This study compared potential evapotranspiration, computed from climatological data by each of six empirical methods, with pan evaporation adjusted to equivalent lake evaporation by regional coefficients. The six methods tested were the Thornthwaite, U.S. Weather Bureau (a modification of the Permian method), Lowry-Johnson, Blaney-Criddle, Lane, and Hamon methods. The test was limited to 25 sites in the arid and subhumid parts of Arizona, California, and Nevada, where pan evaporation and concurrent climatological data were available. However, some of the sites lacked complete climatological data for the application of all six methods. Average values of adjusted pan evaporation and computed potential evapotransp4ration were compared for two periods---the calendar year and the 6-month period from May 1 through October 31. The 25 sites sampled a wide range of climatic conditions. Ten sites (group 1) were in a highly arid environment and four (group 2) were in an arid environment that was modified by extensive irrigation. The remaining 11 sites (group 3) were in a subhumid environment. Only the Weather Bureau method gave estimates of potential evapotranspiration that closely agreed with the adjusted pan evaporation at all sites where the method was used. However, lack of climatological data restricted the use of the Weather Bureau method to seven sites. Results obtained by use of the Thornthwaite, Lowry-Johnson, and Hamon methods were consistently low. Results obtained by use of the Lane method agreed with adjusted pan evaporation at the group 1 sites but were consistently high at the group 2 and 3 sites. During the analysis it became apparent that adjusted pan evaporation in an arid environment (group 1 sites) was a spurious standard for evaluating the reliability of .the methods that were tested. Group 1 data were accordingly not considered when making conclusions as ,to which of the six methods tested was best. The results of this study for group 2 and 3 data indicated that the Blaney-Criddle method, which uses climatological data that can be readily obtained or deduced, was the most practical of the six methods for estimating potential evapotranspiration. At all 15 sites in the two environments, potential evapotranspiration computed by the Blaney-Criddle method checked the adjusted pan evaporation within ?22 percent. This percentage range is generally considered to be the range of reliability for estimating lake evaporation from evaporation pans.