Abstract: Many consider the Penman-Monteith equation conceptually more general and realistic and more universally applicable relative to other Penman-based estimators of reference canopy evapotranspiration (PET). They have recommended its worldwide adoption. However, its practical advantage over currently-used PET estimators still needs to be demonstrated especially in the semi-arid tropics. The question addressed here was whether Penman-Monteith estimates would better match the only existing set of thirty six, 5-day averaged, lysimeter observations of reference PET in semi-arid Botswana than the currently-used Penman-based estimator and, if not why not. In this preliminary test, both equations showed the same pattern of estimated PET, but the correspondence between measured values and estimates depended on the season. The Penman-Monteith estimates were closer to measured values during the fall and winter while the Botswana Department of Meteorological Services (DMS) equation better estimated the observations during the spring and summer. Sensitivity analyses showed that small and justifiable adjustments of the windspeed coefficient in the aerodynamic term of the DMS equation would enable better correspondence between estimated and observed values. PET estimates using the Penman-Monteith equation were comparatively much less sensitive to variations in the two resistance input parameters (ra and rs) needed for its application. It would require unjustifiable adjustment of these parameters to improve the correspondence between the Penman-Montieth estimates and the observed summer PET in Botswana vis-à-vis the existing DMS Penman-based equation. Overall, the findings underline the need for further long-term, multi-locational lysimeter measurements of reference crop PET in Botswana to more fully evaluate the applicability of the Penman-Monteith equation.