Continued advances in oat (Avena sativa L.) breeding rely in part on t
he improvement of test weight and grain yield while maintaining geneti
c variability. The objective or this study was to determine the effica
cy of three cycles of recurrent selection for high test weight (HT pop
ulation) or Smith-Hazel index of test weight and grain yield (HTG popu
lation) in oat. The recurrent selection procedure consisted of interma
ting selected S0:1 lines, growing the S0 seeds to produce S0:1 lines,
and field evaluating the S0:1 lines. Each cycle took 1 yr to complete
and about 300 lines were evaluated in each population of each cycle. T
wenty S0:1 lines were selected from each population to initiate the ne
xt cycle. Response to selection in a population was evaluated by testi
ng 95 S0:1 lines randomly chosen from among Cycles 0, 1, and 2, the co
mplete Cycle 3 population, parents from all cycles, and a set of check
cultivars. There were significant (P < 0.05) increases in mean test w
eight of 17.8 +/- 3.5 kg m-3 and 6.9 +/- 1.5 kg m-3 per cycle in the H
T and HTG populations, respectively. For grain yield, there was a sign
ificant decrease of 0. 11 +/- 0.03 Mg ha - 1 per cycle in HT populatio
ns, and an increase of 0. 12 +/- 0.07 Mg ha - 1 per cycle in HTG popul
ations. Broad-sense heritabilities and genotypic variances remained hi
gh in all cycles of each population. The results of this study show th
at recurrent selection in oat is an effective breeding procedure for i
mproving test weight and grain yield without significant reductions in
genetic variability.