Slowing of inactivation at positive potentials in a neuronal K+ channel is not due to preferential closed-state inactivation. (October 2000)