A question which I think needs to be worked out (if it hasn't been) is given an Earth like life, what is the theoretical limit of dehydration that would still be compatible with a functional biochemistry? On Earth water is a lot more than a solvent, it is an organizer (by establishing hydrophilic/hydrophobic boundaries), it is a reagent (both hydration and dehydration reactions are important), and it is a linker in H-bridge formation so I can't imagine it being dispensed with, but could its role be minimized? We know how much water we need here and now, but suppose the Earth slowly lost water, could cells evolve to downplay their dependence on water (at least in principle), and if so at what limit would any form of life become untenable? Or would cells evolve new ways to secure water from a dry environment? My guess is that given a highly efficient mechanism for gathering water from a dry environment cells may not need vast reserves of external water. Some biochemical environments exclude water or use it in very limited and controlled conditions (for example the inner lipid bilayer or the hydrophobic core of many proteins. The cell is wet but it is more a gel than soup. The colloidal nature of the cell requires that it burns energyb to protect itself from osmotic damage. A long time ago I saw an article that compared some enzymatic reactions to gas-phase chemistry suggesting that solvent water exclusion was critical to at least some biochemical reactions. This interested me at the time but my brain has decayed along way down its inevitable half-life to senility since then.

regards,

Andrew