The dune trains in the two images above, if viewed as slow flows of sand, silt and perhaps dust operating over aeons, adequately explain the formation of Martian channels and valleys, without liquid water.

It has been assumed by many observers that much of the present Martian surface is either exposed or lightly buried lava or fluvial sediments, mainly on the basis of the layering seen exposed in canyon and crater walls. The cliffs collapse far too readily and symmetrically to be made of lava, in my view. Where are the huge tumbled or partly slumped blocks of lava cliffs?

There is a possible alternative to this. The material may be unconsolidated or poorly consolidated aeolian dust and silt, with the banding the result of pedogenic processes, repeated after each sheet of wind-blown material was laid down.

As there is some atmospheric water vapour on Mars, in winter some water must freeze out of the air at the top of the permafrost layer, and then in summer melt and interact with the lithic material and gas around it to form laterites, carbonates and perhaps bauxites.

Repeated cycles, as the wind deposits new layers, may give the observed layering. This groundwater may of course contribute to a greater or lesser degree to sapping and slumping of the walls in the valleys, channels and craters, as Malin and Edgett have postulated.

Each surface could generate two lateritic horizons, as seen in some parts of Western Australia, corresponding with summer and winter permafrost levels.

Boulder measuring metres or decametres across can be seen to have tumbled down some Martian valley and crater slopes. I have seen just that sort of thing below bauxite-tending-to-laterite cliffs on the west side of Cape York, at the northern most tip of Eastern Australia.

Is it possible that Yogi and Barnacle Bill (Pathfinder rocks) and their neighbours were pisolitic laterites, from a surface fragmented by meteorite impact, and then deflated, with all the small bits blowing away and breaking down?

It would be easier to account for laterite scattered all over the surface than fresh igneous and volcanic rocks, given the obvious oxidation at the surface.

The surface of Mars has rusted, after all, it's red from iron oxide. The alpha proton x-ray spectrometer data show all the right elements, and it is notoriously difficult to identify minerals in a rock and hence rock types, from photos and elemental compositions alone. I shall ask NASA if this alternative can be ruled out with certainty.

The latest excitement in Mars geology is evidence interpreted by Mike Malin and Kenneth Edgett as indicating recent groundwater seepage on a limited number of predominantly pole-facing slopes (Science, 20.6.2000, pp 2330-2335. Evidence of Recent Groundwater Seepage and Surface Runoff on Mars).

It's a superb paper, full of brilliant and detailed observation and deduction. They may be right, but perhaps the fluid was not liquid water. I will now try to apply the above generalisations to these particular cases.

In their Note 1, Malin and Edgett state that the only alternative to liquid water is to speculate that some other completely unknown agent, with properties essentially identical to those of water is or has been at work on Mars.