For 41 soils (pH > 5.0), the Langmuir equation was an excellent model for describing P adsorption from solutions < 10-¯³M P, if it was assumed that adsorption occurs on two types of surface of contrasting bonding energies. In most of these soils, which were very undersaturated with P, > 90% of the existing adsorbed P occurs on the high energy surface. However, the adsorption capacity of the low-energy surface is usually at least double that of the high-energy surface. In 24 calcareous Sherborne soils, the specific surface areas of the carbonate component were an inverse hyperbolic function of the CaCO₃ content, and ranged from 16 to about 500 m²/g. The total surface areas were a slightly inverse and linear function of the CaCO₃ content, and ranged from 4 to 8 m²/g soil. In multiple regression studies on the Sherborne soils, high-energy P adsorption capacity was mainly associated with dithionite-extractable Fe (r² = 0.81), while pH had a large negative effect (r² = 0.71). Low-energy P adsorption was most closely correlated with organic matter content (r² = 0.53) and CaCO₃ surface area (r² = 0.46). From these results, together with results of P adsorption studies on the limestone parent materials, it was postulated that low-energy P adsorption occurs on CaCO₃ surfaces on which organic anions already occupy alternate adsorption sites. The Langmuir equation was used to calculate the theoretical maximum buffer capacities and the equilibrium buffer capacities of two different groups of soils. In multiple regression studies with P intensity or quantity measurements, these estimates of buffer capacity accounted for up to 94% of the variance in P uptake by ryegrass on these soils. The results showed that when soil P is measured as quantity, its availability (ie ease of desorption) is inversely related to the Langmuir bonding energy parameter and the buffer capacity. When measured as intensity, its 'availability' (ie resistence to change) is directly related to the adsorption capacity and buffer capacity.