The kinetics of nonexchangeable -K release using Na TPB solution were studied on three paddy soils. Composite soil samples from 0-20 cm depth of three locations in Mazandaran province were taken, and their physico -chemical and mineralogical properties determined. Ca-saturated soil samples were equilibriated with Na TPB solution for 0.25-96 hr ,but , in one sample the time was extended up to 456 hr. The effect of shaking on the rate of K release and the recovery of K in Na TPB extraction method were also investigated .The K recovery in this method was almost 100% and its reproducibility was high. The rate of K release from soils were different, though their (dominant) clay mineralogy was the same. The rate of release in all samples increased rapidly initially but tended to level off with time. The quantity of K released initially with a high rate constituted a small fraction of total nonexchangeable-K, however, this “form” of K, despite its low quantity, plays an important role in K dynamics and fertility of soils. The effect of shaking on K-release rate was insignificant. The first-order and zero-order equations did not describe the K-release kinetics in soils, whereas, the parabolic diffusion equation described it more or less adequately, especially in soil No.2. HoweVer, all three equations described the data satisfactorily at initial period (up to 1.5 hr.) in all soils. Due to the conformity of kinetic data to the parabolic diffusion equation, it was concluded that the process of K release from these soils is essentially diffusion-controled. Comparison of correlation coefficients (r) and standard errors of the estimate (SE) indicated that the best model for describing the data in soils No. 1 and 3 was parabolic diffusion equation , and in soil No. 2 was first-order rate equation.
)