EVALUATION OF RAPID CYCLE RECURRENT SELECTION FOR EARLY FLOWERING IN OAT

More efficient methods of recurrent selection are important to oat (Av ena sativa L.) breeding. Although recurrent selection methods have bee n utilized successfully for the improvement of a number of traits in o at, rapid-cycle population improvement methods for moderately to highl y heritable traits have not been evaluated. Our objectives were (i) to evaluate effectiveness of rapid-cycle recurrent selection for early f lowering, and (ii) to determine the effect of selection for early flow ering on grain yield, maturity, and plant height. Phenotypic recurrent selection for early flowering was practiced for four cycles. The sche me requires one generation per cycle and up to three cycles can be com pleted in 1 yr. From each of the 30 selected plants in all four cycles , a single F-2:4 line was derived for replicated testing. All derived lines and the original parents were evaluated at Rosemount and St. Pau l, MN, in 1992 and 1993. Analysis of F-2:4 lines from the earliest 12 plants per cycle indicated that selection reduced days-to-flowering by 1.25 days per cycle. Grain yield and plant height were reduced by 25 and 8%, respectively. Although the selection was effective, this popul ation resulted in an unacceptably high frequency of aneuploid and phen otypically off-type progeny. Cytological analysis of the original F-1 hybrids indicated the presence of chromosomal interchanges between som e of the C-0 parents. We hypothesize that the absence of a selfing gen eration, in a closed recurrent selection system enhanced chromosomal i nterchange differences, resulting in a high frequency of off-types and aneuploids in later cycles. Therefore, if interchanges are present be tween breeding lines, recurrent selection without selfing in oat may n ot be a useful method for long-term selection programs.