Regardless of bank angle, there is still the same amount of air passing under and over the wings; they do not know if they are horizontal or not.

I'm a mechanical engineering student. I haven't taken fluid dynamics yet (next semester ), but I believe that the dihedral issue is more based on simple mechanics.

Imagine, for an exaggerated example, that the dihedral was 90 degrees, so that when the plane is flying level, the wings are each at a 45 degree angle. If you bank to the right so that the right wing is now horizontal (45 degree bank right) you will have an unbalanced wing with the weight of the plane completely at the left end. Gravity will pull down on the plane's CG, causing it to roll back to the left, until the moment force of gravity is balanced on the left side by the left wing generating some vertical lift.

Now imagine the same plane with no dihedral in a 45 degree bank. Gravity will have a much harder time 'righting' the plane because there is vertical lift being generated by the left wing, and less vertical lift being generated by the right wing.

Unfortunately I haven't flown a model plane yet! This theory would be proved if the dihedral had less self-righting characteristics at higher speeds, which means that as the lift force of the wings increases (in the direction perpendicular to the surface of the wing), the effect of gravity becomes negligible. If you could fly really, really fast in a plane with dihedral, it should have almost no self-righting characteristics at all (if my postulate was correct).

Let me know what you think!