphere would swiftly have undergone a self-amplifying, runaway freeze out to form two great polar caps of dry ice -- for as part of Mars' CO2 atmosphere froze there, the remaining air would have become even thinner and even less capable of warming itself.

Then, after Mars' CO2 was entirely frozen into those two great polar caps, some of the dry ice at their bases would gradually have been melted into liquid by Mars' geothermal warmth and gradually trickled away to spread through the rest of the planet's subsurface, providing a ready and waiting supply of deeply buried liquid CO2 to fuel the later catastrophic gas outbursts.

(This might also explain the evidence James Head has found that Hesperian Mars had a much bigger south polar cap which has since disappeared, leaving traces of its existence in the form of what looks like subglacial melting ponds and "esker" trails of sediment deposited by flowing subglacial streams.

The only difference would be that the ancient polar cap and its buried basal liquid would be CO2 rather than water, as Head thinks.)

There were other objections to Hoffmann's theory at the LPSC.

Natalie Cabrol presented evidence that several Martian craters with gullies running down their slopes also show clear evidence of flat, wide sediment layers deposited on their floors, which she thinks could only have been spread that far by sediment drifting in liquid water rather than being sprayed out as cryoclastic clouds by gas jets.

That is, she thinks that liquid water "ponded" in these craters for prolonged periods at some point in Martian history -- perhaps even fairly recently -- before disappearing.

Hoffmann, however, disagrees and thinks that cryoclastic debris clouds really could have spread out that widely.

There may not be any way to decide this debate without actually landing at those sites and looking.

Finally, Stewart and Nimmo object that any jet of CO2 gas produced by released liquid CO2 suddenly flashing into vapor at the gully sites would produce such a powerful blast of gas that it would carve a relatively wide, straight and short gully -- as gas outbursts from the side of Earth volcanoes do, and for that matter as the great Martian catastrophic outbursts seem to have done.

They argue that the long, thin, tapering and sometimes meandering slope gullies could only have been carved by liquid.

Again, Hoffmann disagrees, saying that a gentler, longer-lasting jet of released CO2 would gradually have eroded a thin, gradually lengthening gully -- but, again, this question involves such complex questions regarding soil mechanics that the only way to decide it may be to check out the gullies with Martian landers.

Finally, there's another disturbing mystery surrounding the slope gullies -- the only ones MGS has found so far look as through they were formed very recently.

Their features are sharp; and the "aprons" of debris at their feet have virtually no craters on them -- even tiny ones -- and have often flowed over such other surfaces as wind-blown dunes and "polygons" of cracked surface soil which must themselves be very recent.

But while this can be explained if recent (maybe even early) climate changes are what triggers the gullies, so far MGS' photos, strangely, have not revealed any older-looking gullies.

At the Conference, G. Leone tried to explain this on the grounds that the actual rock erosion rate on Mars since its atmosphere turned thin has been extremely slow, and said that even the youngest-looking gullies may be hundreds of millions of years old, this doesn't at all explain their aprons of totally uncratered landslide debris covering up such extremely fresh surfaces as dunes.

It would be an incredible coincidence if they had been created by some one-time freak occurrence on Mars that just happened to occur only a few hundred thousand years ago, and this of course would be flat-out impossible if they are indeed caused by a hundred-thousand year climate cycle.

Where are the older gullies? It may be that -- as Kenneth Tanaka said last year -- "older shallow discharges may also have occurred... and may be obscured or just not recognized yet." At this point we simply don't know.

One related note: The strange, banded "layered terrain" that stretches for hundreds of kilometers around the edges of Mars' polar caps -- on which Mars Polar Lander tried unsuccessfully to land in 1999 -- is generally thought to consist of layers of varying mixtures of ice and windblown dust laid down by those same 100,000-year Martian obliquity climate cycles.

However, S.M. Milkovich noted at the Conference that it now seems that "individual layers probably formed on timescales too short to be due entirely to obliquity cycles" -- and also that the fine-scale MGS photos also indicate that there are very great differences in fine texture between the layers, suggesting that their deposition was affected not just by repetitive climate cycles but by other factors such as volcanic eruptions in Mars' recent past that may have dumped large amounts of ash and water vapor into the air.

This may also indicate that Mars is still undergoing relatively short-term climate changes -- perhaps lasting only a few thousand years each -- whose cause is still unknown.