Morton (1969, 1975) proposed that the actual and potential evapotranspiration may behave in a complementary fashion about some definable constant as the regional surface dries from an initial saturated condition. He supported this hypothesis with an analysis of five-year means of water balance data for drainage basins in the United States, Canada and Ireland. In this paper we examine the validity of the assumptions used in the derivation of this complementary relationship. We approach the problem using a physically realistic diagnostic model of the energy fluxes within a closed system. From a data set derived from a series of model simulations, the probable range of error attributable to each assumption is evaluated. For the experimental design considered in the original hypothesis and replicated in our diagnostic model, we conclude that specifying the sum of actual and potential evapotranspiration as a constant is physically unrealistic. The errors associated with this complementary relationship may be as large as the magnitude of the variable being estimated.