There are many requirements for dispersion models in the Clean Air Act Amendments of 1977 including prevention of significant deterioration, air quality maintenance plans, and new source permits. The historical perspective of dispersion modeling includes both theoretical and practical developments, rural and urban field studies including tracer studies, plume rise, and urban modeling. Photochemical modeling includes both grid point and trajectory models. Off-the-shelf models include level terrain models for elevated point sources, urban point and area sources, and transportation sources; and complex terrain models for point sources. Models are needed for photochemical pollutants, sulfates, and nitrates; they should treat long range transport, topographic effects, build-up of urban concentrations, and aerodynamic downwash. Emphasis is placed herein upon proper interpretation of vertical temperature structure measurements, and the difficulties in determining the magnitude and locations of the highest and second highest concentrations in a year. Modeling improvements can be expected with representative on-site measurements of input parameters, and improved dispersion parameter schemes which must be verified with appropriate field investigations. Terms related to model evaluation are defined. A primary source of error, in comparing model estimates with monitored air quality, is not including the actual change of wind direction with height which determines elevated plume position. Measures for model evaluation will differ depending upon model use. The paper concludes asking for 1) encouragement of model development, 2) improvement of model inputs through on-site measurements, and 3) detailed field investigations.Mr. Turner is presently on assignment from the National