The influence of five clay concentrations as contaminants into the herbicide solution was determined for glyphosate from pot experiments. Yellow nutsedge plants at IS± 2 leaf stage were sprayed with two rates of glyphosate (1.44 and 2.88 kg ai·ha·1) in a volume equivalent to 200 L·ha·1. Each rate was applied with increasing amount of soil into the herbicide solution, to get concentrations of 0; 3,270; 6,540; 9,800 and 13,080 ppm. Each treatment was replicated five times. An absolute check replicated five times was used. Leaf damage, fresh weight of aerial portion and tuber number per pot due to glyphosate were evaluated 48 DAA. Results showed a decreased glyphosate activity when applied at the lower rate, and a soil concentration greater than 6,540 ppm independently of the parameter evaluated. On the other hand, the presence of soil in the herbicide solution, regardless of its concentration, did not affect activity of high rate glyphosate treatments.