As part of a study of the phytotoxic risk of spreading of contaminated sediments "on soil", we carried out a laboratory experiment assessing the impact of water draining from sediments in a deposit scenario on the peripheral vegetation. The plant tested were the Chinese cabbage (Brassica campestris L. var. chinensis), maize (Zea mays L.) and ryegrass (Lolium perenne L.). The draining water samples (A1, B1 and C1) were obtained after decantation in laboratory of the sediments. The classification of the water sampled in decreasing order of cumulative contamination was C1 > A1 > B1. The B1 and C1 water samples inhibited the germination of seeds tested to various extents. The percentage of seeds that did not germinate was 1.3-fold times higher for Chinese cabbage with B1 than for the control and 2.3-fold times higher for ryegrass with C1 than for the control. Seeds watered with B1 had lower moisture contents than the control: 10% lower for ryegrass and maize and 50% lower for Chinese cabbage. An increase (about 1.5 microg/ml) in total soluble protein (versus the control) was observed for all three plant species tested in the presence of C1. Glutamine synthetase activity was significantly (1.35 times) higher in ryegrass seeds in the presence of C1 than in the control. We also observed changes in the specific activity of phosphoenolpyruvate carboxylase, which increased for ryegrass and decreased for maize as the concentration of contaminants in the water increased. The results show (i) the necessity to use different plant species to evaluate the toxic effect of sediment deposited on soil on the vegetation, and (ii) that soon as on germination an evaluation of an impact is possible.