The effectiveness of three different techniques, designed to improve the bending fatigue life in comparison to conventionally processed gas-carburized 8620 steel, were evaluated with modified Brugger bending fatigue specimens and actual ring and pinion gears. The bending fatigue samples were machined from forged gear blanks from the same lot of material used for the pinion gear tests, and all processing of laboratory samples and gears was done together. Fatigue data were obtained on standard as-carburized parts and after three special processing histories: shotpeening to increase surface residual stresses; double heat treating to refined austenite grain size; and vacuum carburizing to minimize intergranular oxidation. Standard room-temperature S-N curves and endurance limits were obtained with the laboratory samples. The pinions were run as part of a complete gear set on a laboratory dynamometer and data were obtained at two imposed torque levels. The number of cycles to failure was used to evaluate the effects of processing history. Based on laboratory endurance limits, shown in parentheses, the processing histories ranked as follows: shot-peened (1410 MPa), vacuum carburized (1210 MPa), reheated to refine grain size (1140 MPa), and asgas-carburized (1000MPa). In the gear set tests, shot peening also proved to be the most effective way to improve fatigue life at both imposed torque levels. The results of this study show that data on laboratory samples can be used to interpret the fatigue performance of gears.