man (1983) and Østrem (1988) studied variation in seed yield components. Bonin and Goplen (1966) reported
Reed canarygrass (Phalaris arundinacea L.) (RCG) is usuallyheritability of seed yield components, and Jonassen grown for forage, but it is also a potential crop for bioenergy and paper pulp production since under long-day conditions it produces (1994) detailed effects of various treatments on seed substantial biomass. Problems related to seed production, including yield. Baltensperger and Kalton (1958;1959) and Bonin low seed yield and early seed shattering, should be addressed before and Goplen (1963) reported variability in seed shatits cultivation can be increased at high latitudes. The objectives of tering, Landgraff and Junttila (1979), Junttila and Nilsen this study were to evaluate the extent and nature of variation in seed (1980), and Berg (1992) studied variation in seed germiproduction traits and factors contributing to high seed yield. Seed nation and seed dormancy, and Sahramaa and Hö mmö yield, seed shattering, seed weight, germination, panicle weight, and (2000) recorded variation in seed production in Finland.
panicle length were studied in Finland in 1995, and panicle numberResults of earlier studies showed that there is variabetween 1995 and 1997. The research material included eight cultivars, tion in RCG for seed production traits. Since RCG is 14 breeding lines, and 53 Finnish wild populations, which were divided an outbreeder, it is highly heterozygous and variable into 10 groups according to their origin. Results indicated that populations of reed canarygrass differed markedly in seed production traits. for many traits. Finland is a potential area for forage Cultivars, breeding lines, and the southernmost wild groups from the grass seed production (Niemelä inen and Niskanen, 2002). coastal area had the best seed production traits. The northernmost RCG is native to Finland and geographical differentiawild group was superior only in resistance to seed shattering. The tion in seed production traits was assumed because of best time for harvest was later for cultivars than for breeding lines the substantial distances separating populations. Harand wild populations. Seed production traits were correlated, which vest time has been shown to be difficult to determine would allow simultaneous improvement of those traits. The variation because stands mature unevenly, both among and within in traits recorded in this study indicated that there is potential to panicles (Evans and Ely, 1941). Seeds also shatter soon breed new reed canarygrass cultivars with good seed production from after maturity and hence seed yield is often low with local populations from particular geographic areas and breeding mauneven germination. The objectives of this study were terials.to evaluate the extent and nature of variation in seed production traits and factors contributing to high seed yield. Panicles were manually shed three times after each collec-