I present some puzzling cases regarding knowledge and its relation to rational credence. They seem to entail a failure of an apparently correct principle: if S knows P, then the epistemic justi...