To study the indirect effects, seed increase nurseries for each population (RCB, r = 4) were established during 2007/08 and 2008/09 with spaced planting. Individual plant observations for the above mentioned traits were taken during reproductive growth in the spring. For all correlated responses, C1 has shown significant change from C0 with exception of seed yield in 2008/09, while all responses remained constant for other selection cycles (C2, C3). With three cycles of selection, selected populations shifted their heading date 8-d earlier with more erect growth habit and increased seed yield than the C0. Out of these three correlated responses, seed yield has shown population x evaluation year interaction.
For second objective seed was harvested from 24 random plants per isolation block from seed increase nursery (2007/08) for a total of 96 entries per population, where each selected plant represented a half-sib family. Two blocks per cycle of short 1-m rows on 30-cm row spacing were seeded in late October 2008 and harvested 750 GDD later in mid-January. Another two blocks were seeded to evaluate heading date. ANOVA among HS-families confirmed the results from the solid-seeded sward studies that indicated progress from selection as well as the presence of genetic variation. Heritability estimates for the first cut of C0 = 0.54, C1 = 0.43, and C2 = 0.45 indicated that (1) narrow sense heritability was high enough to make progress, and (2) all three populations had similar heritability. Heritability for heading date ranged from 0.45 to 0.81, depending on the cycle; C0 had the highest heritability. We conclude that the a priori assumption of sufficient genetic variation for winter productivity was fully warranted.