Gravar-mail: Enhancement of conditioned reinforcement by uncertainty