Soils from locations near Amsterdam, Dublin, Exeter (SW England), Göttingen, Paris and in the Grizedale Forest (NW England) were placed in lysimeters in a Norway spruce (Picea abies) stand in the Grizedale Forest and the soil solutions, rainfall and throughfall were monitored for 54 weeks. Two-way data tables for SO4, NH4, NO3, Al, Mg, Ca and pH were analysed by median polishing, in which each observed data value is decomposed into a typical value plus a row effect plus a column effect plus a residual. The row effects suggested that SO4 was the main acidifying influence in the Göttingen soil, but in the other soils SO4 had less effect than did NO3. The column effects showed little variation over the 54 weeks. Low values for NO3 in March-May and September were associated with periods of low or zero rainfall. The residuals contained from 27% (Al) to 59% (NO3) of the variation. Some were interpretable and contained useful information. Where the residuals are considered to represent uninterpretable random variation (‘noise’) they can be subtracted from the original data to give corrected data for other types of analysis. In median polishing, changing a few values in the input data, even drastically, will result in unusual residuals only in the cells with altered values. Hence, it is a useful technique for screening two-way data tables obtained from field experiments.