This is the "White Mars" model, which totally rejects the view of early Mars as a relatively warm and friendly place for microbial life.

In Hoffman's view, early-day Mars -- most of the time -- was even more savagely hostile than modern Mars, simply because it was even colder and more airless.

He points out that virtually everyone accepts the fact that the Sun, during its young days, was a good deal less intense than it is now -- indeed, 4 billion years ago it was only about 80% as bright, and has been gradually brightening since.

Even if you assume that early Mars had a CO2 atmosphere several times denser than Earth's, it's very hard to visualize a greenhouse effect strong enough to raise Mars' surface temperature in those days anywhere near the melting point of water -- and so the "warm and wet" Mars proponents have had to call in such additional factors as large traces of volcanic sulfur dioxide (a stronger greenhouse gas) in the air, or the possibility that the layer of frozen CO2 clouds that would have formed in Mars' stratosphere would produce a strong additional greenhouse effect (although there is currently a debate raging as to whether those clouds would warm Mars, cool it, or not have much effect one way or the other -- we just don't know enough yet about cloud physics).

Hoffman thinks that, in fact, any such CO2 atmosphere wouldn't even have produced enough greenhouse effect to keep itself from freezing out, starting at the poles -- and so in his view Mars 4 billion years ago was an icebox world with an average temperature of only 78 deg below zero C., and a faint CO2 atmosphere with a surface pressure of only about 1 millibar -- less than 1/10 of its current level.

It never had liquid water on or near its surface, except for very brief times and places.

What it did have from the start -- and still has -- is a great deal of CO2 stored below its surface in various forms: as solid "dry ice", as a "clathrate" mixture of CO2 with water ice, and even as liquid CO2 -- which cannot exist at all at an air pressure less than 5 times that of Earth's (that is, 5 "bars"), but which could reach such pressures as little as 7 meters below Mars' surface, provided it was sealed away from the surface near-vacuum by a layer of solid ice.

Then -- as the Sun slowly warmed during Mars' middle "Hesperian" period (roughly 3.8 to 2 billion years ago) -- a dramatic new chapter opened in Mars' history.

The gigantic "catastrophic outflow channels" on Mars have always been one of its biggest mysteries -- they look as though they were titanic one-time eruptions of liquid water from underneath Mars' surface in some spots, producing floods hundreds of times greater than any that are ever known to have ever occurred on Earth, which carved great valleys hundreds of kilometers long and dozens of km wide before quickly dying out, after which the water they had poured out onto the surface either froze into ice or "sublimated" (evaporated directly from ice into vapor, since Mars' air pressure by the Hesperian had already dropped to the same wispy near-vacuum level it's at today).

But scientists have always had a great deal of trouble explaining just where those gigantic eruptions of high-pressure water came from -- especially since the patches of collapsed, "chaotic" ground from which they seem to have gushed usually don't look anywhere near big enough to contain amounts of water capable of carving such huge flood channels.

(After all, there's only so much liquid water you can cram into the pores of subsurface rock or soil.)

Hoffman proposes a radical new solution to this "Volumetric Paradox": the stuff that gushed out in such gigantic, forceful flows wasn't liquid water at all, but huge high-pressure jets of CO2 gas, released from Mars' great underground deposits of liquid CO2 (which he calls "liquefiers").

When liquid CO2 is heated above its boiling point of minus 57 deg C, by a spasm of local volcanic activity or a giant meteor impact -- or when an impact or landslide cracks the layer of ice and rock that have sealed it off from Mars' surface near-vacuum -- it explodes violently into a volume of gas 500 times the volume of the original liquid.

This could produce a tremendous eruption of high-speed gas, capable of sweeping millions of tons of surface rock and soil along with it in a rolling cloud -- just as eruptions of hot gas from volcanoes sometimes do, sweeping gigantic "pyroclastic" clouds of red-hot ash and rock along with them.

(One such cloud, erupting from the side of Mount Pelee in 1902, wiped out the entire town of St. Pierre.)

Hoffman's clouds would be an even bigger, ice-cold version of such clouds; he calls them "cryoclastic".

But they would be capable of plowing huge valleys in Mars' surface just like huge floods of liquid -- and, even given Mars' low air pressure, such a huge cloud of high-pressure gas and solid rocky debris could roll across Mars' surface for hundreds of km before finally dispersing into the surrounding air and dumping all of its solid rocky burden (just as pyroclastic clouds from volcanoes can travel astonishing distances on Earth before finally dispersing).

They would have become commoner during the Hesperian, as the slowly brightening Sun warmed Mars' near-surface layers more -- and then they would gradually have died away, as Mars' underground reservoirs of liquid CO2, left over from the cold Noachian age, were finally depleted.

Moreover, an initially small liquid CO2 release could have collapsed and shattered enough surrounding rock to release other nearby pockets of liquid CO2, thus producing a gigantic, self-amplifying runaway reaction.

It's a dramatic picture, and it explains three major puzzles of Martian history simultaneously:

1. The fact that early Mars looks as though it had large amounts of liquid water pouring across its surface, despite the fact that it's hard to see how the fainter early Sun could have warmed its surface above freezing;

2. The fact that Mars, to have any chance of being warm enough for surface liquid water, must have had a strong greenhouse effect from a very thick early CO2 atmosphere -- but such an atmosphere probably could have been completely destroyed only by being converted into great beds of carbonate minerals, which don't seem to exist on Mars today. (Nor are there many other mineralogical signs that early Mars ever had large amounts of surface water.)

3. The fact that its most remarkable geological features, the "catastrophic outflows", seem to have erupted from areas much too small to contain them if the erupting material was liquid water.

In one paper, he disagreed with Jeffrey Kargel's proposal that the cryoclastic gas blasts could have been released not from reservoirs of liquid CO2, but from subsurface deposits of "CO2 clathrate" (CO2 incorporated into solid ground ice).

His objection is that such clathrate would have broken down and released its stored CO2 only if it had been melted -- which is rather hard, since it has a melting point fully 10 degrees C. higher than regular water ice -- or if it had been exposed to the surface near-vacuum by rupture of the overlying ground, in which case it tends to break down and release its gas in a slow, gentle way that wouldn't produce a high-powered eruption.

In another paper, he suggested that the great smooth plains of sediment that fill Mars' northern lowland basin may have been deposited not on the floor of an ancient ocean, or by great amounts of material washed into the basin by ancient floods of water across the neighboring highlands, but by an even more titanic Hesperian version of the same "runaway cryoclastic eruptions" that he thinks plowed the outflow channels, again triggered by the slowly warming Sun.

"Collapse of the crustal dichotomy on Mars began as one or more small chaos zones, eating back into an extensive regional CO2 liquefier.

Collapse was prolonged and explosive... As the dichotomy [the boundary cliff between the new collapsed lowland and the surrounding highlands] collapsed, the slurries would have poured into the northern lowlands, filling and covering pre-existing cratered terrain and emplacing thick deposits on a very rapid timescale... The northern plains would have been an extensive, mobile 'sea' of fluidized debris -- a 'Mud Ocean' [with the "mud" actually being dry tides of shattered rock and soil borne along on a constant, boiling eruption of new CO2 gas releases].

Its surface would have been wreathed in clouds of escaping CO2 from fumaroles, fissures, and vents.

The margin of the Mud Ocean would have been like the flow front of an aa lava -- advancing in a bulldozer-like carpet, burying and engulfing the surrounding terrain, and steaming from the escape of CO2.

The speed of advance would have been, like lava flows, very variable... Long stand stills and episodes of slow creep would alternate with short intervals of hundreds of meters of advance per day, or more." By the time this fantastic spectacle was finally complete -- in just a few tens of thousands of years -- Mars' initial northern basin would have been eaten away at the edges enough to cover one-third of the planet, and all that shattered edge material would have been spread smoothly across its floor for hundreds of km.

It's an astonishing picture, and a growing number of scientists agree with him that many of Mars' unearthly and mysterious geological features are explained by the fact that Mars -- unlike Earth -- is cold enough for large amounts of CO2 to be stored in its subsurface.

This includes many who still think that liquid water may also have played a significant part -- it may be, for instance, that the catastrophic outflows did mix a large outpouring of liquid water with their huge CO2 gas eruptions, turning Mars into a gigantic "soda fountain".

At the new LPSC Conference, Heinz-Peter Jons suggested that the sinuous ridges seen in Mars' southern polar regions may not have been deposited by ancient glaciers (as James Head thinks), but may actually be "ripple marks" of frozen mud left behind by a giant eruption of volcanically heated underground liquid water that actually froze solid and then dried out in mid-flow on Noachian Mars' supercold surface.

(Even Hoffman says he has some doubts about this idea.) And Kenneth Tanaka proposed that much of the enormous Marineris Valley itself -- the 3000-km long gash, linked to a network of branching side canyons, that is Mars' single most notable surface feature, and which most theorists think is a "stretch mark" produced by the swelling of Mars' vast volcanic "Tharsis Bulge" -- is instead in large part another vast area of cryoclastic eruptions triggered by Tharsis' volcanic heat: "We propose that [magma] intrusions and volcanic eruptions in the Valles Marineris region led to the formation of catastrophic debris flows... charged by CO2 to account for their ability to travel thousands of kilometers to the northern plains... We feel that the evidence is sufficiently compelling to suggest that magmatically induced catastrophic erosion has been a major yet under appreciated geologic process on Mars."

But Hoffman's "White Mars" model is very bad news for anyone hoping for life on ancient Mars.

He says flatly that all belief in significant amounts of liquid water on or near the surface of Noachian Mars, in which microbes could have evolved, is nonsense -- early Mars was even more frigid and airless than Mars today, and there was almost no liquid water anywhere above pores in the rock of Mars' warm interior kilometers below its surface.

There was plenty of water above there -- but it was all frozen solid, except for a few small local areas temporarily heated by local volcanism.

This wouldn't quite rule out the possibility of Martian life -- it might conceivably have been able to evolve in those underground depths, or in small local areas where volcanic activity did allow near-surface liquid water to exist -- but it would obviously make it vastly more unlikely.

For this reason, astrobiologists have been looking eagerly for potential loopholes in his theory -- and, in fact, there are a few.

For one thing, there's the fact that the craters on Mars' remaining ancient Noachian surfaces are far more eroded than even moderately younger ones on Hesperian terrain, which is one of the primary clues suggesting to geologists that Noachian Mars did have an atmosphere hundreds of times denser, which then disappeared fairly rapidly during the Hesperian.

Then there's the puzzle of the valley networks, most of which were formed during the Noachian, with their production slacking off and finally ceasing during the Hesperian (save a few that seem to have formed on the flanks of volcanoes during Mars' most recent "Amazonian" era).

They don't look at all like the catastrophic outflow channels; they really do look more like riverbeds -- very narrow, meandering channels with multiple tributaries merging into larger streams, that look as though they were carved out in a leisurely way by fluid flowing slowly along (or just under) the surface.