CONVERGENT-DIVERGENT SELECTION FOR SEED PRODUCTION AND FORAGE TRAITS IN ORCHARDGRASS .1. DIRECT SELECTION RESPONSES

Simultaneous improvement of forage traits and seed yield in orchardgra ss (Dactylis glomerata L.) has been problematic because of geographic separation of forage and seed production locations. Most attempts to s imultaneously improve both Forage and seed traits have relied on a sin gle location. The objective of this study was to conduct and evaluate two cycles of convergent-divergent (CID) and local selection for panic le seed weight (PSW) and forage traits. Selection was conducted at fou r locations (Ames, IA; Mount Vernon, MO; Rock Springs, PA; and Arlingt on, WI) in four base populations (I79DT, MO2, PLS4, and WO11). Phenoty pic selection on a spaced-plant basis was applied at a 0.25 selection intensity for forage traits, followed by a 0.25 selection intensity fo r PSW. All populations were evaluated as spaced plants at the Iowa and Pennsylvania locations in 1990 and 1991. Both selection methods were equally effective at increasing PSW, with responses averaging 1.0 to 1 8.2% cycle(-1). Realized heritabilities for PSW were highly variable a mong the four populations (mean of 0.03-0.23) and were lower for Cycle 2 than for Cycle 1 (mean of 0.02 vs. 0.16). For C/D selection, IA, MO , PA, and WI selection locations increased PSW by 25, 39, 94, and 33 m g panicle(-1) respectively. Three of the four populations had signific ant genetic changes of 2 to 4 d later heading, despite no direct selec tion pressure for heading date. There were very few significant and bi ologically meaningful genotype x location interactions for any trait. Therefore, neither selection method resulted in plants with differenti al adaptation when evaluated as spaced plants.