Requirements today for machines have moved beyond functionality intoefficiency and reliability, gears are no exception. The presented work focuseson the analysis of the measurement, evolution and effect of running-in on geardrives from a surface roughness and efficiency point of view. With no consen-sus on a definition or observation of running-in of gear drives, measurementsof both efficiency and surface transformation during the predefined running-inis explored. A verified methodology on how to separate form, waviness androughness is presented. Two finishing methods, namely generation groundand superfinished, are analyzed in terms of efficiency and surface characteris-tics as manufactured, after running-in and after efficiency testing in order todetermine the effects of load level during running-in.Results show that separation of form can be achieved with a carefullychosen polynomial, while waviness is more subject to how the user definesa cut-off wavelength for the surface roughness. Ground gears show distinctsmoothening in terms of surface roughness at high running-in load, and nogeneral trend for low load. This behavior is also reflected in the efficiencysince higher loads gave overall lower efficiency after running-in when com-pared to lower loads. Superfinished gears in contrast show no running-in ef-fects in terms of efficiency. Additionally, ground gears exhibit general changein friction and surface roughness during the first cycles of running-in whenanalyzing high load. Overall gains in efficiency can be obtained from running-in; however, at most speeds improvements from polishing a surface, in thiscase superfinishing, proved to lead to higher efficiency. / <p>QC 20150922</p>