Core Ideas
Mixtures of legumes and grasses were dynamic in composition over time.
Forage mixtures with more legume had less biomass but had stable protein yield.
Alfalfa persisted better than clover, and was favored when grown with bunchgrasses.
When grown with meadow brome, alfalfa biomass in mixtures was 18 to 42% after 5 yr.
Species identity regulated forage productivity more than the proportion legume seeded.
To evaluate the comparative role of different legume–grass mixtures on long‐term forage yield, alfalfa (AL) (Medicago sativa L.) or clover (CL) (Trifolium sp.) were seeded at two sites in various combinations (0, 11, 22, 33, 67, or 100 kg 100 kg−1 of mix) with either rhizomatous grasses (smooth brome [Bromus inermis Leyss.] and Kentucky bluegrass [Poa pratensis L.]) or the bunchgrass meadow brome [Bromus riparius Rehmann]) to determine forage yield, quality, and legume persistence over 5 yr. Peak herbage yields were similar among swards initially seeded with 11 to 33 kg 100 kg−1 legumes, and occurred during the first 3 yr. Pure legume swards were slow to reach maximum biomass and increased crude protein yield (CPY) throughout the study. Regardless of companion grass identity, alfalfa‐based swards had more biomass and CPY than those containing clover. The proportion of alfalfa in mixed swards stabilized after 4 to 5 yr, and both legumes had greater abundance when grown with the bunchgrass rather than rhizomatous grasses. Although alfalfa was suppressed by rhizomatous grasses, their combined biomass was greater than in alfalfa–bunchgrass mixes, leading to similar CPY. Clover did not persist beyond 3 yr. The primary benefit of legumes in mixtures was an improvement in forage quality (i.e., CPY) rather than biomass. When establishing forage, species identity was critical to short‐term forage maximization, as was the presence of both legumes and grasses. In contrast, the relative proportion of legumes and grass at seeding had less influence on long‐term forage availability.