Tall fescue has a high degree of self incompatibility. Because of this, every plant in a tall fescue population represents a unique genotype, and breeding procedures developed for cross-pollinated species are most applicable for tall fescue cultivar development. Tall fescue cultivars released to date are populations that consist of a wide array of genotypes, usually with some level of uniformity for morphological and agronomic characteristics. As a result, breeding procedures have focused on increasing the frequency of favorable alleles in a population.
Ecotype selection was one of the earliest methods used to develop grass cultivars, including tall fescue. This technique involves collecting tall fescue germplasm from a source (e.g., an old pasture, roadside, seed field, etc.), testing that germplasm in a number of environments, and finally directly increasing seed of the germplasm for cultivar release. Cultivars such as KY-31 were developed in this manner. Ecotype breeding continues to play an important role in tall fescue breeding, as breeders are able to capitalize on natural selection that, together with the high level of genetic variation present in tall fescue, can act to generate new populations in short periods of time (Valay and van Santen, 1999).
Recurrent selection involves selecting plants for a trait or traits, intermating those plants, and repeating this process for multiple generations (referred to as cycles of selection). Phenotypic selection is similar except that selection is completed after one cycle. In both procedures, selected plants are intermated in isolation, excluding pollen from other genotypes by distance or by removing undesirable plants from the nursery before pollination. Subsequently seed is harvested from selected plants. Selecting both male and female parents in this manner results in twice the genetic gain per cycle obtained when using mass selection. Recurrent selection can be a powerful means of accumulating favorable alleles in a population and often is the method of choice for improving traits with low heritability. Selecting parents based on progeny performance increases the probability of identifying superior genotypes, but it can also increase the amount of time per cycle.