While the rover teams maintain a respectful silence about their ponderings, I'll air my linen in public...

<Description=ON>

The pale rocks are exquisitely fine grained and layered. They are massive but weather to reveal a weak cross-bedded parallel stratigraphy with local undulose bedding. Some layers weather to granular or nodular textures.

The rocks contain two types of interruptions to the bedding:

1: the beads of bluish material. These exist within the layered host without any interruption of the layering. No drape, no forceful displacement.

2: A triangular array of what looks like cracks but on close inspection appear to be the voids left by flat crystal plates growing in a semi-regular 3D interlocking array. Again, no forceful replacement or apparent disturbance of the layering. These plates appear to be disc-like with tapered edges/ends

Interpretation of sequence of events:

1) a mixed sulphate(?) is deposited in exquisite thin layers
2) a network of interlocking plates grows by segregation of one of the constituents of the matrix. This is similar to Desert Rose crystallisation on Earth.
3) the beads grow around random centres by segregation of the metallic(?) constituents of the matrix
4) The plates are removed to form voids (This may occur through deep surface weathering, or through fluid solution)
5) The beads weather out to form a lag deposit

Most of the above is descriptive, and minimally interpretive, IMHO.

Now, the $64,000 question:- what was the environment of deposition?

<Interpretation=ON>
<Speculation=HIGH>

I favour a sabkha analogue, with a regional water table (yes, water!) slowly welling-up over a regional extent. In other white rock locations the upwelling is more focussed in central peaks of craters or offcentre locations.

As the saturated brines reach the thin atmosphere, they evaporate by day and freeze by night. (the interlocking "crystal" plates could actually be ice lenses formed in the immediate near-surface).

Crystallisation proceeds to add layers to the surface by direct chemiocal precipitation, either of paper-thin layers or of granules/nodules at/near the surface.

Once a thick body has formed and been buried, diagenesis/early metamorphism accomplishes step 2 (if not already formed at surface) and step 3.

In this sabkha-like environment we can generate layered material over the regional extent of Meridiani, given groundwater activity and available minerals. At no time need there be pools of liquid water - merely a moist subsurface.

(one could posit that the whitish soils exposed during trenching by both rovers are a similar process occurring slowly on Modern Mars during favourable climatic periods as a type of caliche within the soil horizon, but I'm not sure if I want to go this far just yet!)

The "sabkha" surface would be self-smoothing, draping and infilling relief and building a layered cap burying earlier terrain. It would be self-healing from impact scars and incredibly smooth on length-scales from cm to hundreds of km.

Within impact craters, either regional groundwater tables intercept the crater base and begin accumulating layers of material, or focussed fluid activity builds a mound which can potentially overtop the crater walls, as observed by Malin and Edgett for some accumulations.

This is all very "wet", I must admit. Nonetheless, it's at the dry saline end of wet, not the gushy splashy end.

<Speculation=NORMAL">
<Interpretation=OFF>

I reserve the right to rework all of the above, in light of better elemental and mineralogical data than "high sulphur"

(And if RObert CLark could refrain from saying "I told you so", we might be able to get more data to improve or refute this hypothesis)