Search

The in parameters, variances, covariances and heritabilities, caused by the Bulmer effect have been predicted for an artificial breeding population of perennial ryegrass using recursion formulae. This enabled anticipation of the decrease in the expected response to multitrait index selection, using two different sets of economic weights. This theoretical decrease in selection efficiency was found to be quite substantial; it is approximately halved after 4 cycles of individual recurrent selection with a fairly high selection intensity of 5 %. However the synthetic population is improved for 4 out of 5 agronomic traits and could theoretically reach quite a high level of performance after 5 cycles. The choice of economic weights has relatively little influence on the expected genetic progress of each trait. [ABSTRACT FROM AUTHOR]

To predict any genetic response to selection, two breeding populations of perennial ryegrass, consisting of half-sib families, were evaluated as spaced plants from 1987 to 1989 for several agronomical traits. In each population, the best plants selected through a multitrait Smith and Hazel's index were planted in a polycross to produce the second generation half-sib families. Both generations were evaluated together from 1991 to 1993 as spaced plants. The difference between generations was the observed genetic response, which was compared to that predicted. Rust tolerance, aftermath heading and persistence were significantly improved in both populations. In trial plots used to assess the yield, no effect of mass selection was noticed. Thus it is concluded that mass selection failed to increase productivity but produced some improvements in disease tolerance and aftermath heading. [ABSTRACT FROM AUTHOR]