Seven dinitroaniline herbicides were tested in the field for 3 yr at rates up to twice that normally used for annual grass control. Trifluralin (α,α,α-trifluoro-2,6-dinitro-N,N-dipropyl-p-toluidine) and profluralin [N-(cyclopropylmethyl)-α,α,α-trifluoro-2,6-dinitro-N-propyl-p-toluidine] gave 95 and 94% control of johnsongrass [Sorghum halepense (L.) Pers.] from rhizomes, respectively, after the third year at the double rate. These two herbicides controlled johnsongrass from rhizomes better than nitralin [4-methylsulfonyl)-2,6-dinitro-N,N-dipropylaniline], butralin [4-(1,1-dimethylethyl)-N-(1-methylpropyl)-2,6-dinitrobenzenamine] and pendimethalin [N-(1-ethylpropyl)-3,4-dimethyl-2,6-dinitrobenzenamine] at comparable rates. After 3 yr, seed cotton (Gossypium hirsutum L. ‘Stoneville 213’) yields were higher with double rates of trifluralin, profluralin, fluchloralin [N-(2-chloroethyl)-2,6-dinitro-N-propyl-4-(trifluoromethyl)aniline], and pendimethalin, than with nitralin, dinitramine (N4,N4-diethyl-α,α,α-trifluoro-3,5-dinitrotoluene-2,4-diamine), and butralin. In greenhouse studies, reduction of cotton raproot length by the dinitroaniline herbicides incorporated at the normal rate of use followed the order: butralin ≤ pendimethalin ≤ fluchloralin ≤ trifluralin < profluralin < dinitramine. All herbicides except pendimethalin at 0.28 and 0.56 kg/ha reduced lateral root production of cotton. The dinitroaniline herbicides at normal rates of use reduced fresh weight of cotton seedlings from 12 to 34%. The mean vapor toxicity of the dinitroaniline herbicides to Japanese millet [Echinochloa crus-galli (Roxb.) Wight ‘frumentacea’] shoots averaged from no injury with oryzalin (3,5-dinitro-N4,N4-dipropylsulfanilamide) and nitralin to 66% injury with dinitramine and trifluralin.