* A phylogenetic perspective of community assembly can reveal new insights into how variation within dominant species interacts with the local species pool to influence the structure of restored plant communities. Many studies have examined the effect of dominant species in structuring plant communities, but few have investigated their effect on phylogenetic diversity (PD). * We established grassland in a post-agricultural field using two population sources (cultivars and local ecotypes) of three dominant grasses (Sorghastrum nutans, Andropogon gerardii and Schizachyrium scoparium) with three unique pools of subordinate species that varied in PD but not taxonomic or life-form diversity. We tested the effect of the population source treatment on two metrics of community PD (net relatedness index [NRI] and nearest taxon index [NTI]) during the first 4 years of restoration. The NRI measures the overall pairwise phylogenetic distance between all pairs of taxa in a community. By contrast, NTI measures the pairwise distance between closely related taxa in a community. * Population sources had a transitory effect on community phylogenetic structure over time. Local ecotypes decreased the abundance of closely related eudicots, monocots (low +NRI and +NTI values) and volunteer species (−NTI) more than cultivars. However, population sources did not affect ecologically conservative species (i.e. species with intermediate-to-poor ecological tolerance and a high degree of fidelity to prairie habitats). Thus, cultivars might have a positive effect on community phylogenetic diversity more than local ecotypes by decreasing the abundance of a phylogenetically diverse community of less closely related volunteer species. Differences in PD of seed mixes were maintained in the community of high-fidelity species, but did not affect PD of the unsown (volunteer) species in the assembling community. * Synthesis and applications. This is the first experiment to show consequences of using different seed sources on phylogenetic diversity (PD) in grassland restoration. Phylogenetics can reveal the effects of population sources on the abundance of volunteer species not evident through traditional analyses of species diversity. The PD of seed mixes or establishing communities, or other assessments of phylogenetic relationships, by restoration practitioners is recommended as a metric to allow consequences of the evolutionary patterns among species to be included in conservation planning. Increased accessibility of phylogenetic tools will allow the application of PD in restoration monitoring.