The efficiency of visual assessment for yield and its components in winter oat rows

The efficiency of visual assessment of grain yield and yield components in 0–75 m rows of winter oats was, in general, moderately high. Only eight or less lines were needed to be saved in six out of eight cases to retain at least 50% of the best four lines from the set of 40 lines. Assessments of yield/row and tillers/row were more effective than assessments of grains/panicle and 1000-grain weight on the basis of the proportion of variation accounted for by regression of visual assessment on actual performance. An experienced oat breeder was more successful in his assessment than another assessor unfamiliar with the crop. Assessments of yield/row by the experienced assessor were apparently biased towards assessments of grains/row as the single most easily assessed character, while the assessor unfamiliar with the crop attempted to take tillers/row and 1000-grain weight into account in his assessments. Implications to a winter oat breeding programme are indicated.