Due to the increasing concern about the appearance of glyphosate [N-(phosphonomethyl)glycine] and its major metabolite aminomethylphosphonic acid (AMPA) in natural waters, batch laboratory and lysimeter transport studies were performed to assess the potential for leaching of the compounds in two agricultural soils. Unlabeled and 14 C-labeled glyphosate were added at a rate corresponding to 1.54 kg a.i. ha −1 on undisturbed sand and clay columns. Leachate was sampled weekly during a period of 748 d for analyses of glyphosate, AMPA, total 14 C, and particle-bound residues. Topsoil and subsoil samples were used for determination of glyphosate adsorption, glyphosate degradation, and formation of AMPA and its degradation. Th e infl uence of adsorption on glyphosate degradation was confi rmed, giving very slow degradation rate in the clay soil (half-life 110-151 d). Th e kinetics of AMPA residues suggest that although AMPA is always more persistent than glyphosate when formed from glyphosate, its degradation rate can be faster than that of glyphosate. Th e kinetics also suggest that apart from glyphosate being transformed to AMPA, the sarcosine pathway can be just as signifi cant. Th e long persistence of glyphosate was also confi rmed in the lysimeter study, where glyphosate+AMPA residues constituted 59% of the initial amount of glyphosate added to the clay soil 748 d after application. Despite large amounts of precipitation in the autumn and winter after application, however, these residues were mainly located in the topsoil, and only 0.009 and 0.019% of the initial amount of glyphosate added leached during the whole study period in the sand and clay, respectively. No leaching of AMPA occurred in the sand, whereas 0.03 g ha −1 leached in the clay soil.