Genetic variation within eight populations of perennial forage grasses

Information on both mean performance and genetic variation is important in determining the value of germplasm for breeding via recurrent selection. The objectives of this experiment were to evaluate different methods of controlling spatial variation for forage yield, and to estimate genetic variances and heritability in eight perennial forage grass populations. Polycross families of two populations of smooth bromegrass (Bromus inermis Leyss.), two populations of hybrid wheatgrass (Elytrigia x muctonata (Opiz ex. Bercht.) Prokud.), and four populations of orchardgrass (Dactylis glomerata L.) were field tested for 2 years. Spatial variation was present in all three progeny tests, but was well accounted for by either the incomplete block design or a nearest neighbour analysis (NNA) of yield potentials. Low rank correlations between adjusted family means for the incomplete block design vs, the NNA created the unanswered question of which would provide a more appropriate estimate of family mean performance. Genetic variation was detected in six out of eight populations for forage yield, four out of six populations for ground cover, all four orchardgrass populations for maturity and one of four orchardgrass populations for leafspot reaction (caused by Drechslera spp.). Excessive year-family interactions prevented detection of genetic variation for leafspot reaction in the other populations. Heritability estimates ranged from 0.30 to 0.42 for forage yield. In all three species, there was a trend toward a negative relationship between population mean and genetic variance, suggesting that less elite populations may need to contribute other traits to be useful in the short term.