Field studies were conducted in 2001 in Lewiston, NC, and in 2002 at Clayton and Lewiston, NC, to investigate the response of nontransgenic cotton to simulated glyphosate drift in a weed-free environment. Nontransgenic cotton variety ‘Fibermax 989’ was planted in a conventional seedbed at all locations. Glyphosate treatments were applied early postemergence (EPOST) at the four-leaf growth stage of cotton at 0, 8.7, 17.5, 35, 70, 140, 280, 560, and 1,120 g ai/ha and represent 0, 0.78, 1.55, 3.13, 6.25, 12.5, 25, 50, and 100% of the commercial use rate, respectively. Rates as low as 140 g/ha caused lint yield reductions depending on year and location. When averaged over all locations, lint yield reductions of 4, 49, 72, and 87% compared with nontreated cotton were observed with glyphosate rates of 140, 280, 560, and 1,120 g/ha, respectively. Visual injury and shikimic acid accumulation were evident at glyphosate rates greater or equal to 70 g/ha. Collectively, visual injury and shikimic acid accumulation at 7 d after EPOST treatment might be used as a diagnostic indicator to predict potential yield reductions from simulated glyphosate drift.