Red clover (Trifolium pratense L) is an increasingly important forage legume for sustainable grassland systems, producing high dry matter yields of quality forage, however currently available varieties lack persistence particularly under grazing. The focus of the IBERS red clover breeding programme is on identifying the factors contributing to the poor persistence of red clover swards and applying this information to the selection of improved germplasm leading to the development of new varieties with greater persistence. Twelve red clover varieties and selection lines were grown in mixed swards with hybrid ryegrass or a mixture of hybrid ryegrass and perennial ryegrass over three harvest years. Three harvests were taken in each year with a Haldrup forage harvester at a cutting height of 5cm and total dry matter yields and dry matter yields of the red clover fraction quantified. Significant differences in DM yield of the red clover varieties was observed and generally yield in year 3 was lower than in year 1 however varieties differed in the extent of this decline yield in year 3 less of the new varieties greater than the control varieties Milvus and Merviot. The implication of these results for the future use of red clover in sustainable grassland systems is discussed.