Information on the availability of different soil phosphorus (P) forms is useful for crop production. Phosphorus contents of 12 Iranian calcareous soils from upper-, mid-, and lower-slope positions of two and and two semiarid toposequences were fractionated to various organic and inorganic pools, and correlations of the P fractions with wheat responses were investigated. Among the inorganic P (IP) fractions, apatite type (Ca-10-P) and dicalcium phosphate equivalents (Ca-2-P) possessed the highest and the lowest amounts of P reserve in the soils, respectively. On average, about 20% of the total P was found in organic form (OP), of which 32% was labile (LOP), 51% was moderately labile (MLOP), and 17% was nonlabile (NLOP). The amounts of the soil P fractions were considerably influenced by the positions of the soils on the landscapes. The maximum contents of soil IP, Ca-2-P, Fe-P (iron-bound P), and Ca-10-P were observed in the lower-slope positions. The amount of soil available [0.5 M sodium bicarbonate (NaHCO3) extractable] P was significantly correlated with Ca2P (r = 0.895), Fe-P (r = 0.760), and Occl-P (iron-occluded P) (r = 0.897). Direct correlation studies, however, showed that wheat shoot dry-matter yield (DMY) was significantly affected by the amounts of Ca-2-P, Fe-P, OP, LOP, and MLOP fractions both at early (4 weeks) and late (10 weeks) stages of growth. All organic and inorganic P fractions, except Al-P (aluminum-bound P), Ca-8-P (octacalcium phosphate equivalents), and NLOP, also showed significant relations to the amount and/or concentration of P in wheat tissues at 4 and 10 weeks after sowing. Among the measured soil properties, the amount of organic carbon was the most affecting factor on the size of the P fractions.