Experiments

Heritability

The most obvious thing that needs doing is to demonstrate that information can be transmitted through a sequence of clay replicators. This is clearly possible when the layers of the clay are still joined together - but it that the information can be transmitted reliably when the crystals break is less well demonstrated.

I suggest the use of crystals with a acicular (needle-like) habitat - i.e. long thin crystals rather than short, flat ones.

If the crystals are fixed at one end, breaks are more likely to arise at the point where they are fixed - and - for a number of reasons, these breaks are less likely to result in a good site for new growth from the existing template.

I suggest attempting to use a "free fall" environment for the crystals - one where they are not liable to attach to the walls of the container too much.

Selecting for cross-section shape

However, for the sake of simplicity, I've tried to think of a selection criterion which doesn't depend on the presence of organics.

Inclined plane

Selection for large crystals is an important reason for using an inclined plane.

Seeking large crystals should make detection easier. It should also mean that the most successful organisms are likely to exhibit a non-trivial phenotype.

The water should be allowed to move down the slope (rather than remaning fixed (e.g in a pipe). The crystals should be kept in constant motion by the motion of the water.

The incline should probably be relatively shallow. Too much of a slope and either the water will generate a lot of turbulence as it flows - or the larger crystals will wind up falling at some speed through the water, causing them to be more likely to dissolve again.

There are two main possibilities for selection on such a slope:

• Selecting for rapid motion;
  • If crystals can be selected for rapid speed on a watery inclined plane, then it might be possible to produce an evolved rod with a circular cross section:
    

    These would move fastest - not so much because they would roll along - but since they have the smallest area in contact with the ground.

    A positive feature about trying to obtain an object like this is that there is selection for large size - since large objects will be more affected by gravity - and less affected by friction.

• Selecting for slow motion;
  • Trying to find organisms that tend to stay where they is the other main possibilty.

    If trying this it might be necessary to simultaneously use some sort of sieving system to select for large size - and make sure that small organisms are washed away.

    It is not terribly clear which shape of crystal this sort of selection would tend to produce, but - for the sake of argument - say it looks like this:

    

I'm inclined to think that selecting for rapid motion is the most attractive option.

Adaptation

Distinguishing between these possibilities may prove to be challenging.

Showing that the resulting entities roll better that other crystals (which had not been subject to selection for rapid motion) would not be sufficient - since round grains of sand are likely to roll faster than square ones - and yet are not the product of natural selection.

Showing that the crystals are closely related would show inheritance, though:

Natural selection would produce a small number of species - each with a characteristic size.

Frictional forces would produce a more continuous range of sizes.

This difference should allow the hypothesis of natural selection to be confirmed or rejected.