Translocation of glyphosate (N-(phosphonomethyl)glycine) to plant roots and its impact on detected herbicide residues in sandy loam soil were studied in a glasshouse pot experiment in Finland. Quinoa (Chenopodium quinoa, Willd) plants in two different growing phases (6-8 and 12-14 leaf stages, groups A and B, respectively) were sprayed with nonlabelled glyphosate. Bare soil pots were included as controls (group C). Soil surface contamination with glyphosate was prevented in groups A and B but not in group C. Soil samples were collected 1 h, 8 days and 44 or 53 days after the glyphosate applications. Root samples were taken 8 days after the application from group B. After 8 days from the treatment, 4% of the applied glyphosate was detected in soil and about 12% in roots (group B). One and a half months later 12% and 8% of the applied glyphosate (groups A and B, respectively) was detected in soil samples incubated with roots. The main metabolite of glyphosate, aminomethyl phosphonic acid (AMPA), was not found in root samples. Glyphosate fate was simulated with the PEARL 3.0 model. Simulated concentrations in bare soil pots were very close to the observed ones. However, the model lacks a process description for herbicide transport within a plant and, therefore, the observed and simulated glyphosate residues in soil after canopy applications did not correlate. Simulations highlight the importance of the translocation process in glyphosate fate. We conclude that also in field studies part of the detected glyphosate soil residues must originate from plant roots, and translocation process should be included both in leaching assessments and pesticide fate models.