Varietal differences and year-to-year variations for seed yield and its associated characters in orchardgrass (Dactylis glomerata L.)

Orchardgrass, Dactylis glomerata L., is an important forage grass species in the northern part of Japan. The seed productivity of orchardgrass cultivars is of great importance from a seed multiplication standpoint, but only limited information is available on their seed yield potential in Japan. The objectives of this study were to examine 1) varietal differences in seed yield and its associated characters, 2) relationships between seed yield and yield components and 3)stability of seed yield and yield components in eight orchardgrass cultivars. Field experiments were conducted at the National Grassland Research Institute (36:55N, 139:58E) at Nishinasuno, Japan, during the years 1978-1984. The experiment was randomized complete block design with three replicates. Certified seeds of the eight cultivars were sown every year from 1978 to 1982 and seed yield and its associated characters were measured every second harvest year. The mean daily air temperatures for the period of June 11 to June 21 were lower in 1981 and 1983 than normal. The rainfall for this period in 1981 and 1983 was greater than normal, and accompanied by smaller amounts of solar radiation. There were significant differences among cultivars and years for seed yield, seed weight per panicle, and the number of panicle. The cultivar x year interaction for the number of panicle was significant at the 5 % level, but no significant cultivar x year interaction for seed yield and seed weight per panicle was found. The number of panicles was highest in Chinook and lowest in Kay. One thousand-seed weight was heaviest in Akimidori and 1,000-seed weights of early flowering cultivars were heavier than those of late flowering cultivars. The seed weights per panicle of Akimidori and Potomac were significantly heavier than those of Floreal, Frode, Tammisto and Kay. The seed yields of three early flowering cultivars, Chinook, Akimidori and Potomac, were significantly higher than those of Floreal, Frode, Tammisto and Kay. There were large differences in seed yield and seed weight per panicle among years. Seed yields and seed weights per panicle were high in 1980, 1982 and 1984, but low in 1981 and 1983. The coefficients of variation for seed yield were smallest in Chinook and Akimidori, and largest in Tammisto and Kay, indicating that Chinook and Akimidori are more stable in seed production in the central part of Japan. Averages of coefficients of variation of eight cultivars for seed yields and seed weights per panicle were higher than those for panicle numbers and 1,000-seed weights. One thousand-seed weight of eight cultivars was significantly correlated with seed yield in 1981 (gamma=0.838*], in good harvest years (gamma=0.596*] and in these four years (gamma=0.440]. The number of panicles was not significantly correlated with seed yield in eight cultivars of all production years and in three early flowering cultivars of good harvest years. It was significantly correlated with seed yield in eight cultivars of good (gamma=0.626*] and poor (gamma=0.728*] harvest years only. In contrast, seed weight per panicle was significantly, and positively, correlated with seed yield in eight cultivars of all five production years (gamma=0.905*] and three early flowering cultivars of good harvest years (gamma=0.850*], as well as eight cultivars of good (gamma=0.753*] and poor (gamma=0.944*] harvest years. The results obtained indicate that in orchardgrass there are large differences in seed yield and its associated characters among cultivars, and that seed weight per panicle is the most important determinant of seed yield. This suggests that strategies to improve orchardgrass seed yield should focus on increasing seed weight per panicle