Germinated seeds of wild oat populations that were susceptible (S) or resistant (R) to triallate at the recommended soil-applied rate (1.7 kg/ha) were treated with six triallate concentrations on filter paper in petri dishes. Measurement of shoot length 8 d after treatment provided an accurate indication of differences among populations, and was more reliable than determining shoot fresh weight. ED50values (herbicide concentrations that reduced shoot length by 50% relative to untreated controls), derived from nonlinear regression analysis, indicated four and five levels of response to triallate among eight S and seven R populations, respectively. The ED50values varied from 0.11 to 11 ppm a.i. triallate for the most susceptible to the most resistant populations, respectively. Routine testing of wild oat samples suspected of resistance, at triallate concentrations of 0.5 or 1 ppm in the petri dish bioassay, effectively identified populations that had become resistant to the recommended soil-applied rate.