For forty-one soils (pH > 5.0) from southern England and eastern Australia, the Langmuir equation was an excellent model for describing P adsorption from solutions < I O -~ M P, if it was assumed that adsorption occurs on two types of surface of contrasting bonding energies. For most of these soils, which were relatively undersaturated with P, this equation may be written as : k'xhc k x k c x = -I +k'c+x$-K"c' where x = adsorption, k = adsorption/desorption equilibrium constant, x, = monolayer adsorption capacity, and c = equilibrium solution concentration. The relative magnitude of the parameters for each surface were approximately: x& = 0.3 x h and k' = IOO k". More than 90 per cent of the native adsorbed P occurs on the high-energy surface in most soils.

The two-surface Langmuir equation was used to study P adsorption by 24 calcareous soils (pH 7-27.6; 0-8-24.2 per cent CaCO,) from the Sherborne soil series, which are derived from Jurassic limestone. High-energy P adsorption capacities ( x; ) ranged from 140-345 pg P/g and were most closely correlated with dithionite-soluble Fe. Hydrous oxides therefore appear to provide the principal sites, even in calcareous soils, on which P is strongly adsorbed (k' = 6-51 ml/pg P). The low-energy adsorption capacities ( x; ) ranged from 400-663 pg P/g and were correlated with organic matter contents and the total surface areas of CaC08 but not with per cent CaC08, pH, or dithionite-soluble Fe. Total surface areas of CaCO, in the soils ranged from 4-0 to 8.5 ma/g soil. Low-energy P adsorption capacities agree reasonably with values (100 pg P/m*) for the sorption of phosphate on Jurassic limestones but phosphate was bonded much less strongly by soil carbonates (K" = 0*08-0*45 d / p g P) than by limestones (k S 10-0 ml/pg P).Low-energy P adsorption in these soils is tentatively ascribed to adsorption on sites already occupied by organic anions (and probably also by bicarbonate and silicate ions) which lessen the bonding energy of co-adsorbed P.

This paper describes the accumulation of uranium in soils from superphosphate applied annually to arable and grassland soils. Rates of application of superphosphate were equivalent to about 33 kg P and 15 g U ha-' year-' in three experiments at Rothamsted and to about 37 kg P and 16g

A simple, sensitive method developed for the analysis of geostandards was used to measure the accumulation of Cd in soils from superphosphate applied annually to grassland and arable soils for many years. Rates of application were equivalent to 33 kg P and 5 g Cd ha-' yr-' for 95 yr in three experiments in England and to 37 kgP and 20gCdha-'yr-' for 30 yr in one experiment in New Zealand. Very little Cd accumulated in the surface horizons (0-22.5cm) of either of the arable soils from England; about one-quarter of the applied Cd was detected in the sub-soil (22.5-45.0 cm) of one experiment (Broadbalk) but none in the second (Barnfield). About one-half of the applied Cd was retained in the CL22.5 cm horizon of grassland soils from both England tnd New Zealand. The light ( <2.2 g C I T -~) organic-rich fraction of Park Grass soil from England contained about three times as much Cd as the heavier, mineral-rich fraction. This suggests that when Cd is incorporated into organic matter its mobility is decreased and soil pH then has smaller effects on its mobility. Uptake of Cd by grass-clover pasture in New Zealand averaged only 0.4 g Cd ha-' yr-' or 2% of the amount applied.

The effects of phosphate buffer capacity on the plant-avMlability of labile soil phosphate, when measured as intensity (I) or quantity (Q), are described and tested using results from a greenhouse experiment on 24 Sherborne soils. In multiple regression studies, phosphate buffer capacity with I or Q measurements as independent variables accounted for up to 94% of the variance in P uptake by ryegrass, the maximum buffer capacity being generally more useful than the equilibrium buffer capacity.When the quantity of soil P is measured (Q), its availability (i.e. ease of desorption) to plant roots is inversely related to the Langmuir bonding energy parameter and the buffer capacity. When the intensity of soil P is measured (I), its availability (i.e. resistance to change) is directly related to the adsorption and buffer capacities. The levels of Q or I, therefore, which are optimal for plant uptake vary with the buffer capacity of the soil. There is little or no correlation between the adsorption capacity and the bonding energy in many soils and consequently phosphate buffer capacity is only poorly correlated with the total adsorption capacity.