An injection technique, in which undisturbed soil cores are labelled with ³²P to study dissolution of phosphate fertilisers in the soil, was evaluated in field and glasshouse trials. When ³²P was injected between 0-150 mm depths of the undisturbed soil columns and fertilisers applied at the surface, the amounts of fertiliser P dissolved, as measured by the increases in the exchangeable P pools, were overestimated. Three possible reasons were suggested: (i) the interaction between surface-applied fertiliser, ³²P injected through the whole soil column, and the vertical decline in root density, (ii) the decline of specific activity in the exchangeable P pool due to losses of ³²P to nonexchangeable P pools and continuous addition of P from fertiliser dissolution, and (iii) non-uniform distribution of ³²P vis-a-vis ³¹P phosphate. The injection technique may be employed to assess the effectiveness of phosphate fertilisers by introducing a concept, the fertiliser equivalent (FE). The FE is a measure of the amounts of soil exchangeable P that the fertilisers are equivalent to in supplying P to plants, when applied at the specific location. Soluble single superphosphate (SSP) applied at the surface of undisturbed grassland soil cores (Tekapo fine sandy loam), was much more effective than surface-applied unground North Carolina phosphate rock (NCPR) and 30% acidulated NCPR with phosphoric acid (NCPAPR) within the 56 day period of plant growth. An isotopic dilution method, based on tracer kinetic theory, was developed to study the rates of dissolution (F in) and retention (F out) of phosphate fertilisers in the soil in growth chamber experiments. The estimation of F in and F out required labelling of the soils with carrier-free ³²P and determination of the corresponding values of the specific activities of the exchangeable P pools, SA₁ and SA₂, and the sizes of the exchangeable P pools, Q₁ and Q₂, at times t₁ and t₂. Most of the phosphate in the monocalcium phosphate (MCP) solution entered the exchangeable P pool immediately after addition to the soils (Tekapo fine sandy loam and Craigieburn silt loam), and there was little further phosphate input. With increasing periods of incubation, the phosphate was quickly transformed to less rapidly exchangeable forms. In the soils treated with ground North Carolina phosphate rock (<150 µm, NCPR) or partially acidulated (30%) NCPR with phosphoric acid (NCPAPR), the initial exchangeable P pools were not as large as those in the soils treated with MCP, but were maintained at relatively stable concentrations for extended periods, due to the continuous dissolution of PR materials and to lower rates of pretention. An increase in P-retention caused a slight rise in the rate of PR dissolution, but also a rise in the rate of P-retention by the soil. The rate of dissolution was higher at a lower application rate in relative terms, but smaller in absolute terms. The trends in the changes of plant-available P in the soils, measured by the water extractable P, Bray I P and Olsen P, correspond to those predicted by the F in and F out values. The average rates of dissolution between 1-50 and 50-111 days estimated by the F in, however, were higher than those estimated by extractions with 0.5 M NaOH followed by 1 M HCl, and with 0.5 M BaCl₂/TEA. This is partly because the Fin values reflect a plant growth effect on PR dissolution. The relative agronomic effectiveness of NCPR and NCPAPR with respect to MCP was higher after 50 and 111 days of incubation than after 1 day. The F in values were included in all the two-variable models constructed by stepwise regression to describe the relationship between plant P uptake and soil measurements. The amounts of variation in plant P uptake accounted for by the regression model was significantly improved by including F in in the model. This indicates the importance of fertiliser dissolution rates in affecting soil P supply, when phosphate fertilisers differing in solubility are applied.