Three experimental sites were established at the Bangladesh Agricultural Research Institute (BARI) experimental farms at headquarters at Joydebpur, the Wheat Research Center at Nashipur, and the Regional Agricultural Research Station at Ishwordi. The Joydebpur site is within Agro-ecological Zone (AEZ) 9 (Old Brahmaputra Floodplain), on flood-free, flat, medium highland with fine-textured (silty clay loam) and low-permeability soil. The Nashipur site is in AEZ 1 (Old Himalayan Piedmont Plains), on flood free highland with coarse-textured (sandy loam) and high-permeability soil. The Ishwordi site is in AEZ 11 (High Ganges River Floodplain), on flood-free highland with fine-textured (silty clay loam) and medium-permeability soil (BARC, 1997). Experimental Design, Treatments, and Crop Management Field experiments at each site compared RW cropping systems with maize or mungbean as pre-rice crops under STB and FP fertilizer management. The experiments commenced with wheat sown in late November/early December in 1995 at Joydebpur and Nashipur, and in 1997 at Ishwordi. A nutrient-extractive maize crop was sown at each site before the start of the experiments to identify and minimize any site heterogeneity. A randomized complete block design with four replications was used at each site, with two main treatments: pre-rice crop management (three levels) and fertilizer management (two levels). Pre-rice treatments included mungbean, with its residues either retained or removed, and maize with its residues removed. In the mungbean-residue-retained treatment, after the pods were harvested, the residues were incorporated into the soil; in the residue-removed treatment, all root and shoot residues were removed from the field. In rice, wheat, and maize, all aboveground residues were removed but roots were incorporated into the soil. Fertilizer treatments included the recommended complete fertilizer dose to achieve high yield (STB) and the typical fertilizer regime used by farmers in each location (FP). Fertilizer N, P, and K for rice across sites ranged, respectively, from 87–120, 18–26, and 0–50 kg ha−1 under STB and from 60–80, 9–26, and 17 kg ha−1 under FP. In wheat, the ranges were 120–140, 15–26, and 33–66 kg ha−1 under STB and 60, 11–18, and 25 kg ha−1 under FP, respectively. Maize received 80–120, 18–30, and 0–33 kg ha−1 under STB and 60–80, 9–26, and 17–33 kg ha−1under FP, respectively, while mungbean received 15–20 kg ha−1 N, P, and K under STB but no fertilizer under FP. Each fertilizer treatment was divided into two N rate sub-plots in a splitplot design. In the +N sub-plot, N was applied at the STB or FP rate, while in the –N sub-plot, all the nutrients were applied according to STB or FP, except there was no application of N, thus resulting in 12 treatment combinations. All plots and sub-plots were separated by earthen banks lined with plastic to a depth of 30 cm. Tissue P concentration was measured and analyzed from only seven treatments: six+N treatments (T1–T6) and one –N treatment (control or T7). The control plot was grown to maize without N but with other nutrients as applied by farmers. All plots and subplots were separated by earthen banks lined with plastic to a depth of about 30 cm. Details of the experimental design, treatments, fertilizer doses, planting dates, crop management, and sampling procedures for yields of individual crops and sites, as well as weather description are provided elsewhere (Quayyum et al., 2001, 2002a, 2002b). Chemical Analysis of Soil, Water, and Plant Samples Soil samples were collected from each site (0–15 cm) before (“initial”) establishing the experiments, and after (“final”) harvest of rice in 2000 from six random locations in each subplot (0–15 cm). Straw and grain samples, and the initial and final soil samples, were analyzed for total P by di-acid digestion (Yoshida et al., 1976). Crop P uptake was derived from the P concentrations and the straw and grain yields (Quayyum et al., 2001, 2002a, 2002b). The P concentrations in irrigation and rainwater were determined occasionally and means were used to calculate the amounts of P added to the soil through irrigation and/or rain.