We postulate that phosphorus (P) fertilization may increase above-ground net primary productivity (ANPP) of rotationally grazed rangelands without reducing the legume component, as does N fertilization. In doing so, we evaluated the effect of phosphate fertilization on the production and relative contribution of legumes and grasses of native and old tall fescue (Festuca arundinacea Schreb) grasslands; we recorded annual production, seasonal productivity, and biomass contribution of each component. The experiment was conducted in a commercial farm located in the Flooding Pampa and managed under rotational grazing. Treatments consisted of two fertilization programs (66 (P₆₆) and 29 (P₂₉) kg P · ha⁻¹ supplied as rock phosphate and/or monoammonium phosphate from 1997 to 1999) and a nonfertilized control. A paddock dominated by native grassland and another dominated by old tall fescue grassland were selected. Nine 5-ha plots were established in each paddock, and treatments were randomly assigned. During the experimental period, from October 1998 to October 1999, total above-ground biomass was harvested from each plot before and after each grazing period and separated into components: tall fescue, other C₃ perennial grasses, legumes, C₃ annual grasses, C₄ grasses, forbs, and standing dead material. ANPP of each component was estimated during the warm (October 1998–February 1999) and the cool (March 1999–September 1999) season. In native grassland, phosphate fertilization increased ANPP of C₃ annual grasses and legumes during both the warm and the cool seasons; therefore annual ANPP of the grassland under P₆₆ was 40% higher than under P₂₉ and doubled ANPP of nonfertilized plots. Phosphate fertilization didn't increase total annual ANPP of old tall fescue grassland, but it did increase ANPP of legumes during both seasons.