The widely used herbicide, glyphosate (N-phosphonomethylglycine), which is the active ingredient in several commercial products, e.g. RoundUpTM, seems to be sorbed onto variable-charge surfaces by ligand exchange through the phosphonate group in a way similar to sorption of phosphate (Dideriksen & Stipp 2003; Sheals et al., 2002). The apparent similarity in sorption mechanisms indicates that glyphosate and phosphate may be sorbed onto the same sorption sites and may compete for these sites. Thus, glyphosate and phosphate were strongly sorbed by synthetic iron and aluminium oxides but with suppressed glyphosate sorption in presence of phosphate showing preference of the oxides for phosphate (Gimsing & Borggaard, 2001; Gimsing & Borggaard, 2002). In contrast, sorption of glyphosate and phosphate by pure permanent-charge 2:1-layer silicates was modest and sorption affinities were similar (Gimsing & Borggaard, 2002). Accordingly, Danish soils dominated by permanent-charge clay silicates were found to sorb limited amounts of glyphosate and phosphate, and sorption of the two sorbates appeared additive rather than competitive inasmuch as higher amounts were sorbed when they were added together than when glyphosate or phosphate was added alone (Gimsing et al., 2004). In order to evaluate sorption behaviour of glyphosate and phosphate by variable-charge soils, sorption of glyphosate and phosphate by an Andisol, two Oxisols and an Ultisol from Tanzania was investigated; glyphosate and phosphate were added both separately and in combination. Sorption of both sorbates was found to decrease in the same order as oxalate-extractable Al and Fe, i.e. Andisols > Oxisols > Ultisol. Competition between glyphosate and phosphate was strong with the Andisol and one Oxisol, where little glyphosate was sorbed on phosphate-saturated soil. However, the two other soils showed limited preference for the two sorbates and the sum of glyphosate plus phosphate when present together was considerable higher than when glyphosate and phosphate occurred separately. These results indicate that soils in relation to glyphosate and phosphate sorption possess common as well as more specific sorption sites. Common sites accept both glyphosate and phosphate but with preference for phosphate, while other sites seem available for either glyphosate or phosphate. Depending on their composition, individual soils can possess different numbers and ratios of common and specific sites, and hence exhibit different behaviour toward glyphosate and phosphate alone and in combination. The sorption behavior will be discussed in relation to soil composition with emphasis on clay mineralogy. Knowledge about glyphosate sorption by various soils and the influence of phosphate on this process become more and more urgent because of increasing use of glyphosate in both conventional and gen-modified crop production. Increasing use of glyphosate augments the risk of pollution of the aquatic environment with this xenobiotic. References: (1)Dideriksen, K. & Stipp, S.L.S. (2003). Geochimica et Cosmochimica Acta 67: 3313-3327. (2) Gimsing, A.L. & Borggaard, O.K. (2001). Clays & Clay Minerals 49: 270-275. (3)Gimsing, A.L. & Borggaard, O.K. (2002). Clay Minerals 37: 509-515. Gimsing, A.L., Borggaard, O.K. & Bang, M. (2004). European Journal of Soil Science 55: 183-191. (4) Sheals, J., Sjöberg, S. & Persson, P. (2002). Environmental Science & Technology 36: 3090-3095.