The Mehlich‐III method used for the determination of available P in 82 Quebec soils was compared to five other chemical methods (Bray‐II, Bray‐I, Mehlich‐II, Mehlich‐I and Olsen) and to two anion exchange resin techniques (F‐and HCO‐3) In general, very good correlations were found between Mehlich‐III, Bray‐II and Bray‐I methods on acid soils (r=0.95** and 0.98**). The Mehlich‐III‐P content is somewhat the same as that determined by Bray‐I and corresponds approximately to 80% of Bray‐II‐P. The Mehlich—III extractant was found to be more reliable than Bray‐II and Bray‐I on very acid spodosols having high P‐sorption capacity. The Bray methods, with high NH4F concentration, extracted strongly fixed Al‐P in these soils and may overestimate the available P. In some acid soils containing apatite and in calcareous soils, the Bray‐II and Mehlich‐I (double acid) easily dissolved these Ca‐P compounds and gave exaggerated high P values. The Mehlich‐III and the other methods were less vigourous for these compounds and were more accurate in estimating available P in these soils. On moderately calcarous soils, the Bray‐I and Mehlich‐I are partially neutralized by the free carbonates in the soil and become less effective in extracting soil P. The Mehlich‐III reactant is more buffered and less affected by carbonates in these soils. The Mehlich‐III and the Mehlich‐II were highly correlated for all soils having diverse properties (r=0.98**) and the amount of P extracted by Mehlich‐III was 117% of that extracted by Mehlich‐II.