Improved methods are required to assess the risks posed by the uptake of potentially toxic elements such as selenium (Se), boron (B), and molybdenum (Mo) by vegetation on contaminated sites. In order to develop such methods and assess risk, vegetation was collected from two sites on a soil-capped coal fly ash landfill near Dunkirk, New York, during June of 1991 and June and August of 1992. The mean concentrations (μg g−1 dry weight) of Se and Mo in the shoots did not exceed, respectively, 0.12 and 18.7 in bird's-foot trefoil (Lotus corniculatus L.), 0.06 and 12.1 in red clover (Trifolium pratense L.), 0.07 and 5.3 in timothy (Phleum pratense L.), and 0.09 and 2.2 in a mixture of grasses. These concentrations were greater than those in the same species harvested concurrently from a non-landfill site. The mean concentrations of B at the landfill ranged from 29 to 53 μg g−1 in the legumes and from 2 to 11 μg g−1 in the grasses, less than those at one non-landfill site but greater than those at another. Within the landfill, the concentration of Se in grasses was not correlated with the concentration of Se in soil and fly ash. The concentration of Se in grasses on both landfill sites was double that of grasses on the non-landfill site despite higher mean concentrations of Se in the upper soil (0–15 cm) on the non-landfill site. Therefore grass roots seem to be accessing Se from the ash by means of mass flow or other mechanisms. Based on our findings of significant variation in trace element uptake among species, harvests, and locations within sites, we recommend that contemporaneous transect sampling of at least two species be used to assess uptake of potentially toxic trace elements on landfills or other sites where contamination may occur.