HERITABILITIES, GENETIC-PARAMETERS, AND RESPONSE TO SELECTION IN PEARL-MILLET X ELEPHANTGRASS HEXAPLOID HYBRIDS

Authors
Citation
Da. Diz et Sc. Schank, HERITABILITIES, GENETIC-PARAMETERS, AND RESPONSE TO SELECTION IN PEARL-MILLET X ELEPHANTGRASS HEXAPLOID HYBRIDS, Crop science, 35(1), 1995, pp. 95-101
Citations number
16
Categorie Soggetti
Agriculture
Journal title
ISSN journal
0011183X
Volume
35
Issue
1
Year of publication
1995
Pages
95 - 101
Database
ISI
SICI code
0011-183X(1995)35:1<95:HGARTS>2.0.ZU;2-A
Abstract
Seeded pearl millet [Pennisetum glaucum (L.) R.Br.] x elephantgrass (P ennisetum purpureum Schum.) allohexaploids possess many desirable attr ibutes for forage and biomass production. Estimation of genetic parame ters would be useful in developing appropriate selection strategies. T his study was conducted to obtain single-plant heritability estimates; genetic, phenotypic, and environmental correlations; and expected res ponses to selection for several vegetative and reproductive characteri stics. Data were collected in 1990 and 1992 from a factorial experimen t containing seven S-1 (selfed) families which were grown on a Sparr f ine sand (sandy siliceous, hyperthermic Grossarenic Paleudult). Fewer plants were evaluated in 1992 because of lack of persistence. Due to s ignificant year effects, and because 1992 data was affected by plant l oss, years were analyzed separately. Heritability estimates differed a mong the 11 traits and were very low to moderate. Leaf width had the h ighest heritability for vegetative traits (0.16 in 1992). Heritabiliti es for reproductive characters were moderate in 1990, being highest fo r 100-seed weight (0.30) and seeds per panicle (0.29). Zn 1992, lack o f significance among families for most reproductive characters reduced heritability estimates. Some correlations were consistent over years. Number of tillers was negatively associated genetically with plant he ight and leaf characteristics. Days to flowering was negatively associ ated with vegetative and reproductive traits. Seed-related traits were highly correlated among themselves. Despite moderate to low heritabil ities for seed characters in 1990, expected responses to selection wer e high (17-43% increase) due to large phenotypic variation. For improv ing seed yield per plant, selecting indirectly for days to flowering o r other seed-related traits would be more efficient. Improvement throu gh phenotypic selection should be satisfactory in several traits.