Counterexamples to a commonly quoted statement of the small-mu theorem are exhibited. A detailed proof of a correct statement is provided. It includes the case where the uncertainty is not constrained to be real-rational.