Premium
Heterosis and Combining Ability in Hybrids Derived from Oilseed Rape Cultivars and Inbred Lines
Author(s) -
Brandle J. E.,
McVetty P. B. E.
Publication year - 1989
Publication title -
crop science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.76
H-Index - 147
eISSN - 1435-0653
pISSN - 0011-183X
DOI - 10.2135/cropsci1989.0011183x002900050020x
Subject(s) - hybrid , inbred strain , biology , cultivar , heterosis , mating design , agronomy , horticulture , genetics , gene
Basic genetic information needed to develop hybrid oilseed rape ( Brassica napus spp. oleifera ) cultivars has not been previously reported. Since recent work had shown oilseed rape cultivars to be genetically heterogeneous, inbred line derived hybrids were compared to cultivar derived hybrids to determine if there were differences among inbred lines in terms of their performance in hybrid combination. For inbred line derived hybrids, midparent was compared to general combining ability as a means of predicting hybrid yield. Three inbred lines, from each of three source cultivars, were crossed using a factorial mating design and the resultant hybrids were field evaluated in three environments. The inbred lines, source cultivars, and cultivar hybrids were included as checks. Yields of inbred line derived hybrids were compared to their respective cultivar derived hybrids and some inbred line crosses were found to be significantly ( P < 0.05) higher and others significantly ( P < 0.05) lower yielding. There were also differences in general combining ability (GCA) effects among inbred lines. This result indicates the presence of variability in breeding values among cultivar derived inbred lines and that hybrid oilseed rape breeding programs should be based on inbred line crosses, rather than cultivar crosses. General combining ability (GCA) was significant ( P < 0.01) and accounted for 88% of the cross sums of squares, while specific combining ability (SCA) was nonsignificant ( P > 0.05) indicating that additive genetic effects predominantly influence the expression of yield. Regression of midparent on observed hybrid yield was significant ( P < 0.01), with an r 2 of 0.37. Regression of expected hybrid yield, based on parental GCA, on observed hybrid yield was also significant ( P < 0.01), but with an r 2 of 0.88. Therefore, GCA was more effective than midparent for describing hybrid yield. There was, however, a positive relationship between midparent and hybrid yield indicating that despite the somewhat low r 2 , midparent may still be a useful guide for the selection of parents for testing in hybrid combination.