Seed banks may contribute useful or weedy species that fill gaps in pastures. In a previous study, pastures planted to complex mixtures of forages had a lesser proportion of weedy species in the aboveground vegetation. In this study, we relate changes in the species composition of the seed bank to changes in the aboveground vegetation. In August 2001, four mixtures [two, three, six, and nine species of temperategrasses, legumes and chicory (Cichorium intybus L.)] were established in replicated 1-ha pastures (eight total) in central Pennsylvania. Pastures were grazed by dairy cattle from April to September in 2002 and 2003. Soil cores (1.88-cm diam. by 5-cm depth) were taken in April and October in 2002 and 2003, and in April 2004 to determine the density of germinable seeds. Soil samples were placed in a greenhouse under natural light and controlled temperatures for 12 to 18 mo and germinated seedlings counted regularly. The total density of germinable seeds from all species did not differ among mixtures (P = 0.08). Annual nonleguminous forbs accounted for 79% of the germinable seeds. Yellow woodsorrel (Oxalis stricta L.) was the dominant annual forb. There were significant differences among pastures planted to different mixtures in the density of germinable annual forb seeds; however, these differences likely occurred because of preexisting spatial variation in seed bank composition. Seeded species contributed fewer than 1000 seeds m–2 total in the 3 yr. Kentucky bluegrass (Poa pratensis L.) and white clover (Trifolium repens L.) were the most common forage species in the seed bank. There was little relation between the species composition of the seed bank and the composition of the aboveground vegetation. Data from this study indicate that previous land management had larger effects on the soil seed bank than did planting diverse mixtures of forages.