[mannykannot]

Start with a dry surface and consider the cone of light, scattered from a point on it, that will pass through your iris. If, as shown in the final diagram, the water, when added, forms a flat surface, and you are looking at it fairly perpendicularly, then none of the rays that are totally internally reflected would have been in that cone and ended up entering your eye anyway - they are too oblique. They do not contribute to the brightness either before or after the water is added.

On the other hand, when water is added, the above cone of light is broadened, on account of refraction, so it is no longer the case that all the light within it will enter your eye.

These considerations must be modified (I am not sure how) if the water forms a thin, capilliary-adhering film that conforms to the roughness of the substrate.

Finally, if the substrate is transparent or translucent (e.g. sand (quartz crystals)), the change of refractive index at its surface is lessened in the presence of water, causing more transmission / absorbtion and less reflection / scattering there.