Phosphorus availability in soils is controlled by both the sizes of P pools and the transformation rates among these pools. Rates of gross P mineralization and immobilization are poorly known due- to the limitations of available analytical techniques. We developed a new method to estimate P transformation rates in three forest soils and one grassland soil representing an Alfisol, an Ultisol, and Andisol, and a Mollisol. Three treatments were applied to each soil in order to separate the processes of mineral P solubilization, organic P mineralization, and solution P immobilization. One set of soils was retained as control, a second set was irradiated with GAMMA-rays to stop microbial immobilization, and a third was irradiated and then autoclaved, also stop phosphatase activity. All three sets of samples were then incubated with anion exchange resin bags under aerobic conditions. Differences in resin P among the three treatments were used to estimate gross P mineralization and immobilization rates. Autoclaving did not affect resin-extractable P in any of the soils. Radiation did not alter resin-extractable P in the forest soils but increased resin-extractable P in the grassland soil. This increase was corrected in the calculation of potential P transformation rates. Effects of radiation on phosphatase activity varied with soils but was within 30% of the original values. Rates of P gross mineralization and immobilization ranged from 0.6-3.8 and 0-4.3 mg kg-soil-1 d-1, respectively, for the four soils. The net rates of solubilization of mineral P in the grassland soil were 7-10 times higher than the rates in forest soils. Mineralization of organic P contributed from 20-60% of total available P in the acid forest soils compared with 6% in the grassland soil, suggesting that the P mineralization processes are more important in controlling P availability in these forest ecosystems. This new method does not require an assumption of equilibrium among P pools, and is safer and simpler in operation than isotopic techniques.