Early-generation selection between and within pair crosses in a potato (Solanum tuberosum subsp. tuberosum) breeding programme

 In 1992, 72 seedlings from each of 198 pair crosses were grown in a glasshouse, and the tubers produced by each plant were visually assessed on a 1–9 scale of increasing preference. Three groups of four progenies with high, medium and low mean scores were chosen to progress, without selection via tuber progenies and four-plant plots at a high-grade seed site, to replicated yield trials in the third clonal generation. The three groups maintained their high, medium and low scores for visual preference over the three clonal generations and also had high, medium and low scores in the second and third clonal generations for yield, size and appearance of tubers, all of which were components of visual preference. The three groups were predicted to have 13.6%, 1.8% and 0.2% of their clones exceeding the mean of 13 control cultivars for visual preference in the replicated trials, and 12.1%, 4.9% and 1.4% for yield, and 56.8%, 37.1% and 14.8% for appearance. The experiment confirmed that selection for visual preference within crosses in the seedling and first clonal generations is very ineffective, but that worthwhile progress can be made from selection in the second clonal generation, with correlated responses for faster emergence, earlier maturity, higher yield and greater regularity of shape (appearance). Combining selection of the high group of progenies with selection in the second clonal generation of the best 34 out of the 120 clones in this group, produced a response in visual preference in the third clonal generation of 1.00 compared with a maximum possible of 1.74. Ways of achieving further improvements in early-generation selection are discussed.