Phalaris arundinacea (reed canarygrass) is recognized as a problematic invader of North American marshes, decreasing biodiversity and persisting in the face of control efforts. Less is known about its ecology or management in forested wetlands, providing an opportunity to apply information about factors critical to an invader's control in one wetland type to another. In a potted plant experiment and in the field, we documented strong competitive effects of reed canarygrass on the establishment and early growth of tree seedlings. In the field, we demonstrated the effectiveness of a novel restoration strategy, combining site scarification with late fall applications of pre-emergent herbicides. Treatments delayed reed canarygrass emergence the following spring, creating a window of opportunity for the early growth of native plants in the absence of competition from the grass. They also allowed for follow-up herbicide treatments during the growing season. We documented greater establishment of wetland herbs and tree seedlings in treated areas. Data from small exclosures suggest, however, that deer browsing can limit tree seedling height growth in floodplain restorations. Slower tree growth will delay canopy closure, potentially allowing reed canarygrass re-invasion. Thus, it may be necessary to protect tree seedlings from herbivory to assure forest regeneration.