Desodification of a highly saline-sodic, permeable, sandy loam soil was accomplished in 2x2 m (S,), 4x4 m (S s ) and 6x6 m (S 3 ) plots under continuous and intermittent ponding. The soil contained large amounts of soluble salts and exchangeable sodium throughout the profile to the water table. The chief salts were chlorides and sulphates of sodium, calcium, and magnesium. The leaching curves with respect to desodification did not differ significantly between the S,, S 2 and S 3 plots under continuous ponding but did under intermittent ponding. With intermittent ponding, the efficiency of leaching exchangeable sodium in the S, plots was very significantly higher than that in the S 2 and S 3 plots, i.e. the leaching efficiency decreased sharply with increased plot size. The leaching efficiency in the S, plots was significantly greater with intermittent than with continuous ponding, but the reverse was true in the S 2 and S 3 plots. The results further indicated that there is no need to apply an amendment such as gypsum, mainly because calcium and magnesium present in such soils are adequate to replace the initially high exchangeable sodium during leaching. The displacement of exchangeable sodium in the S x plots under intermittent ponding was nearly piston-like. With increased plot size, it tended to deviate from this behaviour. It is concluded that under practical situations in which reclamation is accomplished in large plots, the practice of leaching the saline constituents with intermittent ponding would be less effective. Thus, leaching of salts down into the soil with a single heavy irrigation would be more desirable.