Plant-soil feedbacks can affect plant community dynamics by influencing processes of coexistence or invasion, or by maintaining alternate stable states. Darwin's naturalization hypothesis suggests that phylogenetic relatedness should be a critical factor governing such feedbacks in invaded communities but is rarely considered in soil feedback studies. We investigated the effects of soil biota from experimentally established native and invaded California grassland communities on resource capture and allocation of three native and three exotic grass species, comprising three tribes, grown in the laboratory. Phylogeny was the single greatest determinant of grass biomass, root:shoot ratio, and growth rate, with presence of soil biota explaining the second greatest proportion of variance in total grass biomass. Similar trends were observed in soil collected from naturally occurring stands of native perennial and exotic annual grasses. Species of similar life history/provenance exhibited similar biomass responses to the same soil community, while more closely related species exhibited similar root:shoot ratio responses to the same soil community. Relationships between the plant community composition of a field plot and species responses to soil inoculum collected from that field plot were idiosyncratic, with many aspects of plant community structure potentially contributing to soil feedbacks. Thus, future studies should explicitly consider both phylogeny and provenance and evaluate soil feedbacks in a community setting.