We studied the effects of hand weeding of second-year plants of the biennial garlic mustard (Alliaria petiolata) on first-year plants (seedlings) and native ground layer vegetation. Garlic mustard is a Eurasian species that has invaded deciduous forest ground layers in eastern North America. Treatments consisted of a control and an early or late weeding of second-year garlic mustard. The early treatment (early March) was applied before garlic mustard seeds had germinated and when most native species were dormant. The late treatment (mid-May) occurred after plants had bolted, flowering was occurring, and most native species and new garlic mustard seedlings were actively growing. Pre-treatment data were obtained in 2004 and treated and control plots were sampled in 2005, 2006, and 2007. No significant treatment effects were observed in 2004 or 2005. In 2006, mean cover of first-year plants was higher in the early weeding treatment than in the late weeding treatment and control. In 2007, mean cover of first-year garlic mustard was higher in the control than in either of the two weeding treatments. There were no significant treatment effects in any year on native vegetation cover, bare ground, or the five most abundant native species. Our data indicate that (1) late weeding of garlic mustard provided more effective control than early weeding because late weeding allows second-year plants to compete with garlic mustard seedlings for a longer period of time and (2) competition between first- and second-year plants is responsible for alternating dominance of first-year and second-year garlic mustard plants.