A network of long-term, on-farm fertilization experiments was established in 2000 by CREA, a consortium of farmers from the Southern Santa Fe region in the northern Pampas of Argentina. The objective was to evaluate direct and residual yield responses to N, P, S, and other nutrients (K, Mg, B, Cu, and Zn) in two crop rotations: corn-double cropped wheat/soybeans (C-W/S; five sites), and corn-full season soybeans-double cropped wheat/soybeans (C-S-W/S; six sites). Fertilization was annually applied to corn, wheat or full season soybeans (no N was applied on soybeans), always in the same plots with nutrient rates equivalent to grain nutrient removal + 10%, except for N. Rates of N were estimated from high-yielding N experiments in the region. Treatments included: 1) Check; 2) PS; 3) NS; 4) NP; 5) NPS; and 6) NPS plus K, Mg, B, Cu and Zn. Grain yield responses to NPS in corn, wheat and full season and double cropped soybean were significant in most of the years/sites. Corn and wheat yield responses to N were related to soil nitrate-N availability at planting. Response to P was observed in all four crops depending on the soil P Bray 1 levels. Response to S was more frequent in corn and double cropped soybeans, and relate to sulfate-S availability at planting. In general, no response to other nutrients was observed. Differences between the NPS and Check treatments have increased over the years as result of improved and depleted soil fertility in the NPS and Check treatments, respectively. Residual effects of balanced NPS fertilization were observed in the soil Bray 1 P levels and soil organic matter concentrations.