This paper finds, in a particularly simple form, the solution to the optimal (infinite-horizon, known distribution) stopping problem when generalised to the case where the reward function is not monotonically related to the random variable under observation. © 1979.