Solutions containing 24 or 120 g ae/L of amine or salt formulations of glyphosate [N-(phosphonomethyl) glycine], triclopyr {[(3,5,6-trichloro-2-pyridinyl)oxy] acetic acid}, and picloram (4-amino-3,5,6-trichloropicolinic acid) were applied to small shrubs with a carpeted roller. Glyphosate reduced the live canopy cover of false broomweed (Ericameria austrotexana M.C. Johnston) by 83 to 87% at one site but only by 37 to 58% at another location. The lower concentration of picloram reduced false broomweed cover by at least 92%, and the higher concentration completely controlled the weed at both study sites. Effectiveness of triclopyr was usually intermediate between glyphosate and picloram. Removing a scraper designed to abrade stems prior to herbicide deposition reduced the effectiveness of picloram but not other herbicides. The herbicides reduced canopy cover of common goldenweed [Isocoma coronopifolia (Gray) Greene ♯3 IOCCO] by 92 to 100%. Common goldenweed was also controlled by 2,4-D [(2,4-dichlorophenoxy)acetic acid] amine applied with the carpeted roller. In glasshouse experiments, 1.2 to 60 g ae/L of 2,4-D amine reduced live canopy dry weight of common goldenweed by 73 to 88% at 5 months after application with a carpeted roller.