Buffelgrass (Pennisetum ciliare (L.) Link), a non-native perennial bunchgrass, invades ecologically intact areas of the Sonoran Desert. It competitively excludes native plants and increases fire frequency and intensity. Since the 1990s, whole buffelgrass plants have been manually uprooted and removed to control the invasion in southern Arizona. Uprooting plants results in bare, disturbed soil which promotes buffelgrass seed germination. This study examined whether leaving entire uprooted buffelgrass plants (thatch) on a field site reduces future buffelgrass establishment compared to removing uprooted plants from the site. A secondary goal was to determine whether light reduction and autoallelopathy were major factors in the negative effect of thatch on buffelgrass seedling density. Field plots with an average of 8,095 kg/ha thatch had 1.9 buffelgrass seedlings/m(2) which was significantly fewer than the 2.9 seedlings/m(2) in plots without thatch. Thatched portions of thatch plots (50% of their total area) had only 0.7 seedlings/m(2). In the greenhouse, which reduced outdoor light intensity by 35.2%, buffelgrass seeds sown in bare soil resulted in significantly higher seedling density than beneath buffelgrass thatch. Potential autoallelopathic chemicals leached from partially decomposed buffelgrass thatch and leached thatch had an intermediate but not significant (p = 0.09) effect on seedling numbers. Results suggest that leaving uprooted buffelgrass plants has the benefit of reducing seedling establishment in the area disturbed by uprooting.