A preliminary dimensional analysis has been undertaken on a large agricultural atomization database, developed by the Spray Drift Task Force, with the goal of recovering fitting parameters that may be used to generate droplet size spectra in the absence of laboratory data. The results presented herein are encouraging but not definitive: correlation coefficients between 0.596 and 0.843, consistently lower than coefficients produced by two alternative methods-a statistical approach using multiple regression analysis, and a neural network representation-also summarized herein, possible reasons for the lower correlation coefficients with dimensional analysis, a technique that should provide direct physical insight and not simply fit the data, are explored.