Looks reasonable to me. One might raise the question of whether markerfacecolor etc kwargs should be checked along with the color kwarg. The problem with doing that, though, is that someone might want to specify markeredgecolor with a kwarg, while letting marker face colors cycle. That suggests that restricting the check to the "color" kwarg is OK.

I think this patch is reasonable. And I agree that this should only be focused on 'color'. If we want to cycle markeredgecolor, we can address that separately.

It is not a question of wanting to cycle markeredgecolor, but rather of detecting when a user doesn't want the cycle to be advanced. Still, at this point, going with the patch seems reasonable. I don't have time for it right now; if you do, go ahead and merge it.