In a pot experiment, the P-efficient wheat (Triticum aestivum L.) cultivar Goldmark was grown in ten soils from South Australia covering a wide range of pH (four acidic, two neutral, and four alkaline soils) with low to moderate P availability. Phosphorus (100 mg P kg(-1)) was supplied as FePO4 to acidic soils, CaHPO4 to alkaline, and 1:1 mixture of FePO4 and CaHPO4 to neutral soils. Phosphorus uptake was correlated with P availability measured by anion-exchange resin and microbial biomass P in the rhizosphere. Growth and P uptake were best in the neutral soils, lower in the acidic, and poorest in the alkaline soils. The good growth in the neutral soils could be explained by a combination of extensive soil exploitation by the roots and high phosphatase activity in the rhizosphere, indicating microbial facilitation of organic-P mineralization. The plant effect (soil exploitation by roots) appeared to dominate in the acidic soils. Alkaline phosphatase and diesterase activities in acidic soils were lower than in neutral soils, but strongly increased in the rhizosphere compared with the bulk soil, suggesting that microorganisms contribute to P uptake in these acidic soils. Shoot and root growth and P uptake per unit root length were lowest in the alkaline soils. Despite high alkaline phosphatase and diesterase activities in the alkaline soils, microbial biomass P was low, suggesting that the enzymes could not mineralize sufficient organic P to meet the demands of plants and microorganisms. Microbial-community composition, assessed by fatty acid methylester (FAME) analysis, was strongly dependent on soil pH, whereas other soil properties (organic-C or CaCO3 content) were less important or not important at all (soil texture).