Development of compatible, persistent, warm-season grass-legume mixtures could increase forage yield and quality during summer months. We established a trial to determine forage yield, quality, species compatibility, and persistence of binary mixtures of warm-season grasses with selected legumes, five of which are native to the central USA. Grass entries were switchgrass (Panicum virgatum L.), sideoats grama (Bouteloua curtipendula Michx.), and indiangrass [Sorghastrum nutans (L.) Nash]. Legume entries were purple prairieclover [Dalea purpurea Vent.; syn. Petalostemon purpureum (Vent.) Rydb.], roundhead lespedeza (Lespedeza capitata Michx.), leadplant (Amorpha canescens Pursh), Illinois bundleflower [Desmanthus illinoensis (Michx.) MacMill., B. Robins. & Fern.], catclaw sensitive brier [Schrankia nuttallii (DC.) Standl.], and cicer milkveteh (Astragalus cicer L.), a cool-season species. Unfertilized grass plots without legumes also were included. The experiment was on a Haynie very fine sandy loam soil (coarse-silty, mixed, calcareous, mesic Typic Udifluvents). All mixtures containing purple prairieclover, roundhead lespedeza, Illinois bundleflower, or catclaw sensitive brier yielded more forage than did grasses grown alone or with leadplant, except for pure switchgrass in 1986. All legumes increased the crude protein concentration of forage compared to that of grass-alone plots, except for leadplant with switchgrass in 1986. Inclusion of catclaw sensitive brier and cicer milkvetch with grasses consistently improved in vitro digestible dry matter concentration (IVDDM), while inclusion of roundhead lespedeza, leadplant, and Illinois bundleflower generally resulted in decreased [VDDM concentration of forage. Purple prairieclover generally did not influence IVDMD of mixtures. Persistence of all legumes was good. Cicer milkvetch was not compatible with these grasses because it developed a thick, dense canopy prior to initiation of growth by these grasses.