An experiment was run to determine the optimum frequency of the rotary knife on a roller gin stand. Six treatment levels of rotary knife frequency, 150, 300, 450, 600, 750, and 900 rpm, were chosen. (The manufacturer's specification for rotary knife frequency on the gin stand tested is 389 rpm.) The factors considered in determining the optimum rate included fiber and cottonseed quality, lint turnout, bale value, and ginning rate. There were no significant differences due to treatment with respect to fiber properties. Overall, fiber strength and elongation by the Stelometer averaged 27.8 g/tex and 6.98%, respectively; fiber span length and uniformity by the Fibrograph averaged 3.52 cm (1.38 in.) and 47.5, respectively; and fiber grade, staple, and micronaire by the USDA-AMS Classing office averaged 2.17, 46.8 32nd-in., and 38.1, respectively. These fiber properties are on samples taken from the bale and included two mill-type/air-jet lint cleaners. With the exception of linters content, there were no significant differences due to treatment on the cottonseed properties. Overall, foreign matter content of the cottonseed averaged 2.16%, and cottonseed grade averaged 114. Linters content ranged from 1.67 to 2.88% on the 300 and 900 rpm treatment, respectively. Lint turnout and bale value were not significantly different due to treatment, averaging 35.3% and $438.71 overall. Ginning rate was significantly different due to treatment; the highest rate occurred on the 300 rpm treatment and was 350 kg/m/h (1.90 bales/h). Using regression analysis, the optimum rotary knife frequency occurs at 269 rpm. At rotary knife frequencies less than 269 rpm, some of the ginning area on the roller is wasted, and roller damage is possible due to tags that cannot be cleared. At rotary knife frequencies greater than 269 rpm, compromises are made between ginning rate, seed cotton carryover, and choke-ups.