Some aspects, by a statistical approach, of the variation of the peak torque and mean cutting power encountered in gear hobbing are presented here. The cutting torque is measured by a purpose built strain gauge dynamometer and a wattmeter evaluates the power consumed. The problem is to determine whether a host of independent variables such as axial feed, module, material hardness, number of teeth in the gear blank and hobbing speed have a significant influence on the dependent variables, maximum cutting torque and average power. Mathematical models are constructed giving the relationship between these variables and the results indicate that the cutting speed has a significant influence on both the peak torque and average power. This significant point has not previously been noted for this type of process.