The capacity of a stirred ball mill to grind to a certain product size efficiently depends strongly on the power intensity in the milling chamber. The dependence of power intensity on stirrer speed, slurry density, bead density and the amount of dispersant added to the feed was investigated using a 6-litre Il-kW horizontal stirred ball mill fitted with perforated disks as stirrers. Dolomite was ground at a fixed volumetric flow rate, but with slurry densities ranging from 65 to 75 percent by weight, stirrer speeds from 805 to 2253 rpm, bead densities from 2.5 to 5.4 g/ml and the dispersant level from 0.5 to 1.5 percent of dry solids. Three levels of the four variables were used in 27 continuous milling tests, set up as one-third of a 3(4) factorial design. All factors affected the power draft in a highly significant way. Speed, and to a lesser extent slurry density, were the dominant factors with significant non-linear effects. A six-term model, incorporating all significant effects, predicted the experimental results with an accuracy of about 12%, Increasing the power accelerated size reduction dramatically with only a small change to the energy efficiency of the process.