Characterization of phosphorus (P) in soils is important both agronomically and environmentally, although the outcome may depend on the technique applied. Consequently, we evaluated fractionation and adsorption, individually and jointly, and relevant ancillary soil attributes, to determine the dominant functional characteristics of soil P in 32 fertilized temperate grassland Inceptisols classified by eight soil series, and by two soil-P index and parent material groups. Residual P was low (30.7%) and organic P (Po) prominent, 42.0% vs. 17.5% for equivalent soils in unfertilized natural ecosystems. Labile fractions comprised 6.8% inorganic P (Pi) and 9.1% Po. The proportional increase in high vs. low index soils (Morgan P > 6.0 mg l(-1) vs. <= 6.0 mg l(-1)) was higher for Pi, and highest for labile and moderately labile fractions. Only moderately labile Pi and Po differed significantly between soils of limestone and non-limestone origin. Oxalate extractable Fe (Fe-ox) and buffering (EBC) were higher in the latter. The equilibrium P concentration (EPC) was substantially higher in the high index group, and EBC and binding energy (k) substantially lower, with no significant difference in sorption maximum (Pmax). EBC equated with weak to strong buffering in different soil series, and conformed better than k to ancillary attributes. Pmax correlated in order Al-ox > clay > OC > Fe-ox, and more broadly reflected sorption attributes than oxalate-based sorption capacity (PSC). Principal component (PC) analysis showed consistent differentiation of P fractions, mostly labile and moderately labile, in PC 1 vs. adsorption and ancillary attributes in PC 2. However, scatterplots of PC scores showed that adsorption characteristics provided better functional differentiation than P fractions for distinguishing individual soil series, which may have implications in selection and interpretation of extractants not only for environmental but also for agronomic soil-tests.