In Morocco,response of dryland wheat (Triticum aestivuro) to ? has not been related to soil ? availability index. The objectives of the experiments reported were to determine the critical NaHCO3‐extractable ? level in soils of Settat Province in Morocco and to compare the effect o‐f banding with broadcasting of concentrated superphosphate (CSP) for wheat, crops. Eighteen ?‐P factorial experiments were conducted on three different soil types (Rendolls, Caloixerolls and Chromoxererts) with different ? availability indexes and lime levels in 1985–86 and 1986–87. The ? rates were 0, 9, and 18 in 1985–86 and an additional 36 kg P/ha rate in 1986–87. The rainfall levels were 300 mm with an excellent distribution and 220 mm with a poor distribution in the above years, respectively. In the latter year, a P placement study was conducted on a Rendoll with a NaHCO3‐P level of 4.5 mg/kg. For the placement study the rates were 0, 9, 18, 36, and 54 kg P/ha. The grain yield response to ? averaged 8% in the wet year and 40% in the drier year. The maximum yields obtained were close to 3 and 1 Mg/ha for the above years, respectively. A critical level of 5 mg/kg of NaHCO3‐P was obtained, separating responsive from nonresponsive soils for a relative yield of 90%. The ? placement study showed that yield response to broadcast ? reached a plateau at 9 kg P/ha, but the response to banded ? continued up to the highest rate. This shows that the banding effect of CSP is more than a simple ? nutrition effect