Choice and reinforcement delay.